Documentation

TauCeti.Analysis.Normed.Ring.Ultra

The closed unit ball and products of elements close to one in ultrametric normed rings #

An ultrametric seminormed ring is a nonarchimedean topological ring: its open balls about zero are open additive subgroups. In an ultrametric seminormed ring with ‖1‖ = 1, the closed unit ball is an open subring. In an ultrametric normed ring, a finite product of elements each within ε ≤ 1 of 1 is again within ε of 1. A series ∑_{n ≥ 1} f(qⁿ w) of terms bounded by their arguments is bounded by its first argument q w, and it depends on w in a ‖q‖-Lipschitz way when its terms are 1-Lipschitz. The rational functions t / (1 - t)², t / (1 - t)³ and t² / (1 - t)³ are Lipschitz on the open unit ball of an ultrametric field.

Integer casts have norm at most ‖1‖, without requiring the seminorm to be normalized. When ‖1‖ = 1, arbitrarily large natural numbers have norm exactly one. Since the forward differences of a monomial are divisible by k!, their norms carry the factor ‖k!‖.

Main results #

Integer casts in an ultrametric seminormed ring have norm at most the norm of one.

In an ultrametric seminormed ring with ‖1‖ = 1, arbitrarily large natural numbers have norm one: of two consecutive natural numbers, one has norm one, since their difference does.

An ultrametric seminormed ring is a nonarchimedean ring. This is the ring form of Mathlib's IsUltrametricDist.nonarchimedeanAddGroup: every neighbourhood of zero contains an open additive subgroup, namely a small open ball.

The closed unit ball of an ultrametric seminormed ring with ‖1‖ = 1 is a subring: it is the submonoid Submonoid.unitClosedBall and the open additive subgroup IsUltrametricDist.closedBall_openAddSubgroup.

Instances For
    @[simp]

    An element lies in the closed unit ball exactly when its norm is at most one.

    The closed unit ball of an ultrametric seminormed ring is open.

    @[simp]

    An element strictly smaller than one does not change the norm of one when subtracted.

    An element at distance less than one from one has norm one.

    Subtracting an element whose norm is not one from one gives norm at least one.

    theorem TauCeti.norm_div_one_sub_pow {K : Type u_1} [NormedDivisionRing K] [IsUltrametricDist K] (z : K) {w : K} (hw : ‖w‖ < 1) (k : ℕ) :
    ‖z / (1 - w) ^ k‖ = ‖z‖

    Dividing by a power of 1 - w with ‖w‖ < 1 does not change norms in an ultrametric normed division ring.

    theorem TauCeti.norm_div_one_sub_cube_sub_div_one_sub_cube_le {K : Type u_1} [NormedField K] [IsUltrametricDist K] {t t' : K} (ht : ‖t‖ < 1) (ht' : ‖t'‖ < 1) :
    ‖t / (1 - t) ^ 3 - t' / (1 - t') ^ 3‖ ≤ ‖t - t'‖

    The map t ↦ t / (1 - t)³ is 1-Lipschitz on the open unit ball of an ultrametric field.

    theorem TauCeti.norm_sq_div_one_sub_cube_sub_sq_div_one_sub_cube_le_max_mul {K : Type u_1} [NormedField K] [IsUltrametricDist K] {t t' : K} (ht : ‖t‖ < 1) (ht' : ‖t'‖ < 1) :
    ‖t ^ 2 / (1 - t) ^ 3 - t' ^ 2 / (1 - t') ^ 3‖ ≤ max ‖t‖ ‖t'‖ * ‖t - t'‖

    On the open unit ball of an ultrametric field, the map t ↦ t² / (1 - t)³ is Lipschitz with constant max ‖t‖ ‖t'‖ between t and t', so it contracts near zero.

    theorem TauCeti.norm_tsum_le_norm_mul {R : Type u_1} {E : Type u_2} [SeminormedRing R] [SeminormedAddCommGroup E] [IsUltrametricDist E] (f : R → E) (hf : ∀ (w : R), ‖w‖ < 1 → ‖f w‖ ≤ ‖w‖) {q w : R} (hq : ‖q‖ ≤ 1) (hw : ‖q * w‖ < 1) :
    ‖∑' (n : ℕ), f (q ^ (n + 1) * w)‖ ≤ ‖q * w‖

