Conversation
d9ec1c8 to
66bbe24
Compare
d9e4f68 to
b4c3d8d
Compare
Co-authored-by: Marie Kerjean <marie.kerjean@cnrs.fr>
Co-authored-by: Reynald Affeldt <reynald.affeldt@aist.go.jp>
b4c3d8d to
3e26e7b
Compare
mkerjean
left a comment
There was a problem hiding this comment.
necessary generalisation and shortening of proofs
|
|
||
| Local Open Scope convex_scope. | ||
|
|
||
| Let standard_ball_convex (x : R^o) (r : R) : convex (ball x r). |
There was a problem hiding this comment.
Isn't this an instance of ball_convex above?
There was a problem hiding this comment.
Yes but ball_convex is proved inside an HB builder where the carrier is an lmodType and has a norm,
but outside the builder we do not have such a structure, making it impossible to factorize. :-(
Looks good as a generalization. |
|
What about |
That looks specific enough. Is that conformant with other names for predicates on sets? |
Motivation for this change
Generalization the definition of convex functions in
convex.v, from a functionf: R -> R^oforR: realTypeto a functionf : E -> R^o, forR: numFieldTypeandE : convex_lmodType.This will allow the use of
convex_functionin the statement of Hahn-Banach theorem, and more generally, the use of semi-norms as convex functions on instances ofconvextvsType.fixes issue #1888
Co-authored-by: Reynald Affeldt reynald.affeldt@aist.go.jp
Checklist
CHANGELOG_UNRELEASED.mdReference: How to document
Merge policy
As a rule of thumb:
all compile are preferentially merged into master.
Reminder to reviewers