Numerical Optimization Oriented Artificial Immune Network with Cloud-based Mutation Operator

Abstract

This paper proposes an Artificial Immune Network with the Cloud-based Mutation Operator (AINetCMO) for complex numerical optimization problems. Introducing the cloud model into the mutation operator is expected to enhance the convergence of AINet-CMO. In the mutation process, the mutation step length can be dynamically adapted to the evolution of candidate antibodies by measure of the cloud model. A series of numerical simulations are arranged to compare such performance indices as solution accuracy and convergence speed between AINet-CMO and other existing algorithms. The results indicate that AINet-CMO outperforms the other three artificial immune systems, i.e., opt-aiNet, IA-AIS and AAIS-2S.