Impressum Nonnegative_Lebesgue_Integration.thy

(*  Title:      HOL/Analysis/Nonnegative_Lebesgue_Integration.thy
  Author: Johannes Hölzl, TU München
  Author: Armin Heller, TU München
*)

section ‹Lebesgue Integration for Nonnegative Functions›

theory Nonnegative_Lebesgue_Integration
  imports Measure_Space Borel_Space
begin

subsection🍋‹tag unimportant› ‹Approximating functions›

lemma AE_upper_bound_inf_ennreal:
  fixes F G::"'a ==> ennreal"
  assumes "∧e. (e::real) > 0 ==> AE x in M. F x ≤ G x + e"
  shows "AE x in M. F x ≤ G x"
proof -
  have "AE x in M. ∀n::nat. F x ≤ G x + ennreal (1 / Suc n)"
    using assms by (auto simp: AE_all_countable)
  then show ?thesis
  proof (eventually_elim)
    fix x assume x: "∀n::nat. F x ≤ G x + ennreal (1 / Suc n)"
    show "F x ≤ G x"
    proof (rule ennreal_le_epsilon)
      fix e :: real assume "0 < e"
      then obtain n where n: "1 / Suc n < e"
        by (blast elim: nat_approx_posE)
      have "F x ≤ G x + 1 / Suc n"
        using x by simp
      also have "… ≤ G x + e"
        using n by (intro add_mono ennreal_leI) auto
      finally show "F x ≤ G x + ennreal e" .
    qed
  qed
qed

lemma AE_upper_bound_inf:
  fixes F G::"'a ==> real"
  assumes "∧e. e > 0 ==> AE x in M. F x ≤ G x + e"
  shows "AE x in M. F x ≤ G x"
proof -
  have "AE x in M. F x ≤ G x + 1/real (n+1)" for n::nat
    by (rule assms, auto)
  then have "AE x in M. ∀n::nat ∈ UNIV. F x ≤ G x + 1/real (n+1)"
    by (rule AE_ball_countable', auto)
  moreover
  {
    fix x assume i: "∀n::nat ∈ UNIV. F x ≤ G x + 1/real (n+1)"
    have "(λn. G x + 1/real (n+1)) <---- G x + 0"
      by (rule tendsto_add, simp, rule LIMSEQ_ignore_initial_segment[OF lim_1_over_n, of 1])
    then have "F x ≤ G x" using i LIMSEQ_le_const by fastforce
  }
  ultimately show ?thesis by auto
qed

lemma not_AE_zero_ennreal_E:
  fixes f::"'a ==> ennreal"
  assumes "¬ (AE x in M. f x = 0)" and [measurable]: "f ∈ borel_measurable M"
  shows "∃A∈sets M. ∃e::real>0. emeasure M A > 0 ∧ (∀x ∈ A. f x ≥ e)"
proof -
  { assume "¬ (∃e::real>0. {x ∈ space M. f x ≥ e} ∉ null_sets M)"
    then have "0 < e ==> AE x in M. f x ≤ e" for e :: real
      by (auto simp: not_le less_imp_le dest!: AE_not_in)
    then have "AE x in M. f x ≤ 0"
      by (intro AE_upper_bound_inf_ennreal[where G="λ_. 0"]) simp
    then have False
      using assms by auto }
  then obtain e::real where e: "e > 0" "{x ∈ space M. f x ≥ e} ∉ null_sets M" by auto
  define A where "A = {x ∈ space M. f x ≥ e}"
  have 1 [measurable]: "A ∈ sets M" unfolding A_def by auto
  have 2: "emeasure M A > 0"
    using e(2) A_def ‹A ∈ sets M› by auto
  have 3: "∧x. x ∈ A ==> f x ≥ e" unfolding A_def by auto
  show ?thesis using e(1) 1 2 3 by blast
qed

lemma not_AE_zero_E:
  fixes f::"'a ==> real"
  assumes "AE x in M. f x ≥ 0"
          "¬(AE x in M. f x = 0)"
      and [measurable]: "f ∈ borel_measurable M"
  shows "∃A e. A ∈ sets M ∧ e>0 ∧ emeasure M A > 0 ∧ (∀x ∈ A. f x ≥ e)"
proof -
  have "∃e. e > 0 ∧ {x ∈ space M. f x ≥ e} ∉ null_sets M"
  proof (rule ccontr)
    assume *: "¬(∃e. e > 0 ∧ {x ∈ space M. f x ≥ e} ∉ null_sets M)"
    {
      fix e::real assume "e > 0"
      then have "{x ∈ space M. f x ≥ e} ∈ null_sets M" using * by blast
      then have "AE x in M. x ∉ {x ∈ space M. f x ≥ e}" using AE_not_in by blast
      then have "AE x in M. f x ≤ e" by auto
    }
    then have "AE x in M. f x ≤ 0" by (rule AE_upper_bound_inf, auto)
    then have "AE x in M. f x = 0" using assms(1) by auto
    then show False using assms(2) by auto
  qed
  then obtain e where e: "e>0" "{x ∈ space M. f x ≥ e} ∉ null_sets M" by auto
  define A where "A = {x ∈ space M. f x ≥ e}"
  have 1 [measurable]: "A ∈ sets M" unfolding A_def by auto
  have 2: "emeasure M A > 0"
    using e(2) A_def ‹A ∈ sets M› by auto
  have 3: "∧x. x ∈ A ==> f x ≥ e" unfolding A_def by auto
  show ?thesis
    using e(1) 1 2 3 by blast
qed

subsection "Simple function"

text ‹
 
 Our simple functions are not restricted to nonnegative real numbers. Instead
 they are just functions with a finite range and are measurable when singleton
 sets are measurable.
 
 ›

definition🍋‹tag important› "simple_function M g ⟷
    finite (g ` space M) ∧
    (∀x ∈ g ` space M. g -` {x} ∩ space M ∈ sets M)"

lemma simple_functionD:
  assumes "simple_function M g"
  shows "finite (g ` space M)" and "g -` X ∩ space M ∈ sets M"
proof -
  show "finite (g ` space M)"
    using assms unfolding simple_function_def by auto
  have "g -` X ∩ space M = g -` (X ∩ g`space M) ∩ space M" by auto
  also have "… = (∪x∈X ∩ g`space M. g-`{x} ∩ space M)" by auto
  finally show "g -` X ∩ space M ∈ sets M" using assms
    by (auto simp del: UN_simps simp: simple_function_def)
qed

lemma measurable_simple_function[measurable_dest]:
  "simple_function M f ==> f ∈ measurable M (count_space UNIV)"
  unfolding simple_function_def measurable_def
proof safe
  fix A assume "finite (f ` space M)" "∀x∈f ` space M. f -` {x} ∩ space M ∈ sets M"
  then have "(∪x∈f ` space M. if x ∈ A then f -` {x} ∩ space M else {}) ∈ sets M"
    by (intro sets.finite_UN) auto
  also have "(∪x∈f ` space M. if x ∈ A then f -` {x} ∩ space M else {}) = f -` A ∩ space M"
    by (auto split: if_split_asm)
  finally show "f -` A ∩ space M ∈ sets M" .
qed simp

lemma borel_measurable_simple_function:
  "simple_function M f ==> f ∈ borel_measurable M"
  by (auto dest!: measurable_simple_function simp: measurable_def)

lemma simple_function_measurable2[intro]:
  assumes "simple_function M f" "simple_function M g"
  shows "f -` A ∩ g -` B ∩ space M ∈ sets M"
proof -
  have "f -` A ∩ g -` B ∩ space M = (f -` A ∩ space M) ∩ (g -` B ∩ space M)"
    by auto
  then show ?thesis using assms[THEN simple_functionD(2)] by auto
qed

lemma simple_function_indicator_representation:
  fixes f ::"'a ==> ennreal"
  assumes f: "simple_function M f" and x: "x ∈ space M"
  shows "f x = (∑y ∈ f ` space M. y * indicator (f -` {y} ∩ space M) x)"
  (is "?l = ?r")
proof -
  have "?r = (∑y ∈ f ` space M.
    (if y = f x then y * indicator (f -` {y} ∩ space M) x else 0))"
    by (auto intro!: sum.cong)
  also have "... = f x * indicator (f -` {f x} ∩ space M) x"
    using assms by (auto dest: simple_functionD)
  also have "... = f x" using x by (auto simp: indicator_def)
  finally show ?thesis by auto
qed

lemma simple_function_notspace:
  "simple_function M (λx. h x * indicator (- space M) x::ennreal)" (is "simple_function M ?h")
proof -
  have "?h ` space M ⊆ {0}" unfolding indicator_def by auto
  hence [simp, intro]: "finite (?h ` space M)" by (auto intro: finite_subset)
  have "?h -` {0} ∩ space M = space M" by auto
  thus ?thesis unfolding simple_function_def by (auto simp add: image_constant_conv)
qed

lemma simple_function_cong:
  assumes "∧t. t ∈ space M ==> f t = g t"
  shows "simple_function M f ⟷ simple_function M g"
proof -
  have "∧x. f -` {x} ∩ space M = g -` {x} ∩ space M"
    using assms by auto
  with assms show ?thesis
    by (simp add: simple_function_def cong: image_cong)
qed

lemma simple_function_cong_algebra:
  assumes "sets N = sets M" "space N = space M"
  shows "simple_function M f ⟷ simple_function N f"
  unfolding simple_function_def assms ..

lemma simple_function_borel_measurable:
  fixes f :: "'a ==> 'x::{t2_space}"
  assumes "f ∈ borel_measurable M" and "finite (f ` space M)"
  shows "simple_function M f"
  using assms unfolding simple_function_def
  by (auto intro: borel_measurable_vimage)

lemma simple_function_iff_borel_measurable:
  fixes f :: "'a ==> 'x::{t2_space}"
  shows "simple_function M f ⟷ finite (f ` space M) ∧ f ∈ borel_measurable M"
  by (metis borel_measurable_simple_function simple_functionD(1) simple_function_borel_measurable)

lemma simple_function_eq_measurable:
  "simple_function M f ⟷ finite (f`space M) ∧ f ∈ measurable M (count_space UNIV)"
  using measurable_simple_function[of M f] by (fastforce simp: simple_function_def)

lemma simple_function_const[intro, simp]:
  "simple_function M (λx. c)"
  by (auto intro: finite_subset simp: simple_function_def)
lemma simple_function_compose[intro, simp]:
  assumes "simple_function M f"
  shows "simple_function M (g ∘ f)"
  unfolding simple_function_def
proof safe
  show "finite ((g ∘ f) ` space M)"
    using assms unfolding simple_function_def image_comp [symmetric] by auto
next
  fix x assume "x ∈ space M"
  let ?G = "g -` {g (f x)} ∩ (f`space M)"
  have *: "(g ∘ f) -` {(g ∘ f) x} ∩ space M =
    (∪x∈?G. f -` {x} ∩ space M)" by auto
  show "(g ∘ f) -` {(g ∘ f) x} ∩ space M ∈ sets M"
    using assms unfolding simple_function_def *
    by (rule_tac sets.finite_UN) auto
qed

lemma simple_function_indicator[intro, simp]:
  assumes "A ∈ sets M"
  shows "simple_function M (indicator A)"
proof -
  have "indicator A ` space M ⊆ {0, 1}" (is "?S ⊆ _")
    by (auto simp: indicator_def)
  hence "finite ?S" by (rule finite_subset) simp
  moreover have "- A ∩ space M = space M - A" by auto
  ultimately show ?thesis unfolding simple_function_def
    using assms by (auto simp: indicator_def [abs_def])
qed

lemma simple_function_Pair[intro, simp]:
  assumes "simple_function M f"
  assumes "simple_function M g"
  shows "simple_function M (λx. (f x, g x))" (is "simple_function M ?p")
  unfolding simple_function_def
proof safe
  show "finite (?p ` space M)"
    using assms unfolding simple_function_def
    by (rule_tac finite_subset[of _ "f`space M × g`space M"]) auto
next
  fix x assume "x ∈ space M"
  have "(λx. (f x, g x)) -` {(f x, g x)} ∩ space M =
      (f -` {f x} ∩ space M) ∩ (g -` {g x} ∩ space M)"
    by auto
  with ‹x ∈ space M› show "(λx. (f x, g x)) -` {(f x, g x)} ∩ space M ∈ sets M"
    using assms unfolding simple_function_def by auto
qed

lemma simple_function_compose1:
  assumes "simple_function M f"
  shows "simple_function M (λx. g (f x))"
  using simple_function_compose[OF assms, of g]
  by (simp add: comp_def)

lemma simple_function_compose2:
  assumes "simple_function M f" and "simple_function M g"
  shows "simple_function M (λx. h (f x) (g x))"
proof -
  have "simple_function M ((λ(x, y). h x y) ∘ (λx. (f x, g x)))"
    using assms by auto
  thus ?thesis by (simp_all add: comp_def)
qed

lemmas simple_function_add[intro, simp] = simple_function_compose2[where h="(+)"]
  and simple_function_diff[intro, simp] = simple_function_compose2[where h="(-)"]
  and simple_function_uminus[intro, simp] = simple_function_compose[where g="uminus"]
  and simple_function_mult[intro, simp] = simple_function_compose2[where h="(*)"]
  and simple_function_div[intro, simp] = simple_function_compose2[where h="(/)"]
  and simple_function_inverse[intro, simp] = simple_function_compose[where g="inverse"]
  and simple_function_max[intro, simp] = simple_function_compose2[where h=max]

lemma simple_function_sum[intro, simp]:
  assumes "∧i. i ∈ P ==> simple_function M (f i)"
  shows "simple_function M (λx. ∑i∈P. f i x)"
proof cases
  assume "finite P" from this assms show ?thesis by induct auto
qed auto

lemma simple_function_ennreal[intro, simp]:
  fixes f g :: "'a ==> real" assumes sf: "simple_function M f"
  shows "simple_function M (λx. ennreal (f x))"
  by (rule simple_function_compose1[OF sf])

lemma simple_function_real_of_nat[intro, simp]:
  fixes f g :: "'a ==> nat" assumes sf: "simple_function M f"
  shows "simple_function M (λx. real (f x))"
  by (rule simple_function_compose1[OF sf])

lemma🍋‹tag important› borel_measurable_implies_simple_function_sequence:
  fixes u :: "'a ==> ennreal"
  assumes u[measurable]: "u ∈ borel_measurable M"
  shows "∃f. incseq f ∧ (∀i. (∀x. f i x < top) ∧ simple_function M (f i)) ∧ u = (SUP i. f i)"
proof -
  define f where [abs_def]:
    "f i x = real_of_int (floor (enn2real (min i (u x)) * 2^i)) / 2^i" for i x

  have [simp]: "0 ≤ f i x" for i x
    by (auto simp: f_def intro!: divide_nonneg_nonneg mult_nonneg_nonneg enn2real_nonneg)

  have *: "2^n * real_of_int x = real_of_int (2^n * x)" for n x
    by simp

  have "real_of_int ⌊real i * 2 ^ i⌋ = real_of_int ⌊i * 2 ^ i⌋" for i
    by (intro arg_cong[where f=real_of_int]) simp
  then have [simp]: "real_of_int ⌊real i * 2 ^ i⌋ = i * 2 ^ i" for i
    unfolding floor_of_nat by simp

  have "incseq f"
  proof (intro monoI le_funI)
    fix m n :: nat and x assume "m ≤ n"
    moreover
    { fix d :: nat
      have "⌊2^d::real⌋ * ⌊2^m * enn2real (min (of_nat m) (u x))⌋ ≤
        ⌊2^d * (2^m * enn2real (min (of_nat m) (u x)))⌋"
        by (rule le_mult_floor) (auto)
      also have "… ≤ ⌊2^d * (2^m * enn2real (min (of_nat d + of_nat m) (u x)))⌋"
        by (intro floor_mono mult_mono enn2real_mono min.mono)
           (auto simp: min_less_iff_disj of_nat_less_top)
      finally have "f m x ≤ f (m + d) x"
        unfolding f_def
        by (auto simp: field_simps power_add * simp del: of_int_mult) }
    ultimately show "f m x ≤ f n x"
      by (auto simp add: le_iff_add)
  qed
  then have inc_f: "incseq (λi. ennreal (f i x))" for x
    by (auto simp: incseq_def le_fun_def)
  then have "incseq (λi x. ennreal (f i x))"
    by (auto simp: incseq_def le_fun_def)
  moreover
  have "simple_function M (f i)" for i
  proof (rule simple_function_borel_measurable)
    have "⌊enn2real (min (of_nat i) (u x)) * 2 ^ i⌋ ≤ ⌊int i * 2 ^ i⌋" for x
      by (cases "u x" rule: ennreal_cases)
         (auto split: split_min intro!: floor_mono)
    then have "f i ` space M ⊆ (λn. real_of_int n / 2^i) ` {0 .. of_nat i * 2^i}"
      unfolding floor_of_int by (auto simp: f_def intro!: imageI)
    then show "finite (f i ` space M)"
      by (rule finite_subset) auto
    show "f i ∈ borel_measurable M"
      unfolding f_def enn2real_def by measurable
  qed
  moreover
  { fix x
    have "(SUP i. ennreal (f i x)) = u x"
    proof (cases "u x" rule: ennreal_cases)
      case top then show ?thesis
        by (simp add: f_def inf_min[symmetric] ennreal_of_nat_eq_real_of_nat[symmetric]
                      ennreal_SUP_of_nat_eq_top)
    next
      case (real r)
      obtain n where "r ≤ of_nat n" using real_arch_simple by auto
      then have min_eq_r: "∀🪙F x in sequentially. min (real x) r = r"
        by (auto simp: eventually_sequentially intro!: exI[of _ n] split: split_min)

      have "(λi. real_of_int ⌊min (real i) r * 2^i⌋ / 2^i) <---- r"
      proof (rule tendsto_sandwich)
        show "(λn. r - (1/2)^n) <---- r"
          by (auto intro!: tendsto_eq_intros LIMSEQ_power_zero)
        show "∀🪙F n in sequentially. real_of_int ⌊min (real n) r * 2 ^ n⌋ / 2 ^ n ≤ r"
          using min_eq_r by eventually_elim (auto simp: field_simps)
        have *: "r * (2 ^ n * 2 ^ n) ≤ 2^n + 2^n * real_of_int ⌊r * 2 ^ n⌋" for n
          using real_of_int_floor_ge_diff_one[of "r * 2^n", THEN mult_left_mono, of "2^n"]
          by (auto simp: field_simps)
        show "∀🪙F n in sequentially. r - (1/2)^n ≤ real_of_int ⌊min (real n) r * 2 ^ n⌋ / 2 ^ n"
          using min_eq_r by eventually_elim (insert *, auto simp: field_simps)
      qed auto
      then have "(λi. ennreal (f i x)) <---- ennreal r"
        by (simp add: real f_def ennreal_of_nat_eq_real_of_nat min_ennreal)
      from LIMSEQ_unique[OF LIMSEQ_SUP[OF inc_f] this]
      show ?thesis
        by (simp add: real)
    qed }
  ultimately show ?thesis
    by (intro exI [of _ "λi x. ennreal (f i x)"]) (auto simp add: image_comp)
qed

