The design of Adaptive Immune Genetic Algorithm based on vector distance

Abstract

Aiming at the problems which exist in Genetic Algorithm (GA), including reduction in diversity, prematureness, weak local searching ability and slow convergence rate, this paper studies the effect of antigens recognition module, immune memory module, antibodies self-adjusting module of Immune Algorithm, and adaptive probability crossover and mutation operator to GA, and proposes Adaptive Immune Genetic Algorithm (AIGA) based on vector distance. After exploration, this paper solves the problems in GA above. This paper takes a controlled object as example to test the effect of each module to GA by simulation. Simulation results show that the four modules effectively improve the drawbacks of GA. At the same time, this paper proves the convergence of the algorithm, and verifies algorithm by testing function. Simulation results show that AIGA is better than GA on global optimization capability. The algorithm can obtain the optimal solution with high fitness value, and also has good convergence stability.