    A one-sided series ∑_{n ≥ 1} f(qⁿ w) with values in an ultrametric group, whose terms are bounded by their arguments, has norm at most ‖q w‖ when ‖q‖ ≤ 1 and ‖q w‖ < 1.

    theorem TauCeti.norm_tsum_sub_tsum_le_norm_mul {R : Type u_1} {E : Type u_2} [SeminormedRing R] [NormedAddCommGroup E] [IsUltrametricDist E] (f : R → E) (hf : ∀ (t t' : R), ‖t‖ < 1 → ‖t'‖ < 1 → ‖f t - f t'‖ ≤ ‖t - t'‖) {q w w' : R} (hq : ‖q‖ ≤ 1) (hw : ‖q * w‖ < 1) (hw' : ‖q * w'‖ < 1) (hs : Summable fun (n : ℕ) => f (q ^ (n + 1) * w)) (hs' : Summable fun (n : ℕ) => f (q ^ (n + 1) * w')) :
    ‖∑' (n : ℕ), f (q ^ (n + 1) * w) - ∑' (n : ℕ), f (q ^ (n + 1) * w')‖ ≤ ‖q * (w - w')‖

    Two convergent one-sided series ∑_{n ≥ 1} f(qⁿ w) and ∑_{n ≥ 1} f(qⁿ w') with values in an ultrametric group, whose terms are 1-Lipschitz on the open unit ball, differ by at most ‖q (w - w')‖ when ‖q‖ ≤ 1 and ‖q w‖, ‖q w'‖ < 1.

    theorem TauCeti.norm_tsum_sub_tsum_add_tsum_sub_tsum_le {K : Type u_1} {E : Type u_2} [NormedField K] [NormedAddCommGroup E] [IsUltrametricDist E] {q u u' : K} (hq : ‖q‖ < 1) (hu : ‖u‖ = 1) (hu' : ‖u'‖ = 1) (f g : K → E) (hf : ∀ (t t' : K), ‖t‖ < 1 → ‖t'‖ < 1 → ‖f t - f t'‖ ≤ ‖t - t'‖) (hg : ∀ (t t' : K), ‖t‖ < 1 → ‖t'‖ < 1 → ‖g t - g t'‖ ≤ ‖t - t'‖) (hfs : Summable fun (n : ℕ) => f (q ^ (n + 1) * u)) (hfs' : Summable fun (n : ℕ) => f (q ^ (n + 1) * u')) (hgs : Summable fun (n : ℕ) => g (q ^ (n + 1) * u⁻¹)) (hgs' : Summable fun (n : ℕ) => g (q ^ (n + 1) * u'⁻¹)) :
    ‖∑' (n : ℕ), f (q ^ (n + 1) * u) - ∑' (n : ℕ), f (q ^ (n + 1) * u') + (∑' (n : ℕ), g (q ^ (n + 1) * u⁻¹) - ∑' (n : ℕ), g (q ^ (n + 1) * u'⁻¹))‖ ≤ ‖q‖ * ‖u - u'‖

    A one-sided series ∑_{n ≥ 1} f(qⁿ u) at u and one ∑_{n ≥ 1} g(qⁿ u⁻¹) at u⁻¹, with values in an ultrametric group, together depend on u in a ‖q‖-Lipschitz way on the unit circle when their terms are 1-Lipschitz on the open unit ball and ‖q‖ < 1. Only the values need to be ultrametric.

    theorem TauCeti.norm_sub_le_of_eq_mul_div {K : Type u_1} [NormedField K] [IsUltrametricDist K] {t t' a a' P : K} (ht : ‖t‖ < 1) (ht' : ‖t'‖ < 1) {k : ℕ} (h : a - a' = (t - t') * P / ((1 - t) ^ k * (1 - t') ^ k)) (hP : ‖P‖ ≤ 1) :
    ‖a - a'‖ ≤ ‖t - t'‖

    A difference of the form (t - t') P / ((1 - t)ᵏ (1 - t')ᵏ) with ‖t‖, ‖t'‖ < 1 and ‖P‖ ≤ 1 has norm at most ‖t - t'‖.