lemma borel_measurable_implies_simple_function_sequence':
  fixes u :: "'a ==> ennreal"
  assumes u: "u ∈ borel_measurable M"
  obtains f where
    "∧i. simple_function M (f i)" "incseq f" "∧i x. f i x < top" "∧x. (SUP i. f i x) = u x"
  using borel_measurable_implies_simple_function_sequence [OF u]
  by (metis SUP_apply)

lemma🍋‹tag important› simple_function_induct
    [consumes 1, case_names cong set mult add, induct set: simple_function]:
  fixes u :: "'a ==> ennreal"
  assumes u: "simple_function M u"
  assumes cong: "∧f g. simple_function M f ==> simple_function M g ==> (AE x in M. f x = g x) ==> P f ==> P g"
  assumes set: "∧A. A ∈ sets M ==> P (indicator A)"
  assumes mult: "∧u c. P u ==> P (λx. c * u x)"
  assumes add: "∧u v. P u ==> P v ==> P (λx. v x + u x)"
  shows "P u"
proof (rule cong)
  from AE_space show "AE x in M. (∑y∈u ` space M. y * indicator (u -` {y} ∩ space M) x) = u x"
  proof eventually_elim
    fix x assume x: "x ∈ space M"
    from simple_function_indicator_representation[OF u x]
    show "(∑y∈u ` space M. y * indicator (u -` {y} ∩ space M) x) = u x" ..
  qed
next
  from u have "finite (u ` space M)"
    unfolding simple_function_def by auto
  then show "P (λx. ∑y∈u ` space M. y * indicator (u -` {y} ∩ space M) x)"
  proof induct
    case empty show ?case
      using set[of "{}"] by (simp add: indicator_def[abs_def])
  qed (auto intro!: add mult set simple_functionD u)
next
  show "simple_function M (λx. (∑y∈u ` space M. y * indicator (u -` {y} ∩ space M) x))"
    apply (subst simple_function_cong)
    apply (rule simple_function_indicator_representation[symmetric])
    apply (auto intro: u)
    done
qed fact

lemma simple_function_induct_nn[consumes 1, case_names cong set mult add]:
  fixes u :: "'a ==> ennreal"
  assumes u: "simple_function M u"
  assumes cong: "∧f g. simple_function M f ==> simple_function M g ==> (∧x. x ∈ space M ==> f x = g x) ==> P f ==> P g"
  assumes set: "∧A. A ∈ sets M ==> P (indicator A)"
  assumes mult: "∧u c. simple_function M u ==> P u ==> P (λx. c * u x)"
  assumes add: "∧u v. simple_function M u ==> P u ==> simple_function M v ==> (∧x. x ∈ space M ==> u x = 0 ∨ v x = 0) ==> P v ==> P (λx. v x + u x)"
  shows "P u"
proof -
  show ?thesis
  proof (rule cong)
    fix x assume x: "x ∈ space M"
    from simple_function_indicator_representation[OF u x]
    show "(∑y∈u ` space M. y * indicator (u -` {y} ∩ space M) x) = u x" ..
  next
    show "simple_function M (λx. (∑y∈u ` space M. y * indicator (u -` {y} ∩ space M) x))"
      apply (subst simple_function_cong)
      apply (rule simple_function_indicator_representation[symmetric])
      apply (auto intro: u)
      done
  next
    from u have "finite (u ` space M)"
      unfolding simple_function_def by auto
    then show "P (λx. ∑y∈u ` space M. y * indicator (u -` {y} ∩ space M) x)"
    proof induct
      case empty show ?case
        using set[of "{}"] by (simp add: indicator_def[abs_def])
    next
      case (insert x S)
      { fix z have "(∑y∈S. y * indicator (u -` {y} ∩ space M) z) = 0 ∨
          x * indicator (u -` {x} ∩ space M) z = 0"
          using insert by (subst sum_eq_0_iff) (auto simp: indicator_def) }
      note disj = this
      from insert show ?case
        by (auto intro!: add mult set simple_functionD u simple_function_sum disj)
    qed
  qed fact
qed

lemma🍋‹tag important› borel_measurable_induct
    [consumes 1, case_names cong set mult add seq, induct set: borel_measurable]:
  fixes u :: "'a ==> ennreal"
  assumes u: "u ∈ borel_measurable M"
  assumes cong: "∧f g. f ∈ borel_measurable M ==> g ∈ borel_measurable M ==> (∧x. x ∈ space M ==> f x = g x) ==> P g ==> P f"
  assumes set: "∧A. A ∈ sets M ==> P (indicator A)"
  assumes mult': "∧u c. c < top ==> u ∈ borel_measurable M ==> (∧x. x ∈ space M ==> u x < top) ==> P u ==> P (λx. c * u x)"
  assumes add: "∧u v. u ∈ borel_measurable M==> (∧x. x ∈ space M ==> u x < top) ==> P u ==> v ∈ borel_measurable M ==> (∧x. x ∈ space M ==> v x < top) ==> (∧x. x ∈ space M ==> u x = 0 ∨ v x = 0) ==> P v ==> P (λx. v x + u x)"
  assumes seq: "∧U. (∧i. U i ∈ borel_measurable M) ==> (∧i x. x ∈ space M ==> U i x < top) ==> (∧i. P (U i)) ==> incseq U ==> u = (SUP i. U i) ==> P (SUP i. U i)"
  shows "P u"
  using u
proof (induct rule: borel_measurable_implies_simple_function_sequence')
  fix U assume U: "∧i. simple_function M (U i)" "incseq U" "∧i x. U i x < top" and sup: "∧x. (SUP i. U i x) = u x"
  have u_eq: "u = (SUP i. U i)"
    using u by (auto simp add: image_comp sup)

  have not_inf: "∧x i. x ∈ space M ==> U i x < top"
    using U by (auto simp: image_iff eq_commute)

  from U have "∧i. U i ∈ borel_measurable M"
    by (simp add: borel_measurable_simple_function)

  show "P u"
    unfolding u_eq
  proof (rule seq)
    fix i show "P (U i)"
      using ‹simple_function M (U i)› not_inf[of _ i]
    proof (induct rule: simple_function_induct_nn)
      case (mult u c)
      show ?case
      proof cases
        assume "c = 0 ∨ space M = {} ∨ (∀x∈space M. u x = 0)"
        with mult(1) show ?thesis
          by (intro cong[of "λx. c * u x" "indicator {}"] set)
             (auto dest!: borel_measurable_simple_function)
      next
        assume "¬ (c = 0 ∨ space M = {} ∨ (∀x∈space M. u x = 0))"
        then obtain x where "space M ≠ {}" and x: "x ∈ space M" "u x ≠ 0" "c ≠ 0"
          by auto
        with mult(3)[of x] have "c < top"
          by (auto simp: ennreal_mult_less_top)
        then have u_fin: "x' ∈ space M ==> u x' < top" for x'
          using mult(3)[of x'] ‹c ≠ 0› by (auto simp: ennreal_mult_less_top)
        then have "P u"
          by (rule mult)
        with u_fin ‹c < top› mult(1) show ?thesis
          by (intro mult') (auto dest!: borel_measurable_simple_function)
      qed
    qed (auto intro: cong intro!: set add dest!: borel_measurable_simple_function)
  qed fact+
qed

lemma simple_function_If_set:
  assumes sf: "simple_function M f" "simple_function M g" and A: "A ∩ space M ∈ sets M"
  shows "simple_function M (λx. if x ∈ A then f x else g x)" (is "simple_function M ?IF")
proof -
  define F where "F x = f -` {x} ∩ space M" for x
  define G where "G x = g -` {x} ∩ space M" for x
  show ?thesis unfolding simple_function_def
  proof safe
    have "?IF ` space M ⊆ f ` space M ∪ g ` space M" by auto
    from finite_subset[OF this] assms
    show "finite (?IF ` space M)" unfolding simple_function_def by auto
  next
    fix x assume "x ∈ space M"
    then have *: "?IF -` {?IF x} ∩ space M = (if x ∈ A
      then ((F (f x) ∩ (A ∩ space M)) ∪ (G (f x) - (G (f x) ∩ (A ∩ space M))))
      else ((F (g x) ∩ (A ∩ space M)) ∪ (G (g x) - (G (g x) ∩ (A ∩ space M)))))"
      using sets.sets_into_space[OF A] by (auto split: if_split_asm simp: G_def F_def)
    have [intro]: "∧x. F x ∈ sets M" "∧x. G x ∈ sets M"
      unfolding F_def G_def using sf[THEN simple_functionD(2)] by auto
    show "?IF -` {?IF x} ∩ space M ∈ sets M" unfolding * using A by auto
  qed
qed

lemma simple_function_If:
  assumes sf: "simple_function M f" "simple_function M g" and P: "{x∈space M. P x} ∈ sets M"
  shows "simple_function M (λx. if P x then f x else g x)"
proof -
  have "{x∈space M. P x} = {x. P x} ∩ space M" by auto
  with simple_function_If_set[OF sf, of "{x. P x}"] P show ?thesis by simp
qed

lemma simple_function_subalgebra:
  assumes "simple_function N f"
  and N_subalgebra: "sets N ⊆ sets M" "space N = space M"
  shows "simple_function M f"
  using assms unfolding simple_function_def by auto

lemma simple_function_comp:
  assumes T: "T ∈ measurable M M'"
    and f: "simple_function M' f"
  shows "simple_function M (λx. f (T x))"
proof (intro simple_function_def[THEN iffD2] conjI ballI)
  have "(λx. f (T x)) ` space M ⊆ f ` space M'"
    using T unfolding measurable_def by auto
  then show "finite ((λx. f (T x)) ` space M)"
    using f unfolding simple_function_def by (auto intro: finite_subset)
  fix i assume i: "i ∈ (λx. f (T x)) ` space M"
  then have "i ∈ f ` space M'"
    using T unfolding measurable_def by auto
  then have "f -` {i} ∩ space M' ∈ sets M'"
    using f unfolding simple_function_def by auto
  then have "T -` (f -` {i} ∩ space M') ∩ space M ∈ sets M"
    using T unfolding measurable_def by auto
  also have "T -` (f -` {i} ∩ space M') ∩ space M = (λx. f (T x)) -` {i} ∩ space M"
    using T unfolding measurable_def by auto
  finally show "(λx. f (T x)) -` {i} ∩ space M ∈ sets M" .
qed

subsection "Simple integral"

definition🍋‹tag important› simple_integral :: "'a measure ==> ('a ==> ennreal) ==> ennreal" (‹integral🪙S›) where
  "integral🪙S M f = (∑x ∈ f ` space M. x * emeasure M (f -` {x} ∩ space M))"

syntax
  "_simple_integral" :: "pttrn ==> ennreal ==> 'a measure ==> ennreal"
    (‹(‹open_block notation=‹binder integral›\∫🪙S _. _ ∂_)› [60,61] 110)
 syntax_consts
  "_simple_integral" == simple_integral
 translations
  "∫🪙S x. f ∂M" == "CONST simple_integral M (%x. f)"
 
 lemma simple_integral_cong:
  assumes "∧t. t ∈ space M ==> f t = g t"
  shows "integral🪙S M f = integral🪙S M g"
 proof -
  have "f ` space M = g ` space M"
  "∧x. f -` {x} ∩ space M = g -` {x} ∩ space M"
  using assms by (auto intro!: image_eqI)
  thus ?thesis unfolding simple_integral_def by simp
 qed
 
 lemma simple_integral_const[simp]:
  "(∫🪙Sx. c ∂M) = c * (emeasure M) (space M)"
 proof (cases "space M = {}")
  case True thus ?thesis unfolding simple_integral_def by simp
 next
  case False hence "(λx. c) ` space M = {c}" by auto
  thus ?thesis unfolding simple_integral_def by simp
 qed
 
 lemma simple_function_partition:
  assumes f: "simple_function M f" and g: "simple_function M g"
  assumes sub: "∧x y. x ∈ space M ==> y ∈ space M ==> g x = g y ==> f x = f y"
  assumes v: "∧x. x ∈ space M ==> f x = v (g x)"
  shows "integral🪙S M f = (∑y∈g ` space M. v y * emeasure M {x∈space M. g x = y})"
  (is "_ = ?r")
 proof -
  from f g have [simp]: "finite (f`space M)" "finite (g`space M)"
  by (auto simp: simple_function_def)
  from f g have [measurable]: "f ∈ measurable M (count_space UNIV)" "g ∈ measurable M (count_space UNIV)"
  by (auto intro: measurable_simple_function)
 
  { fix y assume "y ∈ space M"
  then have "f ` space M ∩ {i. ∃x∈space M. i = f x ∧ g y = g x} = {v (g y)}"
  by (auto cong: sub simp: v[symmetric]) }
  note eq = this
 
  have "integral🪙S M f =
  (∑y∈f`space M. y * (∑z∈g`space M.
  if ∃x∈space M. y = f x ∧ z = g x then emeasure M {x∈space M. g x = z} else 0))"
  unfolding simple_integral_def
  proof (safe intro!: sum.cong ennreal_mult_left_cong)
  fix y assume y: "y ∈ space M" "f y ≠ 0"
  have [simp]: "g ` space M ∩ {z. ∃x∈space M. f y = f x ∧ z = g x} =
  {z. ∃x∈space M. f y = f x ∧ z = g x}"
  by auto
  have eq:"(∪i∈{z. ∃x∈space M. f y = f x ∧ z = g x}. {x ∈ space M. g x = i}) =
  f -` {f y} ∩ space M"
  by (auto simp: eq_commute cong: sub rev_conj_cong)
  have "finite (g`space M)" by simp
  then have "finite {z. ∃x∈space M. f y = f x ∧ z = g x}"
  by (rule rev_finite_subset) auto
  then show "emeasure M (f -` {f y} ∩ space M) =
  (∑z∈g ` space M. if ∃x∈space M. f y = f x ∧ z = g x then emeasure M {x ∈ space M. g x = z} else 0)"
  apply (simp add: sum.If_cases)
  apply (subst sum_emeasure)
  apply (auto simp: disjoint_family_on_def eq)
  done
  qed
  also have "… = (∑y∈f`space M. (∑z∈g`space M.
  if ∃x∈space M. y = f x ∧ z = g x then y * emeasure M {x∈space M. g x = z} else 0))"
  by (auto intro!: sum.cong simp: sum_distrib_left)
  also have "… = ?r"
  by (subst sum.swap)
  (auto intro!: sum.cong simp: sum.If_cases scaleR_sum_right[symmetric] eq)
  finally show "integral🪙S M f = ?r" .
 qed
 
 lemma simple_integral_add[simp]:
  assumes f: "simple_function M f" and "∧x. 0 ≤ f x" and g: "simple_function M g" and "∧x. 0 ≤ g x"
  shows "(∫🪙Sx. f x + g x ∂M) = integral🪙S M f + integral🪙S M g"
 proof -
  have "(∫🪙Sx. f x + g x ∂M) =
  (∑y∈(λx. (f x, g x))`space M. (fst y + snd y) * emeasure M {x∈space M. (f x, g x) = y})"
  by (intro simple_function_partition) (auto intro: f g)
  also have "… = (∑y∈(λx. (f x, g x))`space M. fst y * emeasure M {x∈space M. (f x, g x) = y}) +
  (∑y∈(λx. (f x, g x))`space M. snd y * emeasure M {x∈space M. (f x, g x) = y})"
  using assms(2,4) by (auto intro!: sum.cong distrib_right simp: sum.distrib[symmetric])
  also have "(∑y∈(λx. (f x, g x))`space M. fst y * emeasure M {x∈space M. (f x, g x) = y}) = (∫🪙Sx. f x ∂M)"
  by (intro simple_function_partition[symmetric]) (auto intro: f g)
  also have "(∑y∈(λx. (f x, g x))`space M. snd y * emeasure M {x∈space M. (f x, g x) = y}) = (∫🪙Sx. g x ∂M)"
  by (intro simple_function_partition[symmetric]) (auto intro: f g)
  finally show ?thesis .
 qed
 
 lemma simple_integral_sum[simp]:
  assumes "∧i x. i ∈ P ==> 0 ≤ f i x"
  assumes "∧i. i ∈ P ==> simple_function M (f i)"
  shows "(∫🪙Sx. (∑i∈P. f i x) ∂M) = (∑i∈P. integral🪙S M (f i))"
 proof cases
  assume "finite P"
  from this assms show ?thesis
  by induct (auto)
 qed auto
 
 lemma simple_integral_mult[simp]:
  assumes f: "simple_function M f"
  shows "(∫🪙Sx. c * f x ∂M) = c * integral🪙S M f"
 proof -
  have "(∫🪙Sx. c * f x ∂M) = (∑y∈f ` space M. (c * y) * emeasure M {x∈space M. f x = y})"
  using f by (intro simple_function_partition) auto
  also have "… = c * integral🪙S M f"
  using f unfolding simple_integral_def
  by (subst sum_distrib_left) (auto simp: mult.assoc Int_def conj_commute)
  finally show ?thesis .
 qed
 
 lemma simple_integral_mono_AE:
  assumes f[measurable]: "simple_function M f" and g[measurable]: "simple_function M g"
  and mono: "AE x in M. f x ≤ g x"
  shows "integral🪙S M f ≤ integral🪙S M g"
 proof -
  let ?μ = "λP. emeasure M {x∈space M. P x}"
  have "integral🪙S M f = (∑y∈(λx. (f x, g x))`space M. fst y * ?μ (λx. (f x, g x) = y))"
  using f g by (intro simple_function_partition) auto
  also have "… ≤ (∑y∈(λx. (f x, g x))`space M. snd y * ?μ (λx. (f x, g x) = y))"
  proof (clarsimp intro!: sum_mono)
  fix x assume "x ∈ space M"
  let ?M = "?μ (λy. f y = f x ∧ g y = g x)"
  show "f x * ?M ≤ g x * ?M"
  proof cases
  assume "?M ≠ 0"
  then have "0 < ?M"
  by (simp add: less_le)
  also have "… ≤ ?μ (λy. f x ≤ g x)"
  using mono by (force intro: emeasure_mono_AE)
  finally have "¬ ¬ f x ≤ g x"
  by (intro notI) auto
  then show ?thesis
  by (intro mult_right_mono) auto
  qed simp
  qed
  also have "… = integral🪙S M g"
  using f g by (intro simple_function_partition[symmetric]) auto
  finally show ?thesis .
 qed
 
 lemma simple_integral_mono:
  assumes "simple_function M f" and "simple_function M g"
  and mono: "∧ x. x ∈ space M ==> f x ≤ g x"
  shows "integral🪙S M f ≤ integral🪙S M g"
  using assms by (intro simple_integral_mono_AE) auto
 
