利用改进的协同免疫算法(improved CO-evolutionary immune algorithm,ICIA)求解FlowShop调度问题.算法中的疫苗取自迭代N次的局部最优解,并随着每代最优值的变化不断更新.为了克服协同免疫算法初期收敛速度慢的问题,加入了局部搜索算法;针对算法后期求解目标函数值差的问题,提出了一种新的种群选择机制“80/20法则”.通过与遗传算法(geneticalgorithm,GA)和未改进的协同免疫算法(CO-evolutionary immune algorithm,CIA)比较,仿真实验结果验证了ICIA解决FlowShop问题的有效性.
Abstract: An improved co-evolutionary immune algorithm (ICIA) is developed for the Flow Shop scheduling problem. The vaccine of the algorithm is obtained from the local optimal solution iterated N times, and refreshed with the changing of the optimal value of the different generations. Local search algorithm was applied to avoid the slow convergence in the start of the convergence phase. A new selection mechanism "80/20 Principle" was proposed to improve the objective function value difference of the later convergence phase. Computational results show the effectiveness of the ICIA in solving flow shop scheduling problem compared with GA( genetic algorithm) and CIA( co-evolu- tionary immune algorithm).