面向低成品率光刻机调度问题的反向-灾变共生生物搜索算法

An opposition based-catastrophe symbiotic organisms search algorithm for scheduling of in-line steppers in low yield scenarios

针对传统进化算法在低成品率步进式光刻机的有效调度问题上因参数多或局部搜索能力不足导致求解质量欠佳的情况,以最小化最大完工时间为优化目标,提出了1种反向-灾变共生生物搜索算法。在共生生物搜索算法的种群初始化及寄生阶段使用反向学习增加种群多样性,从而扩展搜索广度;在算法陷入局部最优时进行灾变判断,根据判断结果跳出局部最优,在下1次迭代的互利共生与偏利共生阶段引入变邻域下降法增加搜索深度。对不同成品率情况进行多种算法的仿真实验,对比结果表明,所提算法具有更好的求解质量。

Insufficient ability of local search and numerousparametersmake the solution of traditional evolutionary algorithm have lowquality.To solve the problem,this paper presents an Opposition Based-Catastrophe Symbiotic Organisms Search（OB-CSOS）algorithm,in which optimal object is minimizing the makespan.Oppositopn-based learning was used to incresethe population diversity in the initial phase and the parasitism phase of Symbiotic Organisms Search algorithm which help expand the search width.When the algorithm was trapped in local optimal,catastrophe theory was used to jump out of it.In the next iteration,variableneighborhood descent was used in mutualism phase and commensalism phase to increase the search depth.Compared with a variety of algorithms and yields by simulation,the results show that this algorithmhas higher quality of solution in various yield scenarios.