 lemma simple_integral_cong_AE:
  assumes "simple_function M f" and "simple_function M g"
  and "AE x in M. f x = g x"
  shows "integral🪙S M f = integral🪙S M g"
  using assms by (auto simp: eq_iff intro!: simple_integral_mono_AE)
 
 lemma simple_integral_cong':
  assumes sf: "simple_function M f" "simple_function M g"
  and mea: "(emeasure M) {x∈space M. f x ≠ g x} = 0"
  shows "integral🪙S M f = integral🪙S M g"
 proof (intro simple_integral_cong_AE sf AE_I)
  show "(emeasure M) {x∈space M. f x ≠ g x} = 0" by fact
  show "{x ∈ space M. f x ≠ g x} ∈ sets M"
  using sf[THEN borel_measurable_simple_function] by auto
 qed simp
 
 lemma simple_integral_indicator:
  assumes A: "A ∈ sets M"
  assumes f: "simple_function M f"
  shows "(∫🪙Sx. f x * indicator A x ∂M) =
  (∑x ∈ f ` space M. x * emeasure M (f -` {x} ∩ space M ∩ A))"
 proof -
  have eq: "(λx. (f x, indicator A x)) ` space M ∩ {x. snd x = 1} = (λx. (f x, 1::ennreal))`A"
  using A[THEN sets.sets_into_space] by (auto simp: indicator_def image_iff split: if_split_asm)
  have eq2: "∧x. f x ∉ f ` A ==> f -` {f x} ∩ space M ∩ A = {}"
  by (auto simp: image_iff)
 
  have "(∫🪙Sx. f x * indicator A x ∂M) =
  (∑y∈(λx. (f x, indicator A x))`space M. (fst y * snd y) * emeasure M {x∈space M. (f x, indicator A x) = y})"
  using assms by (intro simple_function_partition) auto
  also have "… = (∑y∈(λx. (f x, indicator A x::ennreal))`space M.
  if snd y = 1 then fst y * emeasure M (f -` {fst y} ∩ space M ∩ A) else 0)"
    by (auto simp: indicator_def split: if_split_asm intro!: arg_cong2[where f="(*)
  also have "… = (∑y∈(λx. (f x, 1::ennreal))`A. fst y * emeasure M (f -` {fst y} ∩ space M ∩ A))"
    using assms by (subst sum.If_cases) (auto intro!: simple_functionD(1) simp: eq)
  also have "… = (∑y∈fst`(λx. (f x, 1::ennreal))`A. y * emeasure M (f -` {y} ∩ space M ∩ A))"
    by (subst sum.reindex [of fst]) (auto simp: inj_on_def)
  also have "… = (∑x ∈ f ` space M. x * emeasure M (f -` {x} ∩ space M ∩ A))"
    using A[THEN sets.sets_into_space]
    by (intro sum.mono_neutral_cong_left simple_functionD f) (auto simp: image_comp comp_def eq2)
  finally show ?thesis .
qed

lemma simple_integral_indicator_only[simp]:
  assumes "A ∈ sets M"
  shows "integral🪙S M (indicator A) = emeasure M A"
  using simple_integral_indicator[OF assms, of "λx. 1"] sets.sets_into_space[OF assms]
  by (simp_all add: image_constant_conv Int_absorb1 split: if_split_asm)

lemma simple_integral_null_set:
  assumes "simple_function M u" "∧x. 0 ≤ u x" and "N ∈ null_sets M"
  shows "(∫🪙Sx. u x * indicator N x ∂M) = 0"
proof -
  have "AE x in M. indicator N x = (0 :: ennreal)"
    using ‹N ∈ null_sets M› by (auto simp: indicator_def intro!: AE_I[of _ _ N])
  then have "(∫🪙Sx. u x * indicator N x ∂M) = (∫🪙Sx. 0 ∂M)"
    using assms apply (intro simple_integral_cong_AE) by auto
  then show ?thesis by simp
qed

lemma simple_integral_cong_AE_mult_indicator:
  assumes sf: "simple_function M f" and eq: "AE x in M. x ∈ S" and "S ∈ sets M"
  shows "integral🪙S M f = (∫🪙Sx. f x * indicator S x ∂M)"
  using assms by (intro simple_integral_cong_AE) auto

lemma simple_integral_cmult_indicator:
  assumes A: "A ∈ sets M"
  shows "(∫🪙Sx. c * indicator A x ∂M) = c * emeasure M A"
  using simple_integral_mult[OF simple_function_indicator[OF A]]
  unfolding simple_integral_indicator_only[OF A] by simp

lemma simple_integral_nonneg:
  assumes f: "simple_function M f" and ae: "AE x in M. 0 ≤ f x"
  shows "0 ≤ integral🪙S M f"
proof -
  have "integral🪙S M (λx. 0) ≤ integral🪙S M f"
    using simple_integral_mono_AE[OF _ f ae] by auto
  then show ?thesis by simp
qed

subsection ‹Integral on nonnegative functions›

definition🍋‹tag important› nn_integral :: "'a measure ==> ('a ==> ennreal) ==> ennreal" (‹integral🪙N›) where
  "integral🪙N M f = (SUP g ∈ {g. simple_function M g ∧ g ≤ f}. integral🪙S M g)"

syntax
  "_nn_integral" :: "pttrn ==> ennreal ==> 'a measure ==> ennreal" (‹(‹indent=2 notation=‹binder integral›\∫🪙+(2 _./ _)/ ∂_)› [60,61] 110)
 syntax_consts
  "_nn_integral" == nn_integral
 translations
  "∫🪙+x. f ∂M" == "CONST nn_integral M (λx. f)"
 
 lemma nn_integral_def_finite:
  "integral🪙N M f = (SUP g ∈ {g. simple_function M g ∧ g ≤ f ∧ (∀x. g x < top)}. integral🪙S M g)"
  (is "_ = Sup (?A ` ?f)")
  unfolding nn_integral_def
 proof (safe intro!: antisym SUP_least)
  fix g assume g[measurable]: "simple_function M g" "g ≤ f"
 
  show "integral🪙S M g ≤ Sup (?A ` ?f)"
  proof cases
  assume ae: "AE x in M. g x ≠ top"
  let ?G = "{x ∈ space M. g x ≠ top}"
  have "integral🪙S M g = integral🪙S M (λx. g x * indicator ?G x)"
  proof (rule simple_integral_cong_AE)
  show "AE x in M. g x = g x * indicator ?G x"
  using ae AE_space by eventually_elim auto
  qed (insert g, auto)
  also have "… ≤ Sup (?A ` ?f)"
  using g by (intro SUP_upper) (auto simp: le_fun_def less_top split: split_indicator)
  finally show ?thesis .
  next
  assume nAE: "¬ (AE x in M. g x ≠ top)"
  then have "emeasure M {x∈space M. g x = top} ≠ 0" (is "emeasure M ?G ≠ 0")
  by (subst (asm) AE_iff_measurable[OF _ refl]) auto
  then have "top = (SUP n. (∫🪙Sx. of_nat n * indicator ?G x ∂M))"
  by (simp add: ennreal_SUP_of_nat_eq_top ennreal_top_eq_mult_iff SUP_mult_right_ennreal[symmetric])
  also have "… ≤ Sup (?A ` ?f)"
  using g
  by (safe intro!: SUP_least SUP_upper)
  (auto simp: le_fun_def of_nat_less_top top_unique[symmetric] split: split_indicator
  intro: order_trans[of _ "g x" "f x" for x, OF order_trans[of _ top]])
  finally show ?thesis
  by (simp add: top_unique del: SUP_eq_top_iff Sup_eq_top_iff)
  qed
 qed (auto intro: SUP_upper)
 
 lemma nn_integral_mono_AE:
  assumes ae: "AE x in M. u x ≤ v x" shows "integral🪙N M u ≤ integral🪙N M v"
  unfolding nn_integral_def
 proof (safe intro!: SUP_mono)
  fix n assume n: "simple_function M n" "n ≤ u"
  from ae[THEN AE_E] obtain N
  where N: "{x ∈ space M. ¬ u x ≤ v x} ⊆ N" "emeasure M N = 0" "N ∈ sets M"
  by auto
  then have ae_N: "AE x in M. x ∉ N" by (auto intro: AE_not_in)
  let ?n = "λx. n x * indicator (space M - N) x"
  have "AE x in M. n x ≤ ?n x" "simple_function M ?n"
  using n N ae_N by auto
  moreover
  { fix x have "?n x ≤ v x"
  proof cases
  assume x: "x ∈ space M - N"
  with N have "u x ≤ v x" by auto
  with n(2)[THEN le_funD, of x] x show ?thesis
  by (auto simp: max_def split: if_split_asm)
  qed simp }
  then have "?n ≤ v" by (auto simp: le_funI)
  moreover have "integral🪙S M n ≤ integral🪙S M ?n"
  using ae_N N n by (auto intro!: simple_integral_mono_AE)
  ultimately show "∃m∈{g. simple_function M g ∧ g ≤ v}. integral🪙S M n ≤ integral🪙S M m"
  by force
 qed
 
 lemma nn_integral_mono:
  "(∧x. x ∈ space M ==> u x ≤ v x) ==> integral🪙N M u ≤ integral🪙N M v"
  by (auto intro: nn_integral_mono_AE)
 
 lemma mono_nn_integral: "mono F ==> mono (λx. integral🪙N M (F x))"
  by (auto simp add: mono_def le_fun_def intro!: nn_integral_mono)
 
 lemma nn_integral_cong_AE:
  "AE x in M. u x = v x ==> integral🪙N M u = integral🪙N M v"
  by (auto simp: eq_iff intro!: nn_integral_mono_AE)
 
 lemma nn_integral_cong:
  "(∧x. x ∈ space M ==> u x = v x) ==> integral🪙N M u = integral🪙N M v"
  by (auto intro: nn_integral_cong_AE)
 
 lemma nn_integral_cong_simp:
  "(∧x. x ∈ space M =simp=> u x = v x) ==> integral🪙N M u = integral🪙N M v"
  by (auto intro: nn_integral_cong simp: simp_implies_def)
 
 lemma incseq_nn_integral:
  assumes "incseq f" shows "incseq (λi. integral🪙N M (f i))"
 proof -
  have "∧i x. f i x ≤ f (Suc i) x"
  using assms by (auto dest!: incseq_SucD simp: le_fun_def)
  then show ?thesis
  by (auto intro!: incseq_SucI nn_integral_mono)
 qed
 
 lemma nn_integral_eq_simple_integral:
  assumes f: "simple_function M f" shows "integral🪙N M f = integral🪙S M f"
 proof -
  let ?f = "λx. f x * indicator (space M) x"
  have f': "simple_function M ?f" using f by auto
  have "integral🪙N M ?f ≤ integral🪙S M ?f" using f'
  by (force intro!: SUP_least simple_integral_mono simp: le_fun_def nn_integral_def)
  moreover have "integral🪙S M ?f ≤ integral🪙N M ?f"
  unfolding nn_integral_def
  using f' by (auto intro!: SUP_upper)
  ultimately show ?thesis
  by (simp cong: nn_integral_cong simple_integral_cong)
 qed
 
 text ‹Beppo-Levi monotone convergence theorem›
 lemma nn_integral_monotone_convergence_SUP:
  assumes f: "incseq f" and [measurable]: "∧i. f i ∈ borel_measurable M"
  shows "(∫🪙+ x. (SUP i. f i x) ∂M) = (SUP i. integral🪙N M (f i))"
 proof (rule antisym)
  show "(∫🪙+ x. (SUP i. f i x) ∂M) ≤ (SUP i. (∫🪙+ x. f i x ∂M))"
  unfolding nn_integral_def_finite[of _ "λx. SUP i. f i x"]
  proof (safe intro!: SUP_least)
  fix u assume sf_u[simp]: "simple_function M u" and
  u: "u ≤ (λx. SUP i. f i x)" and u_range: "∀x. u x < top"
  note sf_u[THEN borel_measurable_simple_function, measurable]
  show "integral🪙S M u ≤ (SUP j. ∫🪙+x. f j x ∂M)"
  proof (rule ennreal_approx_unit)
  fix a :: ennreal assume "a < 1"
  let ?au = "λx. a * u x"
 
  let ?B = "λc i. {x∈space M. ?au x = c ∧ c ≤ f i x}"
  have "integral🪙S M ?au = (∑c∈?au`space M. c * (SUP i. emeasure M (?B c i)))"
  unfolding simple_integral_def
  proof (intro sum.cong ennreal_mult_left_cong refl)
  fix c assume "c ∈ ?au ` space M" "c ≠ 0"
  { fix x' assume x': "x' ∈ space M" "?au x' = c"
  with ‹c ≠ 0› u_range have "?au x' < 1 * u x'"
  by (intro ennreal_mult_strict_right_mono ‹a < 1›) (auto simp: less_le)
  also have "… ≤ (SUP i. f i x')"
  using u by (auto simp: le_fun_def)
  finally have "∃i. ?au x' ≤ f i x'"
  by (auto simp: less_SUP_iff intro: less_imp_le) }
  then have *: "?au -` {c} ∩ space M = (∪i. ?B c i)"
  by auto
  show "emeasure M (?au -` {c} ∩ space M) = (SUP i. emeasure M (?B c i))"
  unfolding * using f
  by (intro SUP_emeasure_incseq[symmetric])
  (auto simp: incseq_def le_fun_def intro: order_trans)
  qed
  also have "… = (SUP i. ∑c∈?au`space M. c * emeasure M (?B c i))"
  unfolding SUP_mult_left_ennreal using f
  by (intro ennreal_SUP_sum[symmetric])
  (auto intro!: mult_mono emeasure_mono simp: incseq_def le_fun_def intro: order_trans)
  also have "… ≤ (SUP i. integral🪙N M (f i))"
  proof (intro SUP_subset_mono order_refl)
  fix i
  have "(∑c∈?au`space M. c * emeasure M (?B c i)) =
  (∫🪙Sx. (a * u x) * indicator {x∈space M. a * u x ≤ f i x} x ∂M)"
  by (subst simple_integral_indicator)
  (auto intro!: sum.cong ennreal_mult_left_cong arg_cong2[where f=emeasure])
  also have "… = (∫🪙+x. (a * u x) * indicator {x∈space M. a * u x ≤ f i x} x ∂M)"
  by (rule nn_integral_eq_simple_integral[symmetric]) simp
  also have "… ≤ (∫🪙+x. f i x ∂M)"
  by (intro nn_integral_mono) (auto split: split_indicator)
  finally show "(∑c∈?au`space M. c * emeasure M (?B c i)) ≤ (∫🪙+x. f i x ∂M)" .
  qed
  finally show "a * integral🪙S M u ≤ (SUP i. integral🪙N M (f i))"
  by simp
  qed
  qed
 qed (auto intro!: SUP_least SUP_upper nn_integral_mono)
 
 lemma sup_continuous_nn_integral[order_continuous_intros]:
  assumes f: "∧y. sup_continuous (f y)"
  assumes [measurable]: "∧x. (λy. f y x) ∈ borel_measurable M"
  shows "sup_continuous (λx. (∫🪙+y. f y x ∂M))"
  unfolding sup_continuous_def
 proof safe
  fix C :: "nat ==> 'b" assume C: "incseq C"
  with sup_continuous_mono[OF f] show "(∫🪙+ y. f y (Sup (C ` UNIV)) ∂M) = (SUP i. ∫🪙+ y. f y (C i) ∂M)"
  unfolding sup_continuousD[OF f C]
  by (subst nn_integral_monotone_convergence_SUP) (auto simp: mono_def le_fun_def)
 qed
 
 theorem nn_integral_monotone_convergence_SUP_AE:
  assumes f: "∧i. AE x in M. f i x ≤ f (Suc i) x" "∧i. f i ∈ borel_measurable M"
  shows "(∫🪙+ x. (SUP i. f i x) ∂M) = (SUP i. integral🪙N M (f i))"
 proof -
  from f have "AE x in M. ∀i. f i x ≤ f (Suc i) x"
  by (simp add: AE_all_countable)
  from this[THEN AE_E] obtain N
  where N: "{x ∈ space M. ¬ (∀i. f i x ≤ f (Suc i) x)} ⊆ N" "emeasure M N = 0" "N ∈ sets M"
  by auto
  let ?f = "λi x. if x ∈ space M - N then f i x else 0"
  have f_eq: "AE x in M. ∀i. ?f i x = f i x" using N by (auto intro!: AE_I[of _ _ N])
  then have "(∫🪙+ x. (SUP i. f i x) ∂M) = (∫🪙+ x. (SUP i. ?f i x) ∂M)"
  by (auto intro!: nn_integral_cong_AE)
  also have "… = (SUP i. (∫🪙+ x. ?f i x ∂M))"
  proof (rule nn_integral_monotone_convergence_SUP)
  show "incseq ?f" using N(1) by (force intro!: incseq_SucI le_funI)
  { fix i show "(λx. if x ∈ space M - N then f i x else 0) ∈ borel_measurable M"
  using f N(3) by (intro measurable_If_set) auto }
  qed
  also have "… = (SUP i. (∫🪙+ x. f i x ∂M))"
  using f_eq by (force intro!: arg_cong[where f = "λf. Sup (range f)"] nn_integral_cong_AE ext)
  finally show ?thesis .
 qed
 
 lemma nn_integral_monotone_convergence_simple:
  "incseq f ==> (∧i. simple_function M (f i)) ==> (SUP i. ∫🪙S x. f i x ∂M) = (∫🪙+x. (SUP i. f i x) ∂M)"
  using nn_integral_monotone_convergence_SUP[of f M]
  by (simp add: nn_integral_eq_simple_integral[symmetric] borel_measurable_simple_function)
 
 lemma SUP_simple_integral_sequences:
  assumes f: "incseq f" "∧i. simple_function M (f i)"
  and g: "incseq g" "∧i. simple_function M (g i)"
  and eq: "AE x in M. (SUP i. f i x) = (SUP i. g i x)"
  shows "(SUP i. integral🪙S M (f i)) = (SUP i. integral🪙S M (g i))"
  (is "Sup (?F ` _) = Sup (?G ` _)")
 proof -
  have "(SUP i. integral🪙S M (f i)) = (∫🪙+x. (SUP i. f i x) ∂M)"
  using f by (rule nn_integral_monotone_convergence_simple)
  also have "… = (∫🪙+x. (SUP i. g i x) ∂M)"
  unfolding eq[THEN nn_integral_cong_AE] ..
  also have "… = (SUP i. ?G i)"
  using g by (rule nn_integral_monotone_convergence_simple[symmetric])
  finally show ?thesis by simp
 qed
 
 lemma nn_integral_const[simp]: "(∫🪙+ x. c ∂M) = c * emeasure M (space M)"
  by (subst nn_integral_eq_simple_integral) auto
 
