A new adaptive penalty function in the application of genetic algorithm

Abstract

Penalty function is one of the most commonly used method in genetic algorithm(GA) to solve nonlinear constraint optimization problems.For traditional penalty functions,it is always not easy to control penalty factors.In this paper we present a new adaptive penalty function with simpler construction and prove its convergence.Then based on this adaptive penalty function we present a new genetic algorithm,which can make populations quickly access to feasible regions and improve local search capacity of genetic algorithms.Theoretical analysis and simulation results show that this algorithm has stronger stability and better convergence but needs less parameters than other ones.