一种免疫粒子群优化算法及在小波神经网络学习中的应用

粒子群优化算法是一类简单有效的随机全局优化技术。受生物体免疫系统抗体多样性保持机制的启发,将抗体多样性保持机制引入到粒子群优化算法中,并给出了一种免疫粒子群优化算法。该算法在保留高适应度粒子的同时,确保了粒子的多样性,从而改善了粒子群优化算法摆脱局部极值点的能力,提高了算法的收敛速度和精度。该算法应用于函数优化和小波神经网络学习的计算机仿真,结果表明该算法有良好的收敛性能。

AN IMMUNE PARTICLE SWARM OPTIMIZATION AND ITS APPLICATION IN TRAINING WAVELET NEURAL NETWORKS

Particle swarm optimization(PSO) is a simple and effective stochastic global optimization technique.In this paper,the antibody diversity maintaining mechanism is introduced into PSO as enlightened by that in creatures' immune body system,thus an immune particle swarm optimization is proposed.The proposed algorithm reserves both good fitness particles and the diversity of particles in evolution process,which improves the ability of PSO in getting rid of the local extreme value and meliorates its convergent s...