 lemma nn_integral_linear:
  assumes f: "f ∈ borel_measurable M" and g: "g ∈ borel_measurable M"
  shows "(∫🪙+ x. a * f x + g x ∂M) = a * integral🪙N M f + integral🪙N M g"
  (is "integral🪙N M ?L = _")
 proof -
  obtain u
  where "∧i. simple_function M (u i)" "incseq u" "∧i x. u i x < top" "∧x. (SUP i. u i x) = f x"
  using borel_measurable_implies_simple_function_sequence' f(1)
  by auto
  note u = nn_integral_monotone_convergence_simple[OF this(2,1)] this
 
  obtain v where
  "∧i. simple_function M (v i)" "incseq v" "∧i x. v i x < top" "∧x. (SUP i. v i x) = g x"
  using borel_measurable_implies_simple_function_sequence' g(1)
  by auto
  note v = nn_integral_monotone_convergence_simple[OF this(2,1)] this
 
  let ?L' = "λi x. a * u i x + v i x"
 
  have "?L ∈ borel_measurable M" using assms by auto
  from borel_measurable_implies_simple_function_sequence'[OF this]
  obtain l where "∧i. simple_function M (l i)" "incseq l" "∧i x. l i x < top" "∧x. (SUP i. l i x) = a * f x + g x"
  by auto
  note l = nn_integral_monotone_convergence_simple[OF this(2,1)] this
 
  have inc: "incseq (λi. a * integral🪙S M (u i))" "incseq (λi. integral🪙S M (v i))"
  using u v by (auto simp: incseq_Suc_iff le_fun_def intro!: add_mono mult_left_mono simple_integral_mono)
 
  have l': "(SUP i. integral🪙S M (l i)) = (SUP i. integral🪙S M (?L' i))"
  proof (rule SUP_simple_integral_sequences[OF l(3,2)])
  show "incseq ?L'" "∧i. simple_function M (?L' i)"
  using u v unfolding incseq_Suc_iff le_fun_def
  by (auto intro!: add_mono mult_left_mono)
  { fix x
  have "(SUP i. a * u i x + v i x) = a * (SUP i. u i x) + (SUP i. v i x)"
  using u(3) v(3) u(4)[of _ x] v(4)[of _ x] unfolding SUP_mult_left_ennreal
  by (auto intro!: ennreal_SUP_add simp: incseq_Suc_iff le_fun_def add_mono mult_left_mono) }
  then show "AE x in M. (SUP i. l i x) = (SUP i. ?L' i x)"
  unfolding l(5) using u(5) v(5) by (intro AE_I2) auto
  qed
  also have "… = (SUP i. a * integral🪙S M (u i) + integral🪙S M (v i))"
  using u(2) v(2) by auto
  finally show ?thesis
  unfolding l(5)[symmetric] l(1)[symmetric]
  by (simp add: ennreal_SUP_add[OF inc] v u SUP_mult_left_ennreal[symmetric])
 qed
 
 lemma nn_integral_cmult: "f ∈ borel_measurable M ==> (∫🪙+ x. c * f x ∂M) = c * integral🪙N M f"
  using nn_integral_linear[of f M "λx. 0" c] by simp
 
 lemma nn_integral_multc: "f ∈ borel_measurable M ==> (∫🪙+ x. f x * c ∂M) = integral🪙N M f * c"
  unfolding mult.commute[of _ c] nn_integral_cmult by simp
 
 lemma nn_integral_divide: "f ∈ borel_measurable M ==> (∫🪙+ x. f x / c ∂M) = (∫🪙+ x. f x ∂M) / c"
  unfolding divide_ennreal_def by (rule nn_integral_multc)
 
 lemma nn_integral_indicator[simp]: "A ∈ sets M ==> (∫🪙+ x. indicator A x∂M) = (emeasure M) A"
  by (subst nn_integral_eq_simple_integral) (auto simp: simple_integral_indicator)
 
 lemma nn_integral_cmult_indicator: "A ∈ sets M ==> (∫🪙+ x. c * indicator A x ∂M) = c * emeasure M A"
  by (subst nn_integral_eq_simple_integral) (auto)
 
 lemma nn_integral_indicator':
  assumes [measurable]: "A ∩ space M ∈ sets M"
  shows "(∫🪙+ x. indicator A x ∂M) = emeasure M (A ∩ space M)"
 proof -
  have "(∫🪙+ x. indicator A x ∂M) = (∫🪙+ x. indicator (A ∩ space M) x ∂M)"
  by (intro nn_integral_cong) (simp split: split_indicator)
  also have "… = emeasure M (A ∩ space M)"
  by simp
  finally show ?thesis .
 qed
 
 lemma nn_integral_indicator_singleton[simp]:
  assumes [measurable]: "{y} ∈ sets M" shows "(∫🪙+x. f x * indicator {y} x ∂M) = f y * emeasure M {y}"
 proof -
  have "(∫🪙+x. f x * indicator {y} x ∂M) = (∫🪙+x. f y * indicator {y} x ∂M)"
  by (auto intro!: nn_integral_cong split: split_indicator)
  then show ?thesis
  by (simp add: nn_integral_cmult)
 qed
 
 lemma nn_integral_set_ennreal:
  "(∫🪙+x. ennreal (f x) * indicator A x ∂M) = (∫🪙+x. ennreal (f x * indicator A x) ∂M)"
  by (rule nn_integral_cong) (simp split: split_indicator)
 
 lemma nn_integral_indicator_singleton'[simp]:
  assumes [measurable]: "{y} ∈ sets M"
  shows "(∫🪙+x. ennreal (f x * indicator {y} x) ∂M) = f y * emeasure M {y}"
  by (subst nn_integral_set_ennreal[symmetric]) (simp)
 
 lemma nn_integral_add:
  "f ∈ borel_measurable M ==> g ∈ borel_measurable M ==> (∫🪙+ x. f x + g x ∂M) = integral🪙N M f + integral🪙N M g"
  using nn_integral_linear[of f M g 1] by simp
 
 lemma nn_integral_sum:
  "(∧i. i ∈ P ==> f i ∈ borel_measurable M) ==> (∫🪙+ x. (∑i∈P. f i x) ∂M) = (∑i∈P. integral🪙N M (f i))"
  by (induction P rule: infinite_finite_induct) (auto simp: nn_integral_add)
 
 theorem nn_integral_suminf:
  assumes f: "∧i. f i ∈ borel_measurable M"
  shows "(∫🪙+ x. (∑i. f i x) ∂M) = (∑i. integral🪙N M (f i))"
 proof -
  have all_pos: "AE x in M. ∀i. 0 ≤ f i x"
  using assms by (auto simp: AE_all_countable)
  have "(∑i. integral🪙N M (f i)) = (SUP n. ∑i🪙N M (f i))"
  by (rule suminf_eq_SUP)
  also have "… = (SUP n. ∫🪙+x. (∑i∂M)"
  unfolding nn_integral_sum[OF f] ..
  also have "… = ∫🪙+x. (SUP n. ∑i∂M" using f all_pos
  by (intro nn_integral_monotone_convergence_SUP_AE[symmetric])
  (elim AE_mp, auto simp: sum_nonneg simp del: sum.lessThan_Suc intro!: AE_I2 sum_mono2)
  also have "… = ∫🪙+x. (∑i. f i x) ∂M" using all_pos
  by (intro nn_integral_cong_AE) (auto simp: suminf_eq_SUP)
  finally show ?thesis by simp
 qed
 
 lemma nn_integral_bound_simple_function:
  assumes bnd: "∧x. x ∈ space M ==> f x < ∞"
  assumes f[measurable]: "simple_function M f"
  assumes supp: "emeasure M {x∈space M. f x ≠ 0} < ∞"
  shows "nn_integral M f < ∞"
 proof cases
  assume "space M = {}"
  then have "nn_integral M f = (∫🪙+x. 0 ∂M)"
  by (intro nn_integral_cong) auto
  then show ?thesis by simp
 next
  assume "space M ≠ {}"
  with simple_functionD(1)[OF f] bnd have bnd: "0 ≤ Max (f`space M) ∧ Max (f`space M) < ∞"
  by (subst Max_less_iff) (auto simp: Max_ge_iff)
 
  have "nn_integral M f ≤ (∫🪙+x. Max (f`space M) * indicator {x∈space M. f x ≠ 0} x ∂M)"
  proof (rule nn_integral_mono)
  fix x assume "x ∈ space M"
  with f show "f x ≤ Max (f ` space M) * indicator {x ∈ space M. f x ≠ 0} x"
  by (auto split: split_indicator intro!: Max_ge simple_functionD)
  qed
  also have "… < ∞"
  using bnd supp by (subst nn_integral_cmult) (auto simp: ennreal_mult_less_top)
  finally show ?thesis .
 qed
 
 theorem nn_integral_Markov_inequality:
  assumes u: "(λx. u x * indicator A x) ∈ borel_measurable M" and "A ∈ sets M"
  shows "(emeasure M) ({x∈A. 1 ≤ c * u x}) ≤ c * (∫🪙+ x. u x * indicator A x ∂M)"
  (is "(emeasure M) ?A ≤ _ * ?PI")
 proof -
  define u' where "u' = (λx. u x * indicator A x)"
  have [measurable]: "u' ∈ borel_measurable M"
  using u unfolding u'_def .
  have "{x∈space M. c * u' x ≥ 1} ∈ sets M"
  by measurable
  also have "{x∈space M. c * u' x ≥ 1} = ?A"
  using sets.sets_into_space[OF ‹A ∈ sets M›] by (auto simp: u'_def indicator_def)
  finally have "(emeasure M) ?A = (∫🪙+ x. indicator ?A x ∂M)"
  using nn_integral_indicator by simp
  also have "… ≤ (∫🪙+ x. c * (u x * indicator A x) ∂M)"
  using u by (auto intro!: nn_integral_mono_AE simp: indicator_def)
  also have "… = c * (∫🪙+ x. u x * indicator A x ∂M)"
  using assms by (auto intro!: nn_integral_cmult)
  finally show ?thesis .
 qed
 
 lemma Chernoff_ineq_nn_integral_ge:
  assumes s: "s > 0" and [measurable]: "A ∈ sets M"
  assumes [measurable]: "(λx. f x * indicator A x) ∈ borel_measurable M"
  shows "emeasure M {x∈A. f x ≥ a} ≤
  ennreal (exp (-s * a)) * nn_integral M (λx. ennreal (exp (s * f x)) * indicator A x)"
 proof -
  define f' where "f' = (λx. f x * indicator A x)"
  have [measurable]: "f' ∈ borel_measurable M"
  using assms(3) unfolding f'_def by assumption
  have "(λx. ennreal (exp (s * f' x)) * indicator A x) ∈ borel_measurable M"
  by simp
  also have "(λx. ennreal (exp (s * f' x)) * indicator A x) =
  (λx. ennreal (exp (s * f x)) * indicator A x)"
  by (auto simp: f'_def indicator_def fun_eq_iff)
  finally have meas: "… ∈ borel_measurable M" .
 
  have "{x∈A. f x ≥ a} = {x∈A. ennreal (exp (-s * a)) * ennreal (exp (s * f x)) ≥ 1}"
  using s by (auto simp: exp_minus field_simps simp flip: ennreal_mult)
  also have "emeasure M … ≤ ennreal (exp (-s * a)) *
  (∫🪙+x. ennreal (exp (s * f x)) * indicator A x ∂M)"
  by (intro order.trans[OF nn_integral_Markov_inequality] meas) auto
  finally show ?thesis .
 qed
 
 lemma Chernoff_ineq_nn_integral_le:
  assumes s: "s > 0" and [measurable]: "A ∈ sets M"
  assumes [measurable]: "f ∈ borel_measurable M"
  shows "emeasure M {x∈A. f x ≤ a} ≤
  ennreal (exp (s * a)) * nn_integral M (λx. ennreal (exp (-s * f x)) * indicator A x)"
  using Chernoff_ineq_nn_integral_ge[of s A M "λx. -f x" "-a"] assms by simp
 
 lemma nn_integral_noteq_infinite:
  assumes g: "g ∈ borel_measurable M" and "integral🪙N M g ≠ ∞"
  shows "AE x in M. g x ≠ ∞"
 proof (rule ccontr)
  assume c: "¬ (AE x in M. g x ≠ ∞)"
  have "(emeasure M) {x∈space M. g x = ∞} ≠ 0"
  using c g by (auto simp add: AE_iff_null)
  then have "0 < (emeasure M) {x∈space M. g x = ∞}"
  by (auto simp: zero_less_iff_neq_zero)
  then have "∞ = ∞ * (emeasure M) {x∈space M. g x = ∞}"
  by (auto simp: ennreal_top_eq_mult_iff)
  also have "… ≤ (∫🪙+x. ∞ * indicator {x∈space M. g x = ∞} x ∂M)"
  using g by (subst nn_integral_cmult_indicator) auto
  also have "… ≤ integral🪙N M g"
  using assms by (auto intro!: nn_integral_mono_AE simp: indicator_def)
  finally show False
  using ‹integral🪙N M g ≠ ∞› by (auto simp: top_unique)
 qed
 
 lemma nn_integral_PInf:
  assumes f: "f ∈ borel_measurable M" and not_Inf: "integral🪙N M f ≠ ∞"
  shows "emeasure M (f -` {∞} ∩ space M) = 0"
 proof -
  have "∞ * emeasure M (f -` {∞} ∩ space M) = (∫🪙+ x. ∞ * indicator (f -` {∞} ∩ space M) x ∂M)"
  using f by (subst nn_integral_cmult_indicator) (auto simp: measurable_sets)
  also have "… ≤ integral🪙N M f"
  by (auto intro!: nn_integral_mono simp: indicator_def)
  finally have "∞ * (emeasure M) (f -` {∞} ∩ space M) ≤ integral🪙N M f"
  by simp
  then show ?thesis
  using assms by (auto simp: ennreal_top_mult top_unique split: if_split_asm)
 qed
 
 lemma simple_integral_PInf:
  "simple_function M f ==> integral🪙S M f ≠ ∞ ==> emeasure M (f -` {∞} ∩ space M) = 0"
  by (rule nn_integral_PInf) (auto simp: nn_integral_eq_simple_integral borel_measurable_simple_function)
 
 lemma nn_integral_PInf_AE:
  assumes "f ∈ borel_measurable M" "integral🪙N M f ≠ ∞" shows "AE x in M. f x ≠ ∞"
 proof (rule AE_I)
  show "(emeasure M) (f -` {∞} ∩ space M) = 0"
  by (rule nn_integral_PInf[OF assms])
  show "f -` {∞} ∩ space M ∈ sets M"
  using assms by (auto intro: borel_measurable_vimage)
 qed auto
 
 lemma nn_integral_diff:
  assumes f: "f ∈ borel_measurable M"
  and g: "g ∈ borel_measurable M"
  and fin: "integral🪙N M g ≠ ∞"
  and mono: "AE x in M. g x ≤ f x"
  shows "(∫🪙+ x. f x - g x ∂M) = integral🪙N M f - integral🪙N M g"
 proof -
  have diff: "(λx. f x - g x) ∈ borel_measurable M"
  using assms by auto
  have "AE x in M. f x = f x - g x + g x"
  using diff_add_cancel_ennreal mono nn_integral_noteq_infinite[OF g fin] assms by auto
  then have **: "integral🪙N M f = (∫🪙+x. f x - g x ∂M) + integral🪙N M g"
  unfolding nn_integral_add[OF diff g, symmetric]
  by (rule nn_integral_cong_AE)
  show ?thesis unfolding **
  using fin
  by (cases rule: ennreal2_cases[of "∫🪙+ x. f x - g x ∂M" "integral🪙N M g"]) auto
 qed
 
 lemma nn_integral_mult_bounded_inf:
  assumes f: "f ∈ borel_measurable M" "(∫🪙+x. f x ∂M) < ∞" and c: "c ≠ ∞" and ae: "AE x in M. g x ≤ c * f x"
  shows "(∫🪙+x. g x ∂M) < ∞"
 proof -
  have "(∫🪙+x. g x ∂M) ≤ (∫🪙+x. c * f x ∂M)"
  by (intro nn_integral_mono_AE ae)
  also have "(∫🪙+x. c * f x ∂M) < ∞"
  using c f by (subst nn_integral_cmult) (auto simp: ennreal_mult_less_top top_unique not_less)
  finally show ?thesis .
 qed
 
 text ‹Fatou's lemma: convergence theorem on limes inferior›
 
 lemma nn_integral_monotone_convergence_INF_AE':
  assumes f: "∧i. AE x in M. f (Suc i) x ≤ f i x" and [measurable]: "∧i. f i ∈ borel_measurable M"
  and *: "(∫🪙+ x. f 0 x ∂M) < ∞"
  shows "(∫🪙+ x. (INF i. f i x) ∂M) = (INF i. integral🪙N M (f i))"
 proof (rule ennreal_minus_cancel)
  have "integral🪙N M (f 0) - (∫🪙+ x. (INF i. f i x) ∂M) = (∫🪙+x. f 0 x - (INF i. f i x) ∂M)"
  proof (rule nn_integral_diff[symmetric])
  have "(∫🪙+ x. (INF i. f i x) ∂M) ≤ (∫🪙+ x. f 0 x ∂M)"
  by (intro nn_integral_mono INF_lower) simp
  with * show "(∫🪙+ x. (INF i. f i x) ∂M) ≠ ∞"
  by simp
  qed (auto intro: INF_lower)
  also have "… = (∫🪙+x. (SUP i. f 0 x - f i x) ∂M)"
  by (simp add: ennreal_INF_const_minus)
  also have "… = (SUP i. (∫🪙+x. f 0 x - f i x ∂M))"
  proof (intro nn_integral_monotone_convergence_SUP_AE)
  show "AE x in M. f 0 x - f i x ≤ f 0 x - f (Suc i) x" for i
  using f[of i] by eventually_elim (auto simp: ennreal_mono_minus)
  qed simp
  also have "… = (SUP i. nn_integral M (f 0) - (∫🪙+x. f i x ∂M))"
  proof (subst nn_integral_diff[symmetric])
  fix i
  have dec: "AE x in M. ∀i. f (Suc i) x ≤ f i x"
  unfolding AE_all_countable using f by auto
  then show "AE x in M. f i x ≤ f 0 x"
  using dec by eventually_elim (auto intro: lift_Suc_antimono_le[of "λi. f i x" 0 i for x])
  then have "(∫🪙+ x. f i x ∂M) ≤ (∫🪙+ x. f 0 x ∂M)"
  by (rule nn_integral_mono_AE)
  with * show "(∫🪙+ x. f i x ∂M) ≠ ∞"
  by simp
  qed (insert f, auto simp: decseq_def le_fun_def)
  finally show "integral🪙N M (f 0) - (∫🪙+ x. (INF i. f i x) ∂M) =
  integral🪙N M (f 0) - (INF i. ∫🪙+ x. f i x ∂M)"
  by (simp add: ennreal_INF_const_minus)
 qed (insert *, auto intro!: nn_integral_mono intro: INF_lower)
 
