Abstract: The electromagnetic field optimization algorithm is a relatively novel swarm intelligence optimization algorithm, which uses the repulsive force generated by electromagnetic fields of different polarities to move the electromagnetic particles toward the optimal solution. Aimed at the problems of the standard electromagnetic field optimization algorithm easy to fall into local extreme points and poor convergence accuracy while solving the job shop scheduling problem, a multi-strategy guided electromagnetic field optimization algorithm is proposed. Particles in the algorithm are subject to repulsive forces from three different sources. During the iterative process, the particle's guidance method is determined by calculating the electrical difference, cumulative electrical difference, and comprehensive electrical difference of each mobile strategy, and the probability mutation algorithm is used to avoid being trapped into the locally optimal solution. Through simulation experiments of FT and LA series test cases of job shop scheduling problems, the test results of the new algorithm and other algorithms are compared and analyzed. The research shows that the algorithm has higher precision and faster computing speed.

Key words: electromagnetic field optimization algorithm, job shop scheduling, multi-strategy, repulsive force

摘要： 电磁场优化算法是目前一种比较新颖的群智能优化算法,其利用不同极性电磁场所产生的引斥力,使电磁粒子朝最优解移动。针对标准电磁场优化算法在求解作业车间调度问题时容易陷入局部极值点、收敛精度差等问题,提出了一种多策略引导的电磁场优化算法。算法中粒子受到三种不同来源的引斥力,在迭代过程中通过计算每种移动策略的临代电差、累计电差和综合电差来决定粒子的引导方式,并通过概率变异算法来避免陷入局部最优解。通过作业车间调度问题FT、LA系列测试实例仿真实验,对新算法与其他算法的测试结果进行比较分析,研究表明该算法具有更高的求解精度和更快的计算速度。

关键词: 电磁场优化算法, 作业车间调度, 多策略, 引斥力