Transform of Artificial Immune System algorithm optimization based on mathematical test function

Abstract

Artificial Immune System (AIS) is inspired by nature biological immune system. AIS algorithm has ability to improve the global searching during optimization. However, hypermutation of AIS itself cannot always guarantee a better solution for convergence and accuracy. Therefore Genetic Algorithm (GA) has been used efficiently in solving complex optimization problems. The capability of individual algorithm can makes the new algorithm techniques more efficiency by overcome the shortcomings and without losing their own advantages. This paper demonstrates a hybrid algorithm known as Transform of Artificial Immune System (Trans-AIS) by combining AIS and GA algorithm. There are three mathematical test function are used for comparison to achieve the minimum value which are Rastrigin's, DeJong's and Griewank's functions. In this paper, the simulation of the test function results by using AIS and Trans-AIS will compare with optimization results by other researchers. By comparing the results, it is observed that the performance of Trans-AIS is comparable (if not superior) to other researchers' algorithm.