 theorem nn_integral_monotone_convergence_INF_AE:
  fixes f :: "nat ==> 'a ==> ennreal"
  assumes f: "∧i. AE x in M. f (Suc i) x ≤ f i x"
  and [measurable]: "∧i. f i ∈ borel_measurable M"
  and fin: "(∫🪙+ x. f i x ∂M) < ∞"
  shows "(∫🪙+ x. (INF i. f i x) ∂M) = (INF i. integral🪙N M (f i))"
 proof -
  { fix f :: "nat ==> ennreal" and j assume "decseq f"
  then have "(INF i. f i) = (INF i. f (i + j))"
  apply (intro INF_eq)
  apply (rule_tac x="i" in bexI)
  apply (auto simp: decseq_def le_fun_def)
  done }
  note INF_shift = this
  have mono: "AE x in M. ∀i. f (Suc i) x ≤ f i x"
  using f by (auto simp: AE_all_countable)
  then have "AE x in M. (INF i. f i x) = (INF n. f (n + i) x)"
  by eventually_elim (auto intro!: decseq_SucI INF_shift)
  then have "(∫🪙+ x. (INF i. f i x) ∂M) = (∫🪙+ x. (INF n. f (n + i) x) ∂M)"
  by (rule nn_integral_cong_AE)
  also have "… = (INF n. (∫🪙+ x. f (n + i) x ∂M))"
  by (rule nn_integral_monotone_convergence_INF_AE') (insert assms, auto)
  also have "… = (INF n. (∫🪙+ x. f n x ∂M))"
  by (intro INF_shift[symmetric] decseq_SucI nn_integral_mono_AE f)
  finally show ?thesis .
 qed
 
 lemma nn_integral_monotone_convergence_INF_decseq:
  assumes f: "decseq f" and *: "∧i. f i ∈ borel_measurable M" "(∫🪙+ x. f i x ∂M) < ∞"
  shows "(∫🪙+ x. (INF i. f i x) ∂M) = (INF i. integral🪙N M (f i))"
  using nn_integral_monotone_convergence_INF_AE[of f M i, OF _ *] f by (simp add: decseq_SucD le_funD)
 
 theorem nn_integral_liminf:
  fixes u :: "nat ==> 'a ==> ennreal"
  assumes u: "∧i. u i ∈ borel_measurable M"
  shows "(∫🪙+ x. liminf (λn. u n x) ∂M) ≤ liminf (λn. integral🪙N M (u n))"
 proof -
  have "(∫🪙+ x. liminf (λn. u n x) ∂M) = (SUP n. ∫🪙+ x. (INF i∈{n..}. u i x) ∂M)"
  unfolding liminf_SUP_INF using u
  by (intro nn_integral_monotone_convergence_SUP_AE)
  (auto intro!: AE_I2 intro: INF_greatest INF_superset_mono)
  also have "… ≤ liminf (λn. integral🪙N M (u n))"
  by (auto simp: liminf_SUP_INF intro!: SUP_mono INF_greatest nn_integral_mono INF_lower)
  finally show ?thesis .
 qed
 
 theorem nn_integral_limsup:
  fixes u :: "nat ==> 'a ==> ennreal"
  assumes [measurable]: "∧i. u i ∈ borel_measurable M" "w ∈ borel_measurable M"
  assumes bounds: "∧i. AE x in M. u i x ≤ w x" and w: "(∫🪙+x. w x ∂M) < ∞"
  shows "limsup (λn. integral🪙N M (u n)) ≤ (∫🪙+ x. limsup (λn. u n x) ∂M)"
 proof -
  have bnd: "AE x in M. ∀i. u i x ≤ w x"
  using bounds by (auto simp: AE_all_countable)
  then have "(∫🪙+ x. (SUP n. u n x) ∂M) ≤ (∫🪙+ x. w x ∂M)"
  by (auto intro!: nn_integral_mono_AE elim: eventually_mono intro: SUP_least)
  then have "(∫🪙+ x. limsup (λn. u n x) ∂M) = (INF n. ∫🪙+ x. (SUP i∈{n..}. u i x) ∂M)"
  unfolding limsup_INF_SUP using bnd w
  by (intro nn_integral_monotone_convergence_INF_AE')
  (auto intro!: AE_I2 intro: SUP_least SUP_subset_mono)
  also have "… ≥ limsup (λn. integral🪙N M (u n))"
  by (auto simp: limsup_INF_SUP intro!: INF_mono SUP_least exI nn_integral_mono SUP_upper)
  finally (xtrans) show ?thesis .
 qed
 
 lemma nn_integral_LIMSEQ:
  assumes f: "incseq f" "∧i. f i ∈ borel_measurable M"
  and u: "∧x. (λi. f i x) <---- u x"
  shows "(λn. integral🪙N M (f n)) <---- integral🪙N M u"
 proof -
  have "(λn. integral🪙N M (f n)) <---- (SUP n. integral🪙N M (f n))"
  using f by (intro LIMSEQ_SUP[of "λn. integral🪙N M (f n)"] incseq_nn_integral)
  also have "(SUP n. integral🪙N M (f n)) = integral🪙N M (λx. SUP n. f n x)"
  using f by (intro nn_integral_monotone_convergence_SUP[symmetric])
  also have "integral🪙N M (λx. SUP n. f n x) = integral🪙N M (λx. u x)"
  using f by (subst LIMSEQ_SUP[THEN LIMSEQ_unique, OF _ u]) (auto simp: incseq_def le_fun_def)
  finally show ?thesis .
 qed
 
 theorem nn_integral_dominated_convergence:
  assumes [measurable]:
  "∧i. u i ∈ borel_measurable M" "u' ∈ borel_measurable M" "w ∈ borel_measurable M"
  and bound: "∧j. AE x in M. u j x ≤ w x"
  and w: "(∫🪙+x. w x ∂M) < ∞"
  and u': "AE x in M. (λi. u i x) <---- u' x"
  shows "(λi. (∫🪙+x. u i x ∂M)) <---- (∫🪙+x. u' x ∂M)"
 proof -
  have "limsup (λn. integral🪙N M (u n)) ≤ (∫🪙+ x. limsup (λn. u n x) ∂M)"
  by (intro nn_integral_limsup[OF _ _ bound w]) auto
  moreover have "(∫🪙+ x. limsup (λn. u n x) ∂M) = (∫🪙+ x. u' x ∂M)"
  using u' by (intro nn_integral_cong_AE, eventually_elim) (metis tendsto_iff_Liminf_eq_Limsup sequentially_bot)
  moreover have "(∫🪙+ x. liminf (λn. u n x) ∂M) = (∫🪙+ x. u' x ∂M)"
  using u' by (intro nn_integral_cong_AE, eventually_elim) (metis tendsto_iff_Liminf_eq_Limsup sequentially_bot)
  moreover have "(∫🪙+x. liminf (λn. u n x) ∂M) ≤ liminf (λn. integral🪙N M (u n))"
  by (intro nn_integral_liminf) auto
  moreover have "liminf (λn. integral🪙N M (u n)) ≤ limsup (λn. integral🪙N M (u n))"
  by (intro Liminf_le_Limsup sequentially_bot)
  ultimately show ?thesis
  by (intro Liminf_eq_Limsup) auto
 qed
 
 lemma inf_continuous_nn_integral[order_continuous_intros]:
  assumes f: "∧y. inf_continuous (f y)"
  assumes [measurable]: "∧x. (λy. f y x) ∈ borel_measurable M"
  assumes bnd: "∧x. (∫🪙+ y. f y x ∂M) ≠ ∞"
  shows "inf_continuous (λx. (∫🪙+y. f y x ∂M))"
  unfolding inf_continuous_def
 proof safe
  fix C :: "nat ==> 'b" assume C: "decseq C"
  then show "(∫🪙+ y. f y (Inf (C ` UNIV)) ∂M) = (INF i. ∫🪙+ y. f y (C i) ∂M)"
  using inf_continuous_mono[OF f] bnd
  by (auto simp add: inf_continuousD[OF f C] fun_eq_iff monotone_def le_fun_def less_top
  intro!: nn_integral_monotone_convergence_INF_decseq)
 qed
 
 lemma nn_integral_null_set:
  assumes "N ∈ null_sets M" shows "(∫🪙+ x. u x * indicator N x ∂M) = 0"
 proof -
  have "(∫🪙+ x. u x * indicator N x ∂M) = (∫🪙+ x. 0 ∂M)"
  proof (intro nn_integral_cong_AE AE_I)
  show "{x ∈ space M. u x * indicator N x ≠ 0} ⊆ N"
  by (auto simp: indicator_def)
  show "(emeasure M) N = 0" "N ∈ sets M"
  using assms by auto
  qed
  then show ?thesis by simp
 qed
 
 lemma nn_integral_0_iff:
  assumes u [measurable]: "u ∈ borel_measurable M"
  shows "integral🪙N M u = 0 ⟷ emeasure M {x∈space M. u x ≠ 0} = 0"
  (is "_ ⟷ (emeasure M) ?A = 0")
 proof -
  have u_eq: "(∫🪙+ x. u x * indicator ?A x ∂M) = integral🪙N M u"
  by (auto intro!: nn_integral_cong simp: indicator_def)
  show ?thesis
  proof
  assume "(emeasure M) ?A = 0"
  with nn_integral_null_set[of ?A M u] u
  show "integral🪙N M u = 0" by (simp add: u_eq null_sets_def)
  next
  assume *: "integral🪙N M u = 0"
  let ?M = "λn. {x ∈ space M. 1 ≤ real (n::nat) * u x}"
  have "0 = (SUP n. (emeasure M) (?M n ∩ ?A))"
  proof -
  { fix n :: nat
  have "emeasure M {x ∈ ?A. 1 ≤ of_nat n * u x} ≤
  of_nat n * ∫🪙+ x. u x * indicator ?A x ∂M"
  by (intro nn_integral_Markov_inequality) auto
  also have "{x ∈ ?A. 1 ≤ of_nat n * u x} = (?M n ∩ ?A)"
  by (auto simp: ennreal_of_nat_eq_real_of_nat u_eq * )
  finally have "emeasure M (?M n ∩ ?A) ≤ 0"
  by (simp add: ennreal_of_nat_eq_real_of_nat u_eq * )
  moreover have "0 ≤ (emeasure M) (?M n ∩ ?A)" using u by auto
  ultimately have "(emeasure M) (?M n ∩ ?A) = 0" by auto }
  thus ?thesis by simp
  qed
  also have "… = (emeasure M) (∪n. ?M n ∩ ?A)"
  proof (safe intro!: SUP_emeasure_incseq)
  fix n show "?M n ∩ ?A ∈ sets M"
  using u by (auto intro!: sets.Int)
  next
  show "incseq (λn. {x ∈ space M. 1 ≤ real n * u x} ∩ {x ∈ space M. u x ≠ 0})"
  proof (safe intro!: incseq_SucI)
  fix n :: nat and x
  assume *: "1 ≤ real n * u x"
  also have "real n * u x ≤ real (Suc n) * u x"
  by (auto intro!: mult_right_mono)
  finally show "1 ≤ real (Suc n) * u x" by auto
  qed
  qed
  also have "… = (emeasure M) {x∈space M. 0 < u x}"
  proof (safe intro!: arg_cong[where f="(emeasure M)"])
  fix x assume "0 < u x" and [simp, intro]: "x ∈ space M"
  show "x ∈ (∪n. ?M n ∩ ?A)"
  proof (cases "u x" rule: ennreal_cases)
  case (real r) with ‹0 < u x› have "0 < r" by auto
  obtain j :: nat where "1 / r ≤ real j" using real_arch_simple ..
  hence "1 / r * r ≤ real j * r" unfolding mult_le_cancel_right using ‹0 < r› by auto
  hence "1 ≤ real j * r" using real ‹0 < r› by auto
  thus ?thesis using ‹0 < r› real
  by (auto simp: ennreal_of_nat_eq_real_of_nat ennreal_1[symmetric] ennreal_mult[symmetric]
  simp del: ennreal_1)
  qed (insert ‹0 < u x›, auto simp: ennreal_mult_top)
  qed (auto simp: zero_less_iff_neq_zero)
  finally show "emeasure M ?A = 0"
  by (simp add: zero_less_iff_neq_zero)
  qed
 qed
 
 lemma nn_integral_0_iff_AE:
  assumes u: "u ∈ borel_measurable M"
  shows "integral🪙N M u = 0 ⟷ (AE x in M. u x = 0)"
 proof -
  have sets: "{x∈space M. u x ≠ 0} ∈ sets M"
  using u by auto
  show "integral🪙N M u = 0 ⟷ (AE x in M. u x = 0)"
  using nn_integral_0_iff[of u] AE_iff_null[OF sets] u by auto
 qed
 
 lemma AE_iff_nn_integral:
  "{x∈space M. P x} ∈ sets M ==> (AE x in M. P x) ⟷ integral🪙N M (indicator {x. ¬ P x}) = 0"
  by (subst nn_integral_0_iff_AE) (auto simp: indicator_def[abs_def])
 
 lemma nn_integral_less:
  assumes [measurable]: "f ∈ borel_measurable M" "g ∈ borel_measurable M"
  assumes f: "(∫🪙+x. f x ∂M) ≠ ∞"
  assumes ord: "AE x in M. f x ≤ g x" "¬ (AE x in M. g x ≤ f x)"
  shows "(∫🪙+x. f x ∂M) < (∫🪙+x. g x ∂M)"
 proof -
  have "0 < (∫🪙+x. g x - f x ∂M)"
  proof (intro order_le_neq_trans notI)
  assume "0 = (∫🪙+x. g x - f x ∂M)"
  then have "AE x in M. g x - f x = 0"
  using nn_integral_0_iff_AE[of "λx. g x - f x" M] by simp
  with ord(1) have "AE x in M. g x ≤ f x"
  by eventually_elim (auto simp: ennreal_minus_eq_0)
  with ord show False
  by simp
  qed simp
  also have "… = (∫🪙+x. g x ∂M) - (∫🪙+x. f x ∂M)"
  using f by (subst nn_integral_diff) (auto simp: ord)
  finally show ?thesis
  using f by (auto dest!: ennreal_minus_pos_iff[rotated] simp: less_top)
 qed
 
 lemma nn_integral_subalgebra:
  assumes f: "f ∈ borel_measurable N"
  and N: "sets N ⊆ sets M" "space N = space M" "∧A. A ∈ sets N ==> emeasure N A = emeasure M A"
  shows "integral🪙N N f = integral🪙N M f"
 proof -
  have [simp]: "∧f :: 'a ==> ennreal. f ∈ borel_measurable N ==> f ∈ borel_measurable M"
  using N by (auto simp: measurable_def)
  have [simp]: "∧P. (AE x in N. P x) ==> (AE x in M. P x)"
  using N by (auto simp add: eventually_ae_filter null_sets_def subset_eq)
  have [simp]: "∧A. A ∈ sets N ==> A ∈ sets M"
  using N by auto
  from f show ?thesis
  apply induct
  apply (simp_all add: nn_integral_add nn_integral_cmult nn_integral_monotone_convergence_SUP N image_comp)
  apply (auto intro!: nn_integral_cong cong: nn_integral_cong simp: N(2)[symmetric])
  done
 qed
 
 lemma nn_integral_nat_function:
  fixes f :: "'a ==> nat"
  assumes "f ∈ measurable M (count_space UNIV)"
  shows "(∫🪙+x. of_nat (f x) ∂M) = (∑t. emeasure M {x∈space M. t < f x})"
 proof -
  define F where "F i = {x∈space M. i < f x}" for i
  with assms have [measurable]: "∧i. F i ∈ sets M"
  by auto
 
  { fix x assume "x ∈ space M"
  have "(λi. if i < f x then 1 else 0) sums (of_nat (f x)::real)"
  using sums_If_finite[of "λi. i < f x" "λ_. 1::real"] by simp
  then have "(λi. ennreal (if i < f x then 1 else 0)) sums of_nat(f x)"
  unfolding ennreal_of_nat_eq_real_of_nat
  by (subst sums_ennreal) auto
  moreover have "∧i. ennreal (if i < f x then 1 else 0) = indicator (F i) x"
  using ‹x ∈ space M› by (simp add: one_ennreal_def F_def)
  ultimately have "of_nat (f x) = (∑i. indicator (F i) x :: ennreal)"
  by (simp add: sums_iff) }
  then have "(∫🪙+x. of_nat (f x) ∂M) = (∫🪙+x. (∑i. indicator (F i) x) ∂M)"
  by (simp cong: nn_integral_cong)
  also have "… = (∑i. emeasure M (F i))"
  by (simp add: nn_integral_suminf)
  finally show ?thesis
  by (simp add: F_def)
 qed
 
 theorem nn_integral_lfp:
  assumes sets[simp]: "∧s. sets (M s) = sets N"
  assumes f: "sup_continuous f"
  assumes g: "sup_continuous g"
  assumes meas: "∧F. F ∈ borel_measurable N ==> f F ∈ borel_measurable N"
  assumes step: "∧F s. F ∈ borel_measurable N ==> integral🪙N (M s) (f F) = g (λs. integral🪙N (M s) F) s"
  shows "(∫🪙+ψ. lfp f ψ ∂M s) = lfp g s"
 proof (subst lfp_transfer_bounded[where α="λF s. ∫🪙+x. F x ∂M s" and g=g and f=f and P="λf. f ∈ borel_measurable N", symmetric])
  fix C :: "nat ==> 'b ==> ennreal" assume "incseq C" "∧i. C i ∈ borel_measurable N"
  then show "(λs. ∫🪙+x. (SUP i. C i) x ∂M s) = (SUP i. (λs. ∫🪙+x. C i x ∂M s))"
  unfolding SUP_apply[abs_def]
  by (subst nn_integral_monotone_convergence_SUP)
  (auto simp: mono_def fun_eq_iff intro!: arg_cong2[where f=emeasure] cong: measurable_cong_sets)
 qed (auto simp add: step le_fun_def SUP_apply[abs_def] bot_fun_def bot_ennreal intro!: meas f g)
 
