基于种群的多层次迭代贪婪算法优化阻塞流水车间调度问题

Population-based multi-layer iterated greedy algorithm for solving blocking flow shop scheduling problem

针对以最小化最大完工时间为目标的阻塞流水车间调度问题,提出一种有效的基于种群的多层次迭代贪婪算法进行求解。算法基于排列形式进行编码,提出NEH_PF(Nawaz-Enscore-Ham&profile fitting)算法构造初始种群,以提高初始解的质量;设计了多层次迭代贪婪算法,并基于插入、交换策略对个体进行变异,基于Path-Relinking算法对个体进行交叉,以提高个体局部搜索能力,并提出部分交叉策略来更新种群。通过标准实例测试,验证了所提算法的有效性。

Aiming at the blocking flow shop scheduling problem with minimized makespan, an effective population- based multi-layer iterated greedy algorithm was proposed. The permutation based encoding schemes was designed in the algorithm, and Nawaz-Enscore-Ham - Profile Fitting （NEH_PF） algorithm was proposed to construct the ini- tial population for improving the quality of initial results. To enhance the local search abilities of individuals, a multi-layer iterated greedy algorithm was designed, together with the interchange and insert strategies to mutate in- dividuals and path-relinking strategies to cross the individuals. The partly crossing policy was put forward to update the populations. The effectiveness of proposed algorithm was validated through a group of benchmark instances.