Volume 2, Number 3, September 2006, pp. 679-692
J.S. Pang and J. Sun
Key words:
piecewise quadratic program, Nash equilibria, sequential penalization
Mathematices Subject Classification: 90C20, 91A06, 91B50
ONLINE SUBSCRIPTION (Institutional Subscription Only)
Copyright© 2006 Yokohama Publishers
Back

Abstract:
Inspired by an applied model arising from electric power markets with price caps that is discussed in a previous paper [3], this paper studies the Nash equilibrium problem in which the minimizing players' objective functions are sums of composite separable convex piecewise quadratic functions. Based on a fundamental but previously not proven equivalence between a separable convex piecewise quadratic program and a standard convex quadratic program, we show that the nonsmooth Nash equilibrium problem can be equivalently reformulated as a generalized Nash equilibrium problem with coupled linear constraints. We establish the convergence of a sequential penalized Nash algorithm for solving the reformulated generalized Nash problem under a boundedness condition.

Nash equilibria with piecewise quadratic osts