LeanAide GPT-4 documentation

Module: Mathlib.Analysis.NormedSpace.RCLike


norm_smul_inv_norm

theorem norm_smul_inv_norm : โˆ€ {๐•œ : Type u_1} [inst : RCLike ๐•œ] {E : Type u_2} [inst_1 : NormedAddCommGroup E] [inst_2 : NormedSpace ๐•œ E] {x : E}, x โ‰  0 โ†’ โ€–(โ†‘โ€–xโ€–)โปยน โ€ข xโ€– = 1

This theorem states that for any non-zero vector `x` in a normed space `E` over either the real numbers `โ„` or the complex numbers `โ„‚`, if we scale `x` by the reciprocal of its norm, the norm of the resulting vector is 1. In other words, this process normalizes the vector to unit length.

More concisely: For any non-zero vector `x` in a normed space `E` over `โ„` or `โ„‚`, the normalized vector `x / norm x` has norm 1.

LinearMap.bound_of_ball_bound'

theorem LinearMap.bound_of_ball_bound' : โˆ€ {๐•œ : Type u_1} [inst : RCLike ๐•œ] {E : Type u_2} [inst_1 : NormedAddCommGroup E] [inst_2 : NormedSpace ๐•œ E] {r : โ„}, 0 < r โ†’ โˆ€ (c : โ„) (f : E โ†’โ‚—[๐•œ] ๐•œ), (โˆ€ z โˆˆ Metric.closedBall 0 r, โ€–f zโ€– โ‰ค c) โ†’ โˆ€ (z : E), โ€–f zโ€– โ‰ค c / r * โ€–zโ€–

The theorem `LinearMap.bound_of_ball_bound'` states that for any nontrivially normed field `๐•œ`, a normed additive commutative group `E`, and a normed space over `๐•œ` and `E`, if a real number `r` is positive, then for any real number `c` and a linear map `f` from `E` to `๐•œ`, if the norm of the output of `f` is less than or equal to `c` for all points within a closed ball of radius `r` centered at 0, then the norm of the output of `f` for any point in `E` is less than or equal to `c / r` times the norm of that point. In simpler terms, it says that if the output of a linear map is bounded within a certain sphere around the origin, then you can estimate an upper bound for the output of the linear map anywhere, based on the norm of the input.

More concisely: For any normed field `๐•œ`, normed additive commutative group `E`, and normed space over `๐•œ` and `E`, if a linear map `f` from `E` to `๐•œ` is norm-bounded on a closed ball of radius `r` centered at 0, then for all `x` in `E`, `โ€‰โ€–fxโ€–โ€‰โ‰คโ€‰โ€–xโ€–โ€‰โ€–c/rโ€–`.

norm_smul_inv_norm'

theorem norm_smul_inv_norm' : โˆ€ {๐•œ : Type u_1} [inst : RCLike ๐•œ] {E : Type u_2} [inst_1 : NormedAddCommGroup E] [inst_2 : NormedSpace ๐•œ E] {r : โ„}, 0 โ‰ค r โ†’ โˆ€ {x : E}, x โ‰  0 โ†’ โ€–(โ†‘r * (โ†‘โ€–xโ€–)โปยน) โ€ข xโ€– = r

This theorem states that for any real number `r` that is greater than or equal to zero and any nonzero vector `x` in a normed space `E` over a field `๐•œ` (which is either the real numbers `โ„` or the complex numbers `โ„‚`), the norm (or length) of the vector obtained by scaling `x` by the factor `r` divided by the norm of `x` is equal to `r`. Here, `โ€ข` denotes the scalar multiplication operation in the normed space.

More concisely: For any nonzero vector `x` in a normed space `E` over a field `โ„` or `โ„‚`, and any real number `r โ‰ฅ 0`, the norm of `r * x / ||x||` is equal to `r`.