KERNEL-FUNCTION BASED PRIMAL-DUAL ALGORITHMS FOR P*(K) LINEAR COMPLEMENTARITY PROBLEMS

Recently, [Y.Q. Bai, M. El Ghami and C. Roos, SIAM J.Opt.15 (2004) 101–128] investigated a new class of kernel functionswhich differs from the class of self-regular kernel functions. The classis defined by some simple conditions on the growth and the barrierbehavior of the kernel function. In this paper we generalize the anal-ysis presented in the above paper forP ( κ) Linear ComplementarityProblems (LCPs). The analysis for LCPs deviates significantly fromthe analysis for linear optimization. Several new tools and techniquesare derived in this paper