免疫多向二进制粒子群优化算法

Immune Multi-Direction Binary Particle Swarm Optimization Algorithm

摘要: 提出了一种新的BPSO改进算法——免疫多向二进制粒子群优化算法(IMBPSO)，算法中引入了免疫算法的克隆选择算子和免疫系统的免疫记忆特性，保证了算法寻优的快速性以及群体的多样性。此外，通过修改速度更新公式，把粒子群的搜索方向从单向变为多向，从而克服了BPSO算法中易陷入局部最优，进化后期收敛速度慢和精度不高等缺点。通过几个标准函数进行测试与评价，证明了IMBPSO算法具有良好的寻优性能。
>A novel algorithm of BPSO is presented, which is named immune multi-direction binary particle swarm op- timization algorithm (IMBPSO). Operators including immune memory and clone selection of immune algorithm are introduced into BPSO in order to ensure the algorithm to find the best solution quickly and the diversity of colony. Fur- thermore, by modifying the formula of renewal of speed, the particle is translated from single direction into multi-di- rection. So it overcomes the disadvantages of BPSO algorithm, including falling into local best easily, low convergence speed as well as low quality in evolution evening. By testing and estimating with some standard functions, IMBPSO algorithm’s ability in finding the best solution is proved.