Static and adaptive mutation techniques for genetic algorithm: a systematic comparative analysis

In this paper, a systematic comparative analysis is presented on various static and adaptive mutation techniques to understand their nature on genetic algorithm. Three most popular random mutation techniques such as uniform mutation, Gaussian mutation and boundary mutation, two recently introduced individual adaptive mutation techniques, a self-adaptive mutation technique and a deterministic mutation technique are taken to carry out the analysis. A common experimental bench of benchmark test functions is used to test the techniques and the results are analysed. The analysis intends to identify a best mutation technique for every benchmark problem and to understand the dependency behaviour of mutation techniques with other genetic algorithm parameters such as population sizes, crossover rates and number of generations. Based on the analytical results, interesting findings are obtained that would improve the performance of genetic algorithm.