Cyclical monotonicity

From HandWiki

In mathematics, cyclical monotonicity is a generalization of the notion of monotonicity to the case of vector-valued function.[1][2]

Definition

Let , denote the inner product on an inner product space X and let U be a nonempty subset of X. A correspondence f:UX is called cyclically monotone if for every set of points x1,,xm+1U with xm+1=x1 it holds that k=1mxk+1,f(xk+1)f(xk)0.[3]

Properties

  • For the case of scalar functions of one variable the definition above is equivalent to usual monotonicity.
  • Gradients of convex functions are cyclically monotone.
  • In fact, the converse is true.[4] Suppose U is convex and f:Un is a correspondence with nonempty values. Then if f is cyclically monotone, there exists an upper semicontinuous convex function F:U such that f(x)F(x) for every xU, where F(x) denotes the subgradient of F at x.[5]

See also

References

  1. Levin, Vladimir (1 March 1999). "Abstract Cyclical Monotonicity and Monge Solutions for the General Monge–Kantorovich Problem" (in en). Set-Valued Analysis (Germany: Springer Science+Business Media) 7: 7–32. doi:10.1023/A:1008753021652. 
  2. Beiglböck, Mathias (May 2015). "Cyclical monotonicity and the ergodic theorem" (in en). Ergodic Theory and Dynamical Systems (Cambridge University Press) 35 (3): 710–713. doi:10.1017/etds.2013.75. 
  3. Chambers, Christopher P.; Echenique, Federico (2016). Revealed Preference Theory. Cambridge University Press. p. 9. 
  4. Rockafellar, R. Tyrrell, 1935- (2015-04-29). Convex analysis. Princeton, N.J.. ISBN 9781400873173. OCLC 905969889. [page needed]
  5. http://www.its.caltech.edu/~kcborder/Courses/Notes/CyclicalMonotonicity.pdf [bare URL PDF]