Generalized monotonicity of subdifferentials and generalized convexity (Q1367798): Difference between revisions

A function \(f:X\to {\mathbb{R}}\) defined on a Banach space \(X\) is said to be quasiconvex if \[ \forall x,y\in X, \forall t\in [0,1]: f(x+t(y-x)) \leq \max \{f(x),f(y)\} . \] Given a notion of subdifferential \(\partial f\), the authors provide characterizations of the convexity and quasiconvexity of \(f\) in terms of the monotonicity and quasimonotonicity of \(\partial f\). The same kind of result was already considered in the literature for suitable classes of subdifferentials and Banach spaces [see, e.g., \textit{D. Aussel, J.-N. Corvellec} and \textit{M. Lassonde}, Trans. Am. Math. Soc. 347, No. 10, 4147-4161 (1995; Zbl 0849.49016)]. Here the authors consider new settings, like Fréchet and contingent subdifferentials on trustworthy spaces, where only a mild fuzzy calculus rule is satisfied.