论文部分内容阅读
本文提出一种带有启发式社会学习机制的人工免疫系统优化算法(AIS-HSL)。在AIS-HSL中,候选抗体被分成两个群,即精英群(ES)和普通群(CS)。不同的群执行不同的变异机制,即精英群采用自学习机制,普通群采用启发式社会学习(HSL)机制。在HSL机制中,CS中的每个抗体根据亲和力依概率选择ES中的抗体进行学习,以避免陷入局部最优。通过一些比较性实验,来评估AIS-HSL算法的性能。实验结果表明,与传统的opt-ai Net算法和IA-AIS算法相比,本文提出的AIS-HSL算法有着更高的收敛精度和收敛速度。
In this paper, an artificial immune system optimization algorithm (AIS-HSL) with a heuristic social learning mechanism is proposed. In AIS-HSL, candidate antibodies are divided into two groups, Elite (ES) and Normal (CS). Different groups perform different mutation mechanisms, namely elite groups adopt self-learning mechanism and common groups adopt heuristic social learning (HSL) mechanism. In the HSL mechanism, each antibody in CS is selected on the basis of probability by probability of antibodies in ES to learn to avoid falling into local optima. Through some comparative experiments to evaluate the performance of AIS-HSL algorithm. The experimental results show that compared with the traditional opt-ai Net algorithm and IA-AIS algorithm, the proposed AIS-HSL algorithm has higher convergence accuracy and convergence rate.