    theorem TauCeti.norm_div_one_sub_sq_sub_div_one_sub_sq_le {K : Type u_1} [NormedField K] [IsUltrametricDist K] {t t' : K} (ht : ‖t‖ < 1) (ht' : ‖t'‖ < 1) :
    ‖t / (1 - t) ^ 2 - t' / (1 - t') ^ 2‖ ≤ ‖t - t'‖

    The map t ↦ t / (1 - t)² is 1-Lipschitz on the open unit ball of an ultrametric field.

    theorem TauCeti.norm_one_sub_cube_div_sub_one_sub_cube_div_le {K : Type u_1} [NormedField K] [IsUltrametricDist K] {v w : K} (hv : ‖v‖ = 1) (hw : ‖w‖ = 1) :
    ‖(1 - v) ^ 3 / v - (1 - w) ^ 3 / w‖ ≤ ‖v - w‖

    The map v ↦ (1 - v)³ / v is 1-Lipschitz on the unit circle of an ultrametric field.

    theorem TauCeti.norm_mul_sub_one_lt {R : Type u_1} [SeminormedRing R] [NormOneClass R] [IsUltrametricDist R] {a b : R} {ε : ℝ} (hε : ε ≤ 1) (ha : ‖a - 1‖ < ε) (hb : ‖b - 1‖ < ε) :
    ‖a * b - 1‖ < ε

    In an ultrametric normed ring, if a and b are both within ε ≤ 1 of 1, so is a * b.

    In an ultrametric seminormed ring with ‖1‖ = 1, a power of an element x of norm at most one is no farther from 1 than x is: ‖x ^ n - 1‖ ≤ ‖x - 1‖.

    theorem TauCeti.norm_prod_sub_one_lt {ι : Type u_1} {R : Type u_2} [SeminormedCommRing R] [NormOneClass R] [IsUltrametricDist R] (s : Finset ι) (f : ι → R) {ε : ℝ} (hε0 : 0 < ε) (hε : ε ≤ 1) (hf : ∀ i ∈ s, ‖f i - 1‖ < ε) :
    ‖∏ i ∈ s, f i - 1‖ < ε

    In an ultrametric normed commutative ring, a finite product of elements each within ε ≤ 1 of 1 is again within ε of 1 (for 0 < ε).

    theorem TauCeti.norm_prod_sub_prod_le {ι : Type u_1} {R : Type u_2} [SeminormedCommRing R] [NormOneClass R] [IsUltrametricDist R] {s : Finset ι} {f g : ι → R} {M ε : ℝ} (hM : 0 ≤ M) (hε : 0 ≤ ε) (hf : ∀ i ∈ s, ‖f i‖ ≤ M) (hg : ∀ i ∈ s, ‖g i‖ ≤ M) (hfg : ∀ i ∈ s, ‖f i - g i‖ ≤ ε) :
    ‖∏ i ∈ s, f i - ∏ i ∈ s, g i‖ ≤ ε * M ^ (s.card - 1)

    If two finite families are bounded by M and differ by at most ε in each coordinate, their products differ by at most ε * M ^ (s.card - 1). The exponent uses natural subtraction, so the statement includes the empty product and M = 0.

    theorem TauCeti.norm_fwdDiff_iter_pow_mul_pow_le {A : Type u_1} [SeminormedCommRing A] [NormOneClass A] [IsUltrametricDist A] {c : ℝ} {h x : A} (hh : ‖h‖ ≤ c) (hx : ‖x‖ ≤ c) (k m : ℕ) :
    ‖(fwdDiff h)^[k] (fun (y : A) => y ^ m) x‖ * c ^ k ≤ ‖↑k.factorial‖ * ‖h‖ ^ k * c ^ m

    Forward differences of powers. In an ultrametric seminormed commutative ring with ‖1‖ = 1, for a point x and a step h with ‖x‖ ≤ c and ‖h‖ ≤ c, ‖Δₕᵏ (y ↦ yᵐ)(x)‖ cᵏ ≤ ‖k!‖ ‖h‖ᵏ cᵐ.