J4

• 数学 • Previous Articles     Next Articles

A Combined Maximum Entropy Homotopy Method for Solving the KKT Point of a Class of Nonconvex Nonlinear Programming Problems

SU Menglong1, ZHAO Liqin2, L Xianrui1   

  1. 1. Institute of Mathematics, Jilin University, Changchun 130012, China;2. Editorial Department of Journal of Jilin University, Changchun 130021, China
  • Received:2005-11-28 Revised:1900-01-01 Online:2006-09-26 Published:2006-09-26
  • Contact: SU Menglong

Abstract: We utilized the combined maximum entropy homotopy method to solve the general nonconvex nonlinear programming problems. At first, we transformed a programming problem with many constraints into one with a single constraint using the maximum entropy function. Then we constructed a homotopy equ ation to solve the K-K-T system of the programming problems with a single cons traint and hence obtained an efficient pathfollowing method for solving the programming problems with many constraints. At last we gave the proof of its global convergence.

Key words: combined maximum entropy homotopy method, nonconvex no nlinear programming problems, global convergence

CLC Number: 

  • O221.2