An Adaptive Penalty Function Method for Constrained Optimization with Evolutionary Programming
Xinghuo Yu and Baolin Wu
Faculty of Informatics and Communication Central Queensland University Rockhampton Qld 4702, Australia
Received:March 15, 1999Accepted:July 20, 1999Published:March 20, 2000
Keywords:Evolutionary computation, Constraints, Fuction optimization, Penalty, Search
In this paper, we propose a novel adaptive penalty function method for constrained optimization problems using the evolutionary programming technique. This method incorporates an adaptive tuning algorithm that adjusts the penalty parameters according to the population landscape so that it allows fast escape from a local optimum and quick convergence toward a global optimum. The method is simple and computationally effective in the sense that only very few penalty parameters are needed for tuning. Simulation results of five well-known benchmark problems are presented to show the performance of the proposed method.
Cite this article as:X. Yu and B. Wu, “An Adaptive Penalty Function Method for Constrained Optimization with Evolutionary Programming,” J. Adv. Comput. Intell. Intell. Inform., Vol.4 No.2, pp. 164-170, 2000.Data files: