You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
feat(Analysis/Convex/Function): composition of strictly and non-strictly convex functions (leanprover-community#40830)
We currently have composition lemmas for `StrictCon{vex/cave}On` of the form:
```lean
theorem StrictConvexOn.comp :
StrictConvexOn 𝕜 (f '' s) g → StrictMonoOn g (f '' s) →
StrictConvexOn 𝕜 s f → s.InjOn f →
StrictConvexOn 𝕜 s (g ∘ f)
```
If we let either function be convex instead of strictly convex we get:
```lean
theorem ConvexOn.comp_strictConvexOn :
ConvexOn 𝕜 (f '' s) g → StrictMonoOn g (f '' s) →
StrictConvexOn 𝕜 s f →
StrictConvexOn 𝕜 s (g ∘ f)
theorem StrictConvexOn.comp_convexOn :
StrictConvexOn 𝕜 (f '' s) g → MonotoneOn g (f '' s) →
ConvexOn 𝕜 s f → s.InjOn f →
StrictConvexOn 𝕜 s (g ∘ f)
```
In a `Module` these are both stronger than the original version (since `StrictConvexOn` implies `ConvexOn`), but they work over any `SMul`.
0 commit comments