 theorem nn_integral_gfp:
  assumes sets[simp]: "∧s. sets (M s) = sets N"
  assumes f: "inf_continuous f" and g: "inf_continuous g"
  assumes meas: "∧F. F ∈ borel_measurable N ==> f F ∈ borel_measurable N"
  assumes bound: "∧F s. F ∈ borel_measurable N ==> (∫🪙+x. f F x ∂M s) < ∞"
  assumes non_zero: "∧s. emeasure (M s) (space (M s)) ≠ 0"
  assumes step: "∧F s. F ∈ borel_measurable N ==> integral🪙N (M s) (f F) = g (λs. integral🪙N (M s) F) s"
  shows "(∫🪙+ψ. gfp f ψ ∂M s) = gfp g s"
 proof (subst gfp_transfer_bounded[where α="λF s. ∫🪙+x. F x ∂M s" and g=g and f=f
  and P="λF. F ∈ borel_measurable N ∧ (∀s. (∫🪙+x. F x ∂M s) < ∞)", symmetric])
  fix C :: "nat ==> 'b ==> ennreal" assume "decseq C" "∧i. C i ∈ borel_measurable N ∧ (∀s. integral🪙N (M s) (C i) < ∞)"
  then show "(λs. ∫🪙+x. (INF i. C i) x ∂M s) = (INF i. (λs. ∫🪙+x. C i x ∂M s))"
  unfolding INF_apply[abs_def]
  by (subst nn_integral_monotone_convergence_INF_decseq)
  (auto simp: mono_def fun_eq_iff intro!: arg_cong2[where f=emeasure] cong: measurable_cong_sets)
 next
  show "∧x. g x ≤ (λs. integral🪙N (M s) (f top))"
  by (subst step)
  (auto simp add: top_fun_def less_le non_zero le_fun_def ennreal_top_mult
  cong del: if_weak_cong intro!: monoD[OF inf_continuous_mono[OF g], THEN le_funD])
 next
  fix C assume "∧i::nat. C i ∈ borel_measurable N ∧ (∀s. integral🪙N (M s) (C i) < ∞)" "decseq C"
  with bound show "Inf (C ` UNIV) ∈ borel_measurable N ∧ (∀s. integral🪙N (M s) (Inf (C ` UNIV)) < ∞)"
  unfolding INF_apply[abs_def]
  by (subst nn_integral_monotone_convergence_INF_decseq)
  (auto simp: INF_less_iff cong: measurable_cong_sets intro!: borel_measurable_INF)
 next
  show "∧x. x ∈ borel_measurable N ∧ (∀s. integral🪙N (M s) x < ∞) ==>
  (λs. integral🪙N (M s) (f x)) = g (λs. integral🪙N (M s) x)"
  by (subst step) auto
 qed (insert bound, auto simp add: le_fun_def INF_apply[abs_def] top_fun_def intro!: meas f g)
 
 
 text ‹Cauchy--Schwarz inequality for 🍋‹nn_integral›\
 
 lemma sum_of_squares_ge_ennreal:
  fixes a b :: ennreal
  shows "2 * a * b ≤ a🪙2 + b🪙2"
 proof (cases a; cases b)
  fix x y
  assume xy: "x ≥ 0" "y ≥ 0" and [simp]: "a = ennreal x" "b = ennreal y"
  have "0 ≤ (x - y)🪙2"
  by simp
  also have "… = x🪙2 + y🪙2 - 2 * x * y"
  by (simp add: algebra_simps power2_eq_square)
  finally have "2 * x * y ≤ x🪙2 + y🪙2"
  by simp
  hence "ennreal (2 * x * y) ≤ ennreal (x🪙2 + y🪙2)"
  by (intro ennreal_leI)
  thus ?thesis using xy
  by (simp add: ennreal_mult ennreal_power)
 qed auto
 
 lemma Cauchy_Schwarz_nn_integral:
  assumes [measurable]: "f ∈ borel_measurable M" "g ∈ borel_measurable M"
  shows "(∫🪙+x. f x * g x ∂M)🪙2 ≤ (∫🪙+x. f x ^ 2 ∂M) * (∫🪙+x. g x ^ 2 ∂M)"
 proof (cases "(∫🪙+x. f x * g x ∂M) = 0")
  case False
  define F where "F = nn_integral M (λx. f x ^ 2)"
  define G where "G = nn_integral M (λx. g x ^ 2)"
  from False have "¬(AE x in M. f x = 0 ∨ g x = 0)"
  by (auto simp: nn_integral_0_iff_AE)
  hence "¬(AE x in M. f x = 0)" and "¬(AE x in M. g x = 0)"
  by (auto intro: AE_disjI1 AE_disjI2)
  hence nz: "F ≠ 0" "G ≠ 0"
  by (auto simp: nn_integral_0_iff_AE F_def G_def)
 
  show ?thesis
  proof (cases "F = ∞ ∨ G = ∞")
  case True
  thus ?thesis using nz
  by (auto simp: F_def G_def)
  next
  case False
  define F' where "F' = ennreal (sqrt (enn2real F))"
  define G' where "G' = ennreal (sqrt (enn2real G))"
  from False have fin: "F < top" "G < top"
  by (simp_all add: top.not_eq_extremum)
  have F'_sqr: "F'🪙2 = F"
  using False by (cases F) (auto simp: F'_def ennreal_power)
  have G'_sqr: "G'🪙2 = G"
  using False by (cases G) (auto simp: G'_def ennreal_power)
  have nz': "F' ≠ 0" "G' ≠ 0" and fin': "F' ≠ ∞" "G' ≠ ∞"
  using F'_sqr G'_sqr nz fin by auto
  from fin' have fin'': "F' < top" "G' < top"
  by (auto simp: top.not_eq_extremum)
 
  have "2 * (F' / F') * (G' / G') * (∫🪙+x. f x * g x ∂M) =
  F' * G' * (∫🪙+x. 2 * (f x / F') * (g x / G') ∂M)"
  using nz' fin''
  by (simp add: divide_ennreal_def algebra_simps ennreal_inverse_mult flip: nn_integral_cmult)
  also have "F'/ F' = 1"
  using nz' fin'' by simp
  also have "G'/ G' = 1"
  using nz' fin'' by simp
  also have "2 * 1 * 1 = (2 :: ennreal)" by simp
  also have "F' * G' * (∫🪙+ x. 2 * (f x / F') * (g x / G') ∂M) ≤
  F' * G' * (∫🪙+x. (f x / F')🪙2 + (g x / G')🪙2 ∂M)"
  by (intro mult_left_mono nn_integral_mono sum_of_squares_ge_ennreal) auto
  also have "… = F' * G' * (F / F'🪙2 + G / G'🪙2)" using nz
  by (auto simp: nn_integral_add algebra_simps nn_integral_divide F_def G_def)
  also have "F / F'🪙2 = 1"
  using nz F'_sqr fin by simp
  also have "G / G'🪙2 = 1"
  using nz G'_sqr fin by simp
  also have "F' * G' * (1 + 1) = 2 * (F' * G')"
  by (simp add: mult_ac)
  finally have "(∫🪙+x. f x * g x ∂M) ≤ F' * G'"
  by (subst (asm) ennreal_mult_le_mult_iff) auto
  hence "(∫🪙+x. f x * g x ∂M)🪙2 ≤ (F' * G')🪙2"
  by (intro power_mono_ennreal)
  also have "… = F * G"
  by (simp add: algebra_simps F'_sqr G'_sqr)
  finally show ?thesis
  by (simp add: F_def G_def)
  qed
 qed auto
 
 
(* TODO: rename? *)
subsection ‹Integral under concrete measures›

lemma nn_integral_mono_measure:
  assumes "sets M = sets N" "M ≤ N" shows "nn_integral M f ≤ nn_integral N f"
  unfolding nn_integral_def
proof (intro SUP_subset_mono)
  note ‹sets M = sets N›[simp] ‹sets M = sets N›[THEN sets_eq_imp_space_eq, simp]
  show "{g. simple_function M g ∧ g ≤ f} ⊆ {g. simple_function N g ∧ g ≤ f}"
    by (simp add: simple_function_def)
  show "integral🪙S M x ≤ integral🪙S N x" for x
    using le_measureD3[OF ‹M ≤ N›]
    by (auto simp add: simple_integral_def intro!: sum_mono mult_mono)
qed

lemma nn_integral_empty:
  assumes "space M = {}"
  shows "nn_integral M f = 0"
proof -
  have "(∫🪙+ x. f x ∂M) = (∫🪙+ x. 0 ∂M)"
    by(rule nn_integral_cong)(simp add: assms)
  thus ?thesis by simp
qed

lemma nn_integral_bot[simp]: "nn_integral bot f = 0"
  by (simp add: nn_integral_empty)

subsubsection🍋‹tag unimportant› ‹Distributions›

lemma nn_integral_distr:
  assumes T: "T ∈ measurable M M'" and f: "f ∈ borel_measurable (distr M M' T)"
  shows "integral🪙N (distr M M' T) f = (∫🪙+ x. f (T x) ∂M)"
  using f
proof induct
  case (cong f g)
  with T show ?case
    apply (subst nn_integral_cong[of _ f g])
    apply simp
    apply (subst nn_integral_cong[of _ "λx. f (T x)" "λx. g (T x)"])
    apply (simp add: measurable_def Pi_iff)
    apply simp
    done
next
  case (set A)
  then have eq: "∧x. x ∈ space M ==> indicator A (T x) = indicator (T -` A ∩ space M) x"
    by (auto simp: indicator_def)
  from set T show ?case
    by (subst nn_integral_cong[OF eq])
       (auto simp add: emeasure_distr intro!: nn_integral_indicator[symmetric] measurable_sets)
qed (simp_all add: measurable_compose[OF T] T nn_integral_cmult nn_integral_add
                   nn_integral_monotone_convergence_SUP le_fun_def incseq_def image_comp)

subsubsection🍋‹tag unimportant› ‹Counting space›

lemma simple_function_count_space[simp]:
  "simple_function (count_space A) f ⟷ finite (f ` A)"
  unfolding simple_function_def by simp

lemma nn_integral_count_space:
  assumes A: "finite {a∈A. 0 < f a}"
  shows "integral🪙N (count_space A) f = (∑a|a∈A ∧ 0 < f a. f a)"
proof -
  have *: "(∫🪙+x. max 0 (f x) ∂count_space A) =
    (∫🪙+ x. (∑a|a∈A ∧ 0 < f a. f a * indicator {a} x) ∂count_space A)"
    by (auto intro!: nn_integral_cong
             simp add: indicator_def of_bool_def if_distrib sum.If_cases[OF A] max_def le_less)
  also have "… = (∑a|a∈A ∧ 0 < f a. ∫🪙+ x. f a * indicator {a} x ∂count_space A)"
    by (subst nn_integral_sum) (simp_all add: AE_count_space less_imp_le)
  also have "… = (∑a|a∈A ∧ 0 < f a. f a)"
    by (auto intro!: sum.cong simp: one_ennreal_def[symmetric] max_def)
  finally show ?thesis by (simp add: max.absorb2)
qed

lemma nn_integral_count_space_finite:
    "finite A ==> (∫🪙+x. f x ∂count_space A) = (∑a∈A. f a)"
  by (auto intro!: sum.mono_neutral_left simp: nn_integral_count_space less_le)

lemma nn_integral_count_space':
  assumes "finite A" "∧x. x ∈ B ==> x ∉ A ==> f x = 0" "A ⊆ B"
  shows "(∫🪙+x. f x ∂count_space B) = (∑x∈A. f x)"
proof -
  have "(∫🪙+x. f x ∂count_space B) = (∑a | a ∈ B ∧ 0 < f a. f a)"
    using assms(2,3)
    by (intro nn_integral_count_space finite_subset[OF _ ‹finite A›]) (auto simp: less_le)
  also have "… = (∑a∈A. f a)"
    using assms by (intro sum.mono_neutral_cong_left) (auto simp: less_le)
  finally show ?thesis .
qed

lemma nn_integral_bij_count_space:
  assumes g: "bij_betw g A B"
  shows "(∫🪙+x. f (g x) ∂count_space A) = (∫🪙+x. f x ∂count_space B)"
  using g[THEN bij_betw_imp_funcset]
  by (subst distr_bij_count_space[OF g, symmetric])
     (auto intro!: nn_integral_distr[symmetric])

lemma nn_integral_indicator_finite:
  fixes f :: "'a ==> ennreal"
  assumes f: "finite A" and [measurable]: "∧a. a ∈ A ==> {a} ∈ sets M"
  shows "(∫🪙+x. f x * indicator A x ∂M) = (∑x∈A. f x * emeasure M {x})"
proof -
  from f have "(∫🪙+x. f x * indicator A x ∂M) = (∫🪙+x. (∑a∈A. f a * indicator {a} x) ∂M)"
    by (intro nn_integral_cong) (auto simp: indicator_def if_distrib[where f="λa. x * a" for x] sum.If_cases)
  also have "… = (∑a∈A. f a * emeasure M {a})"
    by (subst nn_integral_sum) auto
  finally show ?thesis .
qed

lemma nn_integral_count_space_nat:
  fixes f :: "nat ==> ennreal"
  shows "(∫🪙+i. f i ∂count_space UNIV) = (∑i. f i)"
proof -
  have "(∫🪙+i. f i ∂count_space UNIV) =
    (∫🪙+i. (∑j. f j * indicator {j} i) ∂count_space UNIV)"
  proof (intro nn_integral_cong)
    fix i
    have "f i = (∑j∈{i}. f j * indicator {j} i)"
      by simp
    also have "… = (∑j. f j * indicator {j} i)"
      by (rule suminf_finite[symmetric]) auto
    finally show "f i = (∑j. f j * indicator {j} i)" .
  qed
  also have "… = (∑j. (∫🪙+i. f j * indicator {j} i ∂count_space UNIV))"
    by (rule nn_integral_suminf) auto
  finally show ?thesis
    by simp
qed

lemma nn_integral_enat_function:
  assumes f: "f ∈ measurable M (count_space UNIV)"
  shows "(∫🪙+ x. ennreal_of_enat (f x) ∂M) = (∑t. emeasure M {x ∈ space M. t < f x})"
proof -
  define F where "F i = {x∈space M. i < f x}" for i :: nat
  with assms have [measurable]: "∧i. F i ∈ sets M"
    by auto

  { fix x assume "x ∈ space M"
    have "(λi::nat. if i < f x then 1 else 0) sums ennreal_of_enat (f x)"
      using sums_If_finite[of "λr. r < f x" "λ_. 1 :: ennreal"]
      by (cases "f x") (simp_all add: sums_def of_nat_tendsto_top_ennreal)
    also have "(λi. (if i < f x then 1 else 0)) = (λi. indicator (F i) x)"
      using ‹x ∈ space M› by (simp add: one_ennreal_def F_def fun_eq_iff)
    finally have "ennreal_of_enat (f x) = (∑i. indicator (F i) x)"
      by (simp add: sums_iff) }
  then have "(∫🪙+x. ennreal_of_enat (f x) ∂M) = (∫🪙+x. (∑i. indicator (F i) x) ∂M)"
    by (simp cong: nn_integral_cong)
  also have "… = (∑i. emeasure M (F i))"
    by (simp add: nn_integral_suminf)
  finally show ?thesis
    by (simp add: F_def)
qed

lemma nn_integral_count_space_nn_integral:
  fixes f :: "'i ==> 'a ==> ennreal"
  assumes "countable I" and [measurable]: "∧i. i ∈ I ==> f i ∈ borel_measurable M"
  shows "(∫🪙+x. ∫🪙+i. f i x ∂count_space I ∂M) = (∫🪙+i. ∫🪙+x. f i x ∂M ∂count_space I)"
proof cases
  assume "finite I" then show ?thesis
    by (simp add: nn_integral_count_space_finite nn_integral_sum)
next
  assume "infinite I"
  then have [simp]: "I ≠ {}"
    by auto
  note * = bij_betw_from_nat_into[OF ‹countable I› ‹infinite I›]
  have **: "∧f. (∧i. 0 ≤ f i) ==> (∫🪙+i. f i ∂count_space I) = (∑n. f (from_nat_into I n))"
    by (simp add: nn_integral_bij_count_space[symmetric, OF *] nn_integral_count_space_nat)
  show ?thesis
    by (simp add: ** nn_integral_suminf from_nat_into)
qed

lemma of_bool_Bex_eq_nn_integral:
  assumes unique: "∧x y. x ∈ X ==> y ∈ X ==> P x ==> P y ==> x = y"
  shows "of_bool (∃y∈X. P y) = (∫🪙+y. of_bool (P y) ∂count_space X)"
proof cases
  assume "∃y∈X. P y"
  then obtain y where "P y" "y ∈ X" by auto
  then show ?thesis
    by (subst nn_integral_count_space'[where A="{y}"]) (auto dest: unique)
qed (auto cong: nn_integral_cong_simp)

lemma emeasure_UN_countable:
  assumes sets[measurable]: "∧i. i ∈ I ==> X i ∈ sets M" and I[simp]: "countable I"
  assumes disj: "disjoint_family_on X I"
  shows "emeasure M (∪(X ` I)) = (∫🪙+i. emeasure M (X i) ∂count_space I)"
proof -
  have eq: "∧x. indicator (∪(X ` I)) x = ∫🪙+ i. indicator (X i) x ∂count_space I"
  proof cases
    fix x assume x: "x ∈ ∪(X ` I)"
    then obtain j where j: "x ∈ X j" "j ∈ I"
      by auto
    with disj have "∧i. i ∈ I ==> indicator (X i) x = (indicator {j} i::ennreal)"
      by (auto simp: disjoint_family_on_def split: split_indicator)
    with x j show "?thesis x"
      by (simp cong: nn_integral_cong_simp)
  qed (auto simp: nn_integral_0_iff_AE)

  note sets.countable_UN'[unfolded subset_eq, measurable]
  have "emeasure M (∪(X ` I)) = (∫🪙+x. indicator (∪(X ` I)) x ∂M)"
    by simp
  also have "… = (∫🪙+i. ∫🪙+x. indicator (X i) x ∂M ∂count_space I)"
    by (simp add: eq nn_integral_count_space_nn_integral)
  finally show ?thesis
    by (simp cong: nn_integral_cong_simp)
qed

lemma emeasure_countable_singleton:
  assumes sets: "∧x. x ∈ X ==> {x} ∈ sets M" and X: "countable X"
  shows "emeasure M X = (∫🪙+x. emeasure M {x} ∂count_space X)"
proof -
  have "emeasure M (∪i∈X. {i}) = (∫🪙+x. emeasure M {x} ∂count_space X)"
    using assms by (intro emeasure_UN_countable) (auto simp: disjoint_family_on_def)
  also have "(∪i∈X. {i}) = X" by auto
  finally show ?thesis .
qed

lemma measure_eqI_countable:
  assumes [simp]: "sets M = Pow A" "sets N = Pow A" and A: "countable A"
  assumes eq: "∧a. a ∈ A ==> emeasure M {a} = emeasure N {a}"
  shows "M = N"
proof (rule measure_eqI)
  fix X assume "X ∈ sets M"
  then have X: "X ⊆ A" by auto
  moreover from A X have "countable X" by (auto dest: countable_subset)
  ultimately have
    "emeasure M X = (∫🪙+a. emeasure M {a} ∂count_space X)"
    "emeasure N X = (∫🪙+a. emeasure N {a} ∂count_space X)"
    by (auto intro!: emeasure_countable_singleton)
  moreover have "(∫🪙+a. emeasure M {a} ∂count_space X) = (∫🪙+a. emeasure N {a} ∂count_space X)"
    using X by (intro nn_integral_cong eq) auto
  ultimately show "emeasure M X = emeasure N X"
    by simp
qed simp

