A one-parametric class of merit functions for the second-order cone complementarity problem

Jein-Shan Chen, Shaohua Pan

Research output: Contribution to journalArticle

20 Citations (Scopus)

Abstract

We investigate a one-parametric class of merit functions for the second-order cone complementarity problem (SOCCP) which is closely related to the popular Fischer-Burmeister (FB) merit function and natural residual merit function. In fact, it will reduce to the FB merit function if the involved parameter τ equals 2, whereas as τ tends to zero, its limit will become a multiple of the natural residual merit function. In this paper, we show that this class of merit functions enjoys several favorable properties as the FB merit function holds, for example, the smoothness. These properties play an important role in the reformulation method of an unconstrained minimization or a nonsmooth system of equations for the SOCCP. Numerical results are reported for some convex second-order cone programs (SOCPs) by solving the unconstrained minimization reformulation of the KKT optimality conditions, which indicate that the FB merit function is not the best. For the sparse linear SOCPs, the merit function corresponding to τ=2.5 or 3 works better than the FB merit function, whereas for the dense convex SOCPs, the merit function with τ=0.1, 0.5 or 1.0 seems to have better numerical performance.

Original languageEnglish
Pages (from-to)581-606
Number of pages26
JournalComputational Optimization and Applications
Volume45
Issue number3
DOIs
Publication statusPublished - 2010 Apr 1

Fingerprint

Second-order Cone
Merit Function
Complementarity Problem
Cones
Convex Order
Unconstrained Minimization
Convex Cone
Reformulation
Class
KKT Conditions
Linear Order
Optimality Conditions
System of equations
Smoothness

Keywords

  • Complementarity
  • Jordan product
  • Merit function
  • Second-order cone

ASJC Scopus subject areas

  • Applied Mathematics
  • Computational Mathematics
  • Control and Optimization

Cite this

A one-parametric class of merit functions for the second-order cone complementarity problem. / Chen, Jein-Shan; Pan, Shaohua.

In: Computational Optimization and Applications, Vol. 45, No. 3, 01.04.2010, p. 581-606.

Research output: Contribution to journalArticle

@article{8a7b5f44888f423c8bb0f5120d15efa7,
title = "A one-parametric class of merit functions for the second-order cone complementarity problem",
abstract = "We investigate a one-parametric class of merit functions for the second-order cone complementarity problem (SOCCP) which is closely related to the popular Fischer-Burmeister (FB) merit function and natural residual merit function. In fact, it will reduce to the FB merit function if the involved parameter τ equals 2, whereas as τ tends to zero, its limit will become a multiple of the natural residual merit function. In this paper, we show that this class of merit functions enjoys several favorable properties as the FB merit function holds, for example, the smoothness. These properties play an important role in the reformulation method of an unconstrained minimization or a nonsmooth system of equations for the SOCCP. Numerical results are reported for some convex second-order cone programs (SOCPs) by solving the unconstrained minimization reformulation of the KKT optimality conditions, which indicate that the FB merit function is not the best. For the sparse linear SOCPs, the merit function corresponding to τ=2.5 or 3 works better than the FB merit function, whereas for the dense convex SOCPs, the merit function with τ=0.1, 0.5 or 1.0 seems to have better numerical performance.",
keywords = "Complementarity, Jordan product, Merit function, Second-order cone",
author = "Jein-Shan Chen and Shaohua Pan",
year = "2010",
month = "4",
day = "1",
doi = "10.1007/s10589-008-9182-9",
language = "English",
volume = "45",
pages = "581--606",
journal = "Computational Optimization and Applications",
issn = "0926-6003",
publisher = "Springer Netherlands",
number = "3",

}

TY - JOUR

T1 - A one-parametric class of merit functions for the second-order cone complementarity problem

AU - Chen, Jein-Shan

AU - Pan, Shaohua

PY - 2010/4/1

Y1 - 2010/4/1

N2 - We investigate a one-parametric class of merit functions for the second-order cone complementarity problem (SOCCP) which is closely related to the popular Fischer-Burmeister (FB) merit function and natural residual merit function. In fact, it will reduce to the FB merit function if the involved parameter τ equals 2, whereas as τ tends to zero, its limit will become a multiple of the natural residual merit function. In this paper, we show that this class of merit functions enjoys several favorable properties as the FB merit function holds, for example, the smoothness. These properties play an important role in the reformulation method of an unconstrained minimization or a nonsmooth system of equations for the SOCCP. Numerical results are reported for some convex second-order cone programs (SOCPs) by solving the unconstrained minimization reformulation of the KKT optimality conditions, which indicate that the FB merit function is not the best. For the sparse linear SOCPs, the merit function corresponding to τ=2.5 or 3 works better than the FB merit function, whereas for the dense convex SOCPs, the merit function with τ=0.1, 0.5 or 1.0 seems to have better numerical performance.

AB - We investigate a one-parametric class of merit functions for the second-order cone complementarity problem (SOCCP) which is closely related to the popular Fischer-Burmeister (FB) merit function and natural residual merit function. In fact, it will reduce to the FB merit function if the involved parameter τ equals 2, whereas as τ tends to zero, its limit will become a multiple of the natural residual merit function. In this paper, we show that this class of merit functions enjoys several favorable properties as the FB merit function holds, for example, the smoothness. These properties play an important role in the reformulation method of an unconstrained minimization or a nonsmooth system of equations for the SOCCP. Numerical results are reported for some convex second-order cone programs (SOCPs) by solving the unconstrained minimization reformulation of the KKT optimality conditions, which indicate that the FB merit function is not the best. For the sparse linear SOCPs, the merit function corresponding to τ=2.5 or 3 works better than the FB merit function, whereas for the dense convex SOCPs, the merit function with τ=0.1, 0.5 or 1.0 seems to have better numerical performance.

KW - Complementarity

KW - Jordan product

KW - Merit function

KW - Second-order cone

UR - http://www.scopus.com/inward/record.url?scp=77950521302&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77950521302&partnerID=8YFLogxK

U2 - 10.1007/s10589-008-9182-9

DO - 10.1007/s10589-008-9182-9

M3 - Article

AN - SCOPUS:77950521302

VL - 45

SP - 581

EP - 606

JO - Computational Optimization and Applications

JF - Computational Optimization and Applications

SN - 0926-6003

IS - 3

ER -