High-Performance Optimization of Genetic Algorithms

In this paper we present an approach to the optimization of genetic algorithms using innovative genetic operators. The objective is to explore the high grade of achievements made in the matter of improving the performance and efficiency and decreasing the execution time of a simple genetic algorithm. The implementation has been based on well known genetic operators -selection and mutation. These have been expanded to form brand new schemes named random walk selection and adaptive threshold mutation respectively. Each of them has separately proved its efficiency when it comes to optimizing the applied genetic algorithm for solving NP hard problems. Random walk selection turns out to be a real success in genetic programming with its ability to produce best results for time considered amazing for the problems concerned, e.g. timetabling (Burke et al., 1995 and Muller, 2003). Adaptive threshold mutation is a scheme that is hard to please but when it is, it contributes most to the optimization of the algorithm. A working program has been developed and successfully used for examining the parameters being a subject of optimization. Necessary comparisons and conclusions have been drawn.