Library Coq.Reals.Rprod
Require Import Compare.
Require Import Rbase.
Require Import Rfunctions.
Require Import Rseries.
Require Import PartSum.
Require Import Binomial.
Require Import Lia.
Local Open Scope R_scope.
TT Ak; 0<=k<=N
Fixpoint prod_f_R0 (f:nat -> R) (N:nat) : R :=
match N with
| O => f O
| S p => prod_f_R0 f p * f (S p)
end.
Notation prod_f_SO := (fun An N => prod_f_R0 (fun n => An (S n)) N).
Lemma prod_SO_split :
forall (An:nat -> R) (n k:nat),
(k < n)%nat ->
prod_f_R0 An n =
prod_f_R0 An k * prod_f_R0 (fun l:nat => An (k +1+l)%nat) (n - k -1).
Lemma prod_SO_pos :
forall (An:nat -> R) (N:nat),
(forall n:nat, (n <= N)%nat -> 0 <= An n) -> 0 <= prod_f_R0 An N.
Lemma prod_SO_Rle :
forall (An Bn:nat -> R) (N:nat),
(forall n:nat, (n <= N)%nat -> 0 <= An n <= Bn n) ->
prod_f_R0 An N <= prod_f_R0 Bn N.
match N with
| O => f O
| S p => prod_f_R0 f p * f (S p)
end.
Notation prod_f_SO := (fun An N => prod_f_R0 (fun n => An (S n)) N).
Lemma prod_SO_split :
forall (An:nat -> R) (n k:nat),
(k < n)%nat ->
prod_f_R0 An n =
prod_f_R0 An k * prod_f_R0 (fun l:nat => An (k +1+l)%nat) (n - k -1).
Lemma prod_SO_pos :
forall (An:nat -> R) (N:nat),
(forall n:nat, (n <= N)%nat -> 0 <= An n) -> 0 <= prod_f_R0 An N.
Lemma prod_SO_Rle :
forall (An Bn:nat -> R) (N:nat),
(forall n:nat, (n <= N)%nat -> 0 <= An n <= Bn n) ->
prod_f_R0 An N <= prod_f_R0 Bn N.
Application to factorial
Lemma fact_prodSO :
forall n:nat, INR (fact n) = prod_f_R0 (fun k:nat =>
(match (eq_nat_dec k 0) with
| left _ => 1%R
| right _ => INR k
end)) n.
Lemma le_n_2n : forall n:nat, (n <= 2 * n)%nat.
forall n:nat, INR (fact n) = prod_f_R0 (fun k:nat =>
(match (eq_nat_dec k 0) with
| left _ => 1%R
| right _ => INR k
end)) n.
Lemma le_n_2n : forall n:nat, (n <= 2 * n)%nat.
We prove that (N!)^2<=(2N-k)!*k! forall k in |O;2N|
Lemma RfactN_fact2N_factk :
forall N k:nat,
(k <= 2 * N)%nat ->
Rsqr (INR (fact N)) <= INR (fact (2 * N - k)) * INR (fact k).
Lemma INR_fact_lt_0 : forall n:nat, 0 < INR (fact n).
forall N k:nat,
(k <= 2 * N)%nat ->
Rsqr (INR (fact N)) <= INR (fact (2 * N - k)) * INR (fact k).
Lemma INR_fact_lt_0 : forall n:nat, 0 < INR (fact n).
We have the following inequality : (C 2N k) <= (C 2N N) forall k in |O;2N|