lemma measure_eqI_countable_AE:
  assumes [simp]: "sets M = UNIV" "sets N = UNIV"
  assumes ae: "AE x in M. x ∈ Ω" "AE x in N. x ∈ Ω" and [simp]: "countable Ω"
  assumes eq: "∧x. x ∈ Ω ==> emeasure M {x} = emeasure N {x}"
  shows "M = N"
proof (rule measure_eqI)
  fix A
  have "emeasure N A = emeasure N {x∈Ω. x ∈ A}"
    using ae by (intro emeasure_eq_AE) auto
  also have "… = (∫🪙+x. emeasure N {x} ∂count_space {x∈Ω. x ∈ A})"
    by (intro emeasure_countable_singleton) auto
  also have "… = (∫🪙+x. emeasure M {x} ∂count_space {x∈Ω. x ∈ A})"
    by (intro nn_integral_cong eq[symmetric]) auto
  also have "… = emeasure M {x∈Ω. x ∈ A}"
    by (intro emeasure_countable_singleton[symmetric]) auto
  also have "… = emeasure M A"
    using ae by (intro emeasure_eq_AE) auto
  finally show "emeasure M A = emeasure N A" ..
qed simp

lemma nn_integral_monotone_convergence_SUP_nat:
  fixes f :: "'a ==> nat ==> ennreal"
  assumes chain: "Complete_Partial_Order.chain (≤) (f ` Y)"
  and nonempty: "Y ≠ {}"
  shows "(∫🪙+ x. (SUP i∈Y. f i x) ∂count_space UNIV) = (SUP i∈Y. (∫🪙+ x. f i x ∂count_space UNIV))"
  (is "?lhs = ?rhs" is "integral🪙N ?M _ = _")
proof (rule order_class.order.antisym)
  show "?rhs ≤ ?lhs"
    by (auto intro!: SUP_least SUP_upper nn_integral_mono)
next
  have "∃g. incseq g ∧ range g ⊆ (λi. f i x) ` Y ∧ (SUP i∈Y. f i x) = (SUP i. g i)" for x
    by (rule ennreal_Sup_countable_SUP) (simp add: nonempty)
  then obtain g where incseq: "∧x. incseq (g x)"
    and range: "∧x. range (g x) ⊆ (λi. f i x) ` Y"
    and sup: "∧x. (SUP i∈Y. f i x) = (SUP i. g x i)" by moura
  from incseq have incseq': "incseq (λi x. g x i)"
    by(blast intro: incseq_SucI le_funI dest: incseq_SucD)

  have "?lhs = ∫🪙+ x. (SUP i. g x i) ∂?M" by(simp add: sup)
  also have "… = (SUP i. ∫🪙+ x. g x i ∂?M)" using incseq'
    by(rule nn_integral_monotone_convergence_SUP) simp
  also have "… ≤ (SUP i∈Y. ∫🪙+ x. f i x ∂?M)"
  proof(rule SUP_least)
    fix n
    have "∧x. ∃i. g x n = f i x ∧ i ∈ Y" using range by blast
    then obtain I where I: "∧x. g x n = f (I x) x" "∧x. I x ∈ Y" by moura

    have "(∫🪙+ x. g x n ∂count_space UNIV) = (∑x. g x n)"
      by(rule nn_integral_count_space_nat)
    also have "… = (SUP m. ∑x<m. g x n)"
      by(rule suminf_eq_SUP)
    also have "… ≤ (SUP i∈Y. ∫🪙+ x. f i x ∂?M)"
    proof(rule SUP_mono)
      fix m
      show "∃m'∈Y. (∑x<m. g x n) ≤ (∫🪙+ x. f m' x ∂?M)"
      proof(cases "m > 0")
        case False
        thus ?thesis using nonempty by auto
      next
        case True
        let ?Y = "I ` {..<m}"
        have "f ` ?Y ⊆ f ` Y" using I by auto
        with chain have chain': "Complete_Partial_Order.chain (≤) (f ` ?Y)" by(rule chain_subset)
        hence "Sup (f ` ?Y) ∈ f ` ?Y"
          by(rule ccpo_class.in_chain_finite)(auto simp add: True lessThan_empty_iff)
        then obtain m' where "m' < m" and m': "(SUP i∈?Y. f i) = f (I m')" by auto
        have "I m' ∈ Y" using I by blast
        have "(∑x<m. g x n) ≤ (∑x<m. f (I m') x)"
        proof(rule sum_mono)
          fix x
          assume "x ∈ {..<m}"
          hence "x < m" by simp
          have "g x n = f (I x) x" by(simp add: I)
          also have "… ≤ (SUP i∈?Y. f i) x" unfolding Sup_fun_def image_image
            using ‹x ∈ {..› by (rule Sup_upper [OF imageI])
          also have "… = f (I m') x" unfolding m' by simp
          finally show "g x n ≤ f (I m') x" .
        qed
        also have "… ≤ (SUP m. (∑x<m. f (I m') x))"
          by(rule SUP_upper) simp
        also have "… = (∑x. f (I m') x)"
          by(rule suminf_eq_SUP[symmetric])
        also have "… = (∫🪙+ x. f (I m') x ∂?M)"
          by(rule nn_integral_count_space_nat[symmetric])
        finally show ?thesis using ‹I m' ∈ Y› by blast
      qed
    qed
    finally show "(∫🪙+ x. g x n ∂count_space UNIV) ≤ …" .
  qed
  finally show "?lhs ≤ ?rhs" .
qed

lemma power_series_tendsto_at_left:
  assumes nonneg: "∧i. 0 ≤ f i" and summable: "∧z. 0 ≤ z ==> z < 1 ==> summable (λn. f n * z^n)"
  shows "((λz. ennreal (∑n. f n * z^n)) ---> (∑n. ennreal (f n))) (at_left (1::real))"
proof (intro tendsto_at_left_sequentially)
  show "0 < (1::real)" by simp
  fix S :: "nat ==> real" assume S: "∧n. S n < 1" "∧n. 0 < S n" "S <---- 1" "incseq S"
  then have S_nonneg: "∧i. 0 ≤ S i" by (auto intro: less_imp_le)

  have "(λi. (∫🪙+n. f n * S i^n ∂count_space UNIV)) <---- (∫🪙+n. ennreal (f n) ∂count_space UNIV)"
  proof (rule nn_integral_LIMSEQ)
    show "incseq (λi n. ennreal (f n * S i^n))"
      using S by (auto intro!: mult_mono power_mono nonneg ennreal_leI
                       simp: incseq_def le_fun_def less_imp_le)
    fix n have "(λi. ennreal (f n * S i^n)) <---- ennreal (f n * 1^n)"
      by (intro tendsto_intros tendsto_ennrealI S)
    then show "(λi. ennreal (f n * S i^n)) <---- ennreal (f n)"
      by simp
  qed (auto simp: S_nonneg intro!: mult_nonneg_nonneg nonneg)
  also have "(λi. (∫🪙+n. f n * S i^n ∂count_space UNIV)) = (λi. ∑n. f n * S i^n)"
    by (subst nn_integral_count_space_nat)
       (intro ext suminf_ennreal2 mult_nonneg_nonneg nonneg S_nonneg
              zero_le_power summable S)+
  also have "(∫🪙+n. ennreal (f n) ∂count_space UNIV) = (∑n. ennreal (f n))"
    by (simp add: nn_integral_count_space_nat nonneg)
  finally show "(λn. ennreal (∑na. f na * S n ^ na)) <---- (∑n. ennreal (f n))" .
qed

subsubsection ‹Measures with Restricted Space›

lemma simple_function_restrict_space_ennreal:
  fixes f :: "'a ==> ennreal"
  assumes "Ω ∩ space M ∈ sets M"
  shows "simple_function (restrict_space M Ω) f ⟷ simple_function M (λx. f x * indicator Ω x)"
proof -
  { assume "finite (f ` space (restrict_space M Ω))"
    then have "finite (f ` space (restrict_space M Ω) ∪ {0})" by simp
    then have "finite ((λx. f x * indicator Ω x) ` space M)"
      by (rule rev_finite_subset) (auto split: split_indicator simp: space_restrict_space) }
  moreover
  { assume "finite ((λx. f x * indicator Ω x) ` space M)"
    then have "finite (f ` space (restrict_space M Ω))"
      by (rule rev_finite_subset) (auto split: split_indicator simp: space_restrict_space) }
  ultimately show ?thesis
    unfolding
      simple_function_iff_borel_measurable borel_measurable_restrict_space_iff_ennreal[OF assms]
    by auto
qed

lemma simple_function_restrict_space:
  fixes f :: "'a ==> 'b::real_normed_vector"
  assumes "Ω ∩ space M ∈ sets M"
  shows "simple_function (restrict_space M Ω) f ⟷ simple_function M (λx. indicator Ω x *🪙R f x)"
proof -
  { assume "finite (f ` space (restrict_space M Ω))"
    then have "finite (f ` space (restrict_space M Ω) ∪ {0})" by simp
    then have "finite ((λx. indicator Ω x *🪙R f x) ` space M)"
      by (rule rev_finite_subset) (auto split: split_indicator simp: space_restrict_space) }
  moreover
  { assume "finite ((λx. indicator Ω x *🪙R f x) ` space M)"
    then have "finite (f ` space (restrict_space M Ω))"
      by (rule rev_finite_subset) (auto split: split_indicator simp: space_restrict_space) }
  ultimately show ?thesis
    unfolding simple_function_iff_borel_measurable
      borel_measurable_restrict_space_iff[OF assms]
    by auto
qed

lemma simple_integral_restrict_space:
  assumes Ω: "Ω ∩ space M ∈ sets M" "simple_function (restrict_space M Ω) f"
  shows "simple_integral (restrict_space M Ω) f = simple_integral M (λx. f x * indicator Ω x)"
  using simple_function_restrict_space_ennreal[THEN iffD1, OF Ω, THEN simple_functionD(1)]
  by (auto simp add: space_restrict_space emeasure_restrict_space[OF Ω(1)] le_infI2 simple_integral_def
           split: split_indicator split_indicator_asm
           intro!: sum.mono_neutral_cong_left ennreal_mult_left_cong arg_cong2[where f=emeasure])

lemma nn_integral_restrict_space:
  assumes Ω[simp]: "Ω ∩ space M ∈ sets M"
  shows "nn_integral (restrict_space M Ω) f = nn_integral M (λx. f x * indicator Ω x)"
proof -
  let ?R = "restrict_space M Ω" and ?X = "λM f. {s. simple_function M s ∧ s ≤ f ∧ (∀x. s x < top)}"
  have "integral🪙S ?R ` ?X ?R f = integral🪙S M ` ?X M (λx. f x * indicator Ω x)"
  proof (safe intro!: image_eqI)
    fix s assume s: "simple_function ?R s" "s ≤ f" "∀x. s x < top"
    from s show "integral🪙S (restrict_space M Ω) s = integral🪙S M (λx. s x * indicator Ω x)"
      by (intro simple_integral_restrict_space) auto
    from s show "simple_function M (λx. s x * indicator Ω x)"
      by (simp add: simple_function_restrict_space_ennreal)
    from s show "(λx. s x * indicator Ω x) ≤ (λx. f x * indicator Ω x)"
      "∧x. s x * indicator Ω x < top"
      by (auto split: split_indicator simp: le_fun_def image_subset_iff)
  next
    fix s assume s: "simple_function M s" "s ≤ (λx. f x * indicator Ω x)" "∀x. s x < top"
    then have "simple_function M (λx. s x * indicator (Ω ∩ space M) x)" (is ?s')
      by (intro simple_function_mult simple_function_indicator) auto
    also have "?s' ⟷ simple_function M (λx. s x * indicator Ω x)"
      by (rule simple_function_cong) (auto split: split_indicator)
    finally show sf: "simple_function (restrict_space M Ω) s"
      by (simp add: simple_function_restrict_space_ennreal)

    from s have s_eq: "s = (λx. s x * indicator Ω x)"
      by (auto simp add: fun_eq_iff le_fun_def image_subset_iff
                  split: split_indicator split_indicator_asm
                  intro: antisym)

    show "integral🪙S M s = integral🪙S (restrict_space M Ω) s"
      by (subst s_eq) (rule simple_integral_restrict_space[symmetric, OF Ω sf])
    show "∧x. s x < top"
      using s by (auto simp: image_subset_iff)
    from s show "s ≤ f"
      by (subst s_eq) (auto simp: image_subset_iff le_fun_def split: split_indicator split_indicator_asm)
  qed
  then show ?thesis
    unfolding nn_integral_def_finite by (simp cong del: SUP_cong_simp)
qed

lemma nn_integral_count_space_indicator:
  assumes "NO_MATCH (UNIV::'a set) (X::'a set)"
  shows "(∫🪙+x. f x ∂count_space X) = (∫🪙+x. f x * indicator X x ∂count_space UNIV)"
  by (simp add: nn_integral_restrict_space[symmetric] restrict_count_space)

lemma nn_integral_count_space_eq:
  "(∧x. x ∈ A - B ==> f x = 0) ==> (∧x. x ∈ B - A ==> f x = 0) ==>
    (∫🪙+x. f x ∂count_space A) = (∫🪙+x. f x ∂count_space B)"
  by (auto simp: nn_integral_count_space_indicator intro!: nn_integral_cong split: split_indicator)

lemma nn_integral_ge_point:
  assumes "x ∈ A"
  shows "p x ≤ ∫🪙+ x. p x ∂count_space A"
proof -
  from assms have "p x ≤ ∫🪙+ x. p x ∂count_space {x}"
    by(auto simp add: nn_integral_count_space_finite max_def)
  also have "… = ∫🪙+ x'. p x' * indicator {x} x' ∂count_space A"
    using assms by(auto simp add: nn_integral_count_space_indicator indicator_def intro!: nn_integral_cong)
  also have "… ≤ ∫🪙+ x. p x ∂count_space A"
    by(rule nn_integral_mono)(simp add: indicator_def)
  finally show ?thesis .
qed

subsubsection ‹Measure spaces with an associated density›

definition🍋‹tag important› density :: "'a measure ==> ('a ==> ennreal) ==> 'a measure" where
  "density M f = measure_of (space M) (sets M) (λA. ∫🪙+ x. f x * indicator A x ∂M)"

lemma
  shows sets_density[simp, measurable_cong]: "sets (density M f) = sets M"
    and space_density[simp]: "space (density M f) = space M"
  by (auto simp: density_def)

(* FIXME: add conversion to simplify space, sets and measurable *)
lemma space_density_imp[measurable_dest]:
  "∧x M f. x ∈ space (density M f) ==> x ∈ space M" by auto

lemma
  shows measurable_density_eq1[simp]: "g ∈ measurable (density Mg f) Mg' ⟷ g ∈ measurable Mg Mg'"
    and measurable_density_eq2[simp]: "h ∈ measurable Mh (density Mh' f) ⟷ h ∈ measurable Mh Mh'"
    and simple_function_density_eq[simp]: "simple_function (density Mu f) u ⟷ simple_function Mu u"
  unfolding measurable_def simple_function_def by simp_all

lemma density_cong: "f ∈ borel_measurable M ==> f' ∈ borel_measurable M ==>
  (AE x in M. f x = f' x) ==> density M f = density M f'"
  unfolding density_def by (auto intro!: measure_of_eq nn_integral_cong_AE sets.space_closed)

lemma emeasure_density:
  assumes f[measurable]: "f ∈ borel_measurable M" and A[measurable]: "A ∈ sets M"
  shows "emeasure (density M f) A = (∫🪙+ x. f x * indicator A x ∂M)"
    (is "_ = ?μ A")
  unfolding density_def
proof (rule emeasure_measure_of_sigma)
  show "sigma_algebra (space M) (sets M)" ..
  show "positive (sets M) ?μ"
    using f by (auto simp: positive_def)
  show "countably_additive (sets M) ?μ"
  proof (intro countably_additiveI)
    fix A :: "nat ==> 'a set" assume "range A ⊆ sets M"
    then have "∧i. A i ∈ sets M" by auto
    then have *: "∧i. (λx. f x * indicator (A i) x) ∈ borel_measurable M"
      by auto
    assume disj: "disjoint_family A"
    then have "(∑n. ?μ (A n)) = (∫🪙+ x. (∑n. f x * indicator (A n) x) ∂M)"
       using f * by (subst nn_integral_suminf) auto
    also have "(∫🪙+ x. (∑n. f x * indicator (A n) x) ∂M) = (∫🪙+ x. f x * (∑n. indicator (A n) x) ∂M)"
      using f by (auto intro!: ennreal_suminf_cmult nn_integral_cong_AE)
    also have "… = (∫🪙+ x. f x * indicator (∪n. A n) x ∂M)"
      unfolding suminf_indicator[OF disj] ..
    finally show "(∑i. ∫🪙+ x. f x * indicator (A i) x ∂M) = ∫🪙+ x. f x * indicator (∪i. A i) x ∂M" .
  qed
qed fact

lemma null_sets_density_iff:
  assumes f: "f ∈ borel_measurable M"
  shows "A ∈ null_sets (density M f) ⟷ A ∈ sets M ∧ (AE x in M. x ∈ A ⟶ f x = 0)"
proof -
  { assume "A ∈ sets M"
    have "(∫🪙+x. f x * indicator A x ∂M) = 0 ⟷ emeasure M {x ∈ space M. f x * indicator A x ≠ 0} = 0"
      using f ‹A ∈ sets M› by (intro nn_integral_0_iff) auto
    also have "… ⟷ (AE x in M. f x * indicator A x = 0)"
      using f ‹A ∈ sets M› by (intro AE_iff_measurable[OF _ refl, symmetric]) auto
    also have "(AE x in M. f x * indicator A x = 0) ⟷ (AE x in M. x ∈ A ⟶ f x ≤ 0)"
      by (auto simp add: indicator_def max_def split: if_split_asm)
    finally have "(∫🪙+x. f x * indicator A x ∂M) = 0 ⟷ (AE x in M. x ∈ A ⟶ f x ≤ 0)" . }
  with f show ?thesis
    by (simp add: null_sets_def emeasure_density cong: conj_cong)
qed

