Orthogonal Immune Algorithm with Diversity-based Selection for Numerical Optimization

In this study, we design an Orthogonal Immune Algorithm (OIA) for numerical optimization by incorporating orthogonal initialization, a novel neighborhood orthogonal cloning operator, a static hypermutation operator, and a novel diversity-based selection operator. The OIA is unique in three respects: Firstly, a new selection method based on orthogonal arrays is provided in order to maintain diversity in the population. Secondly, the orthogonal design with quantization technique is introduced to generate initial population. Thirdly, the orthogonal design with the modified quantization technique is introduced into the cloning operator. In order to identify any improvement due to orthogonal initialization, diversity-based selection and neighborhood orthogonal cloning, we modify the OIA via replacing its orthogonal initialization by random initialization; replacing its diversity-based selection by a standard evolutionary operator (mu+lambda)-selection operator; and replacing its neighborhood orthogonal cloning by proportional cloning, and compare the four version algorithms in solving eight benchmark functions and six composition functions.