lemma AE_density:
  assumes f: "f ∈ borel_measurable M"
  shows "(AE x in density M f. P x) ⟷ (AE x in M. 0 < f x ⟶ P x)"
proof
  assume "AE x in density M f. P x"
  with f obtain N where "{x ∈ space M. ¬ P x} ⊆ N" "N ∈ sets M" and ae: "AE x in M. x ∈ N ⟶ f x = 0"
    by (auto simp: eventually_ae_filter null_sets_density_iff)
  then have "AE x in M. x ∉ N ⟶ P x" by auto
  with ae show "AE x in M. 0 < f x ⟶ P x"
    by (rule eventually_elim2) auto
next
  fix N assume ae: "AE x in M. 0 < f x ⟶ P x"
  then obtain N where "{x ∈ space M. ¬ (0 < f x ⟶ P x)} ⊆ N" "N ∈ null_sets M"
    by (auto simp: eventually_ae_filter)
  then have *: "{x ∈ space (density M f). ¬ P x} ⊆ N ∪ {x∈space M. f x = 0}"
    "N ∪ {x∈space M. f x = 0} ∈ sets M" and ae2: "AE x in M. x ∉ N"
    using f by (auto simp: subset_eq zero_less_iff_neq_zero intro!: AE_not_in)
  show "AE x in density M f. P x"
    using ae2
    unfolding eventually_ae_filter[of _ "density M f"] Bex_def null_sets_density_iff[OF f]
    by (intro exI[of _ "N ∪ {x∈space M. f x = 0}"] conjI *) (auto elim: eventually_elim2)
qed

lemma🍋‹tag important› nn_integral_density:
  assumes f: "f ∈ borel_measurable M"
  assumes g: "g ∈ borel_measurable M"
  shows "integral🪙N (density M f) g = (∫🪙+ x. f x * g x ∂M)"
using g proof induct
  case (cong u v)
  then show ?case
    apply (subst nn_integral_cong[OF cong(3)])
    apply (simp_all cong: nn_integral_cong)
    done
next
  case (set A) then show ?case
    by (simp add: emeasure_density f)
next
  case (mult u c)
  moreover have "∧x. f x * (c * u x) = c * (f x * u x)" by (simp add: field_simps)
  ultimately show ?case
    using f by (simp add: nn_integral_cmult)
next
  case (add u v)
  then have "∧x. f x * (v x + u x) = f x * v x + f x * u x"
    by (simp add: distrib_left)
  with add f show ?case
    by (auto simp add: nn_integral_add intro!: nn_integral_add[symmetric])
next
  case (seq U)
  have eq: "AE x in M. f x * (SUP i. U i x) = (SUP i. f x * U i x)"
    by eventually_elim (simp add: SUP_mult_left_ennreal seq)
  from seq f show ?case
    apply (simp add: nn_integral_monotone_convergence_SUP image_comp)
    apply (subst nn_integral_cong_AE[OF eq])
    apply (subst nn_integral_monotone_convergence_SUP_AE)
    apply (auto simp: incseq_def le_fun_def intro!: mult_left_mono)
    done
qed

lemma density_distr:
  assumes [measurable]: "f ∈ borel_measurable N" "X ∈ measurable M N"
  shows "density (distr M N X) f = distr (density M (λx. f (X x))) N X"
  by (intro measure_eqI)
     (auto simp add: emeasure_density nn_integral_distr emeasure_distr
           split: split_indicator intro!: nn_integral_cong)

lemma emeasure_restricted:
  assumes S: "S ∈ sets M" and X: "X ∈ sets M"
  shows "emeasure (density M (indicator S)) X = emeasure M (S ∩ X)"
proof -
  have "emeasure (density M (indicator S)) X = (∫🪙+x. indicator S x * indicator X x ∂M)"
    using S X by (simp add: emeasure_density)
  also have "… = (∫🪙+x. indicator (S ∩ X) x ∂M)"
    by (auto intro!: nn_integral_cong simp: indicator_def)
  also have "… = emeasure M (S ∩ X)"
    using S X by (simp add: sets.Int)
  finally show ?thesis .
qed

lemma measure_restricted:
  "S ∈ sets M ==> X ∈ sets M ==> measure (density M (indicator S)) X = measure M (S ∩ X)"
  by (simp add: emeasure_restricted measure_def)

lemma (in finite_measure) finite_measure_restricted:
  "S ∈ sets M ==> finite_measure (density M (indicator S))"
  by standard (simp add: emeasure_restricted)

lemma emeasure_density_const:
  "A ∈ sets M ==> emeasure (density M (λ_. c)) A = c * emeasure M A"
  by (auto simp: nn_integral_cmult_indicator emeasure_density)

lemma measure_density_const:
  "A ∈ sets M ==> c ≠ ∞ ==> measure (density M (λ_. c)) A = enn2real c * measure M A"
  by (auto simp: emeasure_density_const measure_def enn2real_mult)

lemma density_density_eq:
   "f ∈ borel_measurable M ==> g ∈ borel_measurable M ==>
   density (density M f) g = density M (λx. f x * g x)"
  by (auto intro!: measure_eqI simp: emeasure_density nn_integral_density ac_simps)

lemma distr_density_distr:
  assumes T: "T ∈ measurable M M'" and T': "T' ∈ measurable M' M"
    and inv: "∀x∈space M. T' (T x) = x"
  assumes f: "f ∈ borel_measurable M'"
  shows "distr (density (distr M M' T) f) M T' = density M (f ∘ T)" (is "?R = ?L")
proof (rule measure_eqI)
  fix A assume A: "A ∈ sets ?R"
  { fix x assume "x ∈ space M"
    with sets.sets_into_space[OF A]
    have "indicator (T' -` A ∩ space M') (T x) = (indicator A x :: ennreal)"
      using T inv by (auto simp: indicator_def measurable_space) }
  with A T T' f show "emeasure ?R A = emeasure ?L A"
    by (simp add: measurable_comp emeasure_density emeasure_distr
                  nn_integral_distr measurable_sets cong: nn_integral_cong)
qed simp

lemma density_density_divide:
  fixes f g :: "'a ==> real"
  assumes f: "f ∈ borel_measurable M" "AE x in M. 0 ≤ f x"
  assumes g: "g ∈ borel_measurable M" "AE x in M. 0 ≤ g x"
  assumes ac: "AE x in M. f x = 0 ⟶ g x = 0"
  shows "density (density M f) (λx. g x / f x) = density M g"
proof -
  have "density M g = density M (λx. ennreal (f x) * ennreal (g x / f x))"
    using f g ac by (auto intro!: density_cong measurable_If simp: ennreal_mult[symmetric])
  then show ?thesis
    using f g by (subst density_density_eq) auto
qed

lemma density_1: "density M (λ_. 1) = M"
  by (intro measure_eqI) (auto simp: emeasure_density)

lemma emeasure_density_add:
  assumes X: "X ∈ sets M"
  assumes Mf[measurable]: "f ∈ borel_measurable M"
  assumes Mg[measurable]: "g ∈ borel_measurable M"
  shows "emeasure (density M f) X + emeasure (density M g) X =
           emeasure (density M (λx. f x + g x)) X"
  using assms
  apply (subst (1 2 3) emeasure_density, simp_all) []
  apply (subst nn_integral_add[symmetric], simp_all) []
  apply (intro nn_integral_cong, simp split: split_indicator)
  done

subsubsection ‹Point measure›

definition🍋‹tag important› point_measure :: "'a set ==> ('a ==> ennreal) ==> 'a measure" where
  "point_measure A f = density (count_space A) f"

lemma
  shows space_point_measure: "space (point_measure A f) = A"
    and sets_point_measure: "sets (point_measure A f) = Pow A"
  by (auto simp: point_measure_def)

lemma sets_point_measure_count_space[measurable_cong]: "sets (point_measure A f) = sets (count_space A)"
  by (simp add: sets_point_measure)

lemma measurable_point_measure_eq1[simp]:
  "g ∈ measurable (point_measure A f) M ⟷ g ∈ A → space M"
  unfolding point_measure_def by simp

lemma measurable_point_measure_eq2_finite[simp]:
  "finite A ==>
   g ∈ measurable M (point_measure A f) ⟷
    (g ∈ space M → A ∧ (∀a∈A. g -` {a} ∩ space M ∈ sets M))"
  unfolding point_measure_def by (simp add: measurable_count_space_eq2)

lemma simple_function_point_measure[simp]:
  "simple_function (point_measure A f) g ⟷ finite (g ` A)"
  by (simp add: point_measure_def)

lemma emeasure_point_measure:
  assumes A: "finite {a∈X. 0 < f a}" "X ⊆ A"
  shows "emeasure (point_measure A f) X = (∑a|a∈X ∧ 0 < f a. f a)"
proof -
  have "{a. (a ∈ X ⟶ a ∈ A ∧ 0 < f a) ∧ a ∈ X} = {a∈X. 0 < f a}"
    using ‹X ⊆ A› by auto
  with A show ?thesis
    by (simp add: emeasure_density nn_integral_count_space point_measure_def indicator_def of_bool_def)
qed

lemma emeasure_point_measure_finite:
  "finite A ==> X ⊆ A ==> emeasure (point_measure A f) X = (∑a∈X. f a)"
  by (subst emeasure_point_measure) (auto dest: finite_subset intro!: sum.mono_neutral_left simp: less_le)

lemma emeasure_point_measure_finite_if:
  "finite A ==> emeasure (point_measure A f) X = (if X ⊆ A then ∑a∈X. f a else 0)"
  by (simp add: emeasure_point_measure_finite emeasure_notin_sets sets_point_measure)

lemma measure_point_measure_finite_if:
  assumes "finite A" and "∧x. x ∈ A ==> f x ≥ 0"
  shows "measure (point_measure A f) X = (if X ⊆ A then ∑a∈X. f a else 0)"
  by (simp add: Sigma_Algebra.measure_def assms emeasure_point_measure_finite_if subset_eq sum_nonneg)

lemma emeasure_point_measure_finite2:
  "X ⊆ A ==> finite X ==> emeasure (point_measure A f) X = (∑a∈X. f a)"
  by (subst emeasure_point_measure)
     (auto dest: finite_subset intro!: sum.mono_neutral_left simp: less_le)

lemma null_sets_point_measure_iff:
  "X ∈ null_sets (point_measure A f) ⟷ X ⊆ A ∧ (∀x∈X. f x = 0)"
 by (auto simp: AE_count_space null_sets_density_iff point_measure_def)

lemma AE_point_measure:
  "(AE x in point_measure A f. P x) ⟷ (∀x∈A. 0 < f x ⟶ P x)"
  unfolding point_measure_def
  by (subst AE_density) (auto simp: AE_density AE_count_space point_measure_def)

lemma nn_integral_point_measure:
  "finite {a∈A. 0 < f a ∧ 0 < g a} ==>
    integral🪙N (point_measure A f) g = (∑a|a∈A ∧ 0 < f a ∧ 0 < g a. f a * g a)"
  unfolding point_measure_def
  by (subst nn_integral_density)
     (simp_all add: nn_integral_density nn_integral_count_space ennreal_zero_less_mult_iff)

lemma nn_integral_point_measure_finite:
  "finite A ==> integral🪙N (point_measure A f) g = (∑a∈A. f a * g a)"
  by (subst nn_integral_point_measure) (auto intro!: sum.mono_neutral_left simp: less_le)

subsubsection ‹Uniform measure›

definition🍋‹tag important› "uniform_measure M A = density M (λx. indicator A x / emeasure M A)"

lemma
  shows sets_uniform_measure[simp, measurable_cong]: "sets (uniform_measure M A) = sets M"
    and space_uniform_measure[simp]: "space (uniform_measure M A) = space M"
  by (auto simp: uniform_measure_def)

lemma emeasure_uniform_measure[simp]:
  assumes A: "A ∈ sets M" and B: "B ∈ sets M"
  shows "emeasure (uniform_measure M A) B = emeasure M (A ∩ B) / emeasure M A"
proof -
  from A B have "emeasure (uniform_measure M A) B = (∫🪙+x. (1 / emeasure M A) * indicator (A ∩ B) x ∂M)"
    by (auto simp add: uniform_measure_def emeasure_density divide_ennreal_def split: split_indicator
             intro!: nn_integral_cong)
  also have "… = emeasure M (A ∩ B) / emeasure M A"
    using A B
    by (subst nn_integral_cmult_indicator) (simp_all add: sets.Int divide_ennreal_def mult.commute)
  finally show ?thesis .
qed

lemma measure_uniform_measure[simp]:
  assumes A: "emeasure M A ≠ 0" "emeasure M A ≠ ∞" and B: "B ∈ sets M"
  shows "measure (uniform_measure M A) B = measure M (A ∩ B) / measure M A"
  using emeasure_uniform_measure[OF emeasure_neq_0_sets[OF A(1)] B] A
  by (cases "emeasure M A" "emeasure M (A ∩ B)" rule: ennreal2_cases)
     (simp_all add: measure_def divide_ennreal top_ennreal.rep_eq top_ereal_def ennreal_top_divide)

lemma AE_uniform_measureI:
  "A ∈ sets M ==> (AE x in M. x ∈ A ⟶ P x) ==> (AE x in uniform_measure M A. P x)"
  unfolding uniform_measure_def by (auto simp: AE_density divide_ennreal_def)

lemma emeasure_uniform_measure_1:
  "emeasure M S ≠ 0 ==> emeasure M S ≠ ∞ ==> emeasure (uniform_measure M S) S = 1"
  by (subst emeasure_uniform_measure)
     (simp_all add: emeasure_neq_0_sets emeasure_eq_ennreal_measure divide_ennreal
                    zero_less_iff_neq_zero[symmetric])

lemma nn_integral_uniform_measure:
  assumes f[measurable]: "f ∈ borel_measurable M" and S[measurable]: "S ∈ sets M"
  shows "(∫🪙+x. f x ∂uniform_measure M S) = (∫🪙+x. f x * indicator S x ∂M) / emeasure M S"
proof -
  { assume "emeasure M S = ∞"
    then have ?thesis
      by (simp add: uniform_measure_def nn_integral_density f) }
  moreover
  { assume [simp]: "emeasure M S = 0"
    then have ae: "AE x in M. x ∉ S"
      using sets.sets_into_space[OF S]
      by (subst AE_iff_measurable[OF _ refl]) (simp_all add: subset_eq cong: rev_conj_cong)
    from ae have "(∫🪙+ x. indicator S x / 0 * f x ∂M) = 0"
      by (subst nn_integral_0_iff_AE) auto
    moreover from ae have "(∫🪙+ x. f x * indicator S x ∂M) = 0"
      by (subst nn_integral_0_iff_AE) auto
    ultimately have ?thesis
      by (simp add: uniform_measure_def nn_integral_density f) }
  moreover have "emeasure M S ≠ 0 ==> emeasure M S ≠ ∞ ==> ?thesis"
    unfolding uniform_measure_def
    by (subst nn_integral_density)
       (auto simp: ennreal_times_divide f nn_integral_divide[symmetric] mult.commute)
  ultimately show ?thesis by blast
qed

lemma AE_uniform_measure:
  assumes "emeasure M A ≠ 0" "emeasure M A < ∞"
  shows "(AE x in uniform_measure M A. P x) ⟷ (AE x in M. x ∈ A ⟶ P x)"
proof -
  have "A ∈ sets M"
    using ‹emeasure M A ≠ 0› by (metis emeasure_notin_sets)
  moreover have "∧x. 0 < indicator A x / emeasure M A ⟷ x ∈ A"
    using assms
    by (cases "emeasure M A") (auto split: split_indicator simp: ennreal_zero_less_divide)
  ultimately show ?thesis
    unfolding uniform_measure_def by (simp add: AE_density)
qed

subsubsection🍋‹tag unimportant› ‹Null measure›

lemma null_measure_eq_density: "null_measure M = density M (λ_. 0)"
  by (intro measure_eqI) (simp_all add: emeasure_density)

lemma nn_integral_null_measure[simp]: "(∫🪙+x. f x ∂null_measure M) = 0"
  by (auto simp add: nn_integral_def simple_integral_def SUP_constant bot_ennreal_def le_fun_def
           intro!: exI[of _ "λx. 0"])

lemma density_null_measure[simp]: "density (null_measure M) f = null_measure M"
proof (intro measure_eqI)
  fix A show "emeasure (density (null_measure M) f) A = emeasure (null_measure M) A"
    by (simp add: density_def) (simp only: null_measure_def[symmetric] emeasure_null_measure)
qed simp

subsubsection ‹Uniform count measure›

definition🍋‹tag important› "uniform_count_measure A = point_measure A (λx. 1 / card A)"

lemma
  shows space_uniform_count_measure: "space (uniform_count_measure A) = A"
    and sets_uniform_count_measure: "sets (uniform_count_measure A) = Pow A"
    unfolding uniform_count_measure_def by (auto simp: space_point_measure sets_point_measure)

lemma sets_uniform_count_measure_count_space[measurable_cong]:
  "sets (uniform_count_measure A) = sets (count_space A)"
  by (simp add: sets_uniform_count_measure)

lemma emeasure_uniform_count_measure:
  "finite A ==> X ⊆ A ==> emeasure (uniform_count_measure A) X = card X / card A"
  by (simp add: emeasure_point_measure_finite uniform_count_measure_def divide_inverse ennreal_mult
                ennreal_of_nat_eq_real_of_nat)

lemma emeasure_uniform_count_measure_if:
  "finite A ==> emeasure (uniform_count_measure A) X = (if X ⊆ A then card X / card A else 0)"
  by (simp add: emeasure_notin_sets emeasure_uniform_count_measure sets_uniform_count_measure)

lemma measure_uniform_count_measure:
  "finite A ==> X ⊆ A ==> measure (uniform_count_measure A) X = card X / card A"
  by (simp add: emeasure_point_measure_finite uniform_count_measure_def measure_def enn2real_mult)

lemma measure_uniform_count_measure_if:
  "finite A ==> measure (uniform_count_measure A) X = (if X ⊆ A then card X / card A else 0)"
  by (simp add: measure_uniform_count_measure measure_notin_sets sets_uniform_count_measure)

lemma space_uniform_count_measure_empty_iff [simp]:
  "space (uniform_count_measure X) = {} ⟷ X = {}"
by(simp add: space_uniform_count_measure)

lemma sets_uniform_count_measure_eq_UNIV [simp]:
  "sets (uniform_count_measure UNIV) = UNIV ⟷ True"
  "UNIV = sets (uniform_count_measure UNIV) ⟷ True"
  by(simp_all add: sets_uniform_count_measure)

subsubsection🍋‹tag unimportant› ‹Scaled measure›

lemma nn_integral_scale_measure:
  assumes f: "f ∈ borel_measurable M"
  shows "nn_integral (scale_measure r M) f = r * nn_integral M f"
  using f
proof induction
  case (cong f g)
  thus ?case
    by(simp add: cong.hyps space_scale_measure cong: nn_integral_cong_simp)
next
  case (mult f c)
  thus ?case
    by(simp add: nn_integral_cmult max_def mult.assoc mult.left_commute)
next
  case (add f g)
  thus ?case
    by(simp add: nn_integral_add distrib_left)
next
  case (seq U)
  thus ?case
    by(simp add: nn_integral_monotone_convergence_SUP SUP_mult_left_ennreal image_comp)
qed simp

end