改进的萤火虫算法求解具有学习退化效应的JSP问题

Solving Job-shop Scheduling Problem with Learning and Deterioration Effects by an Improved Firefly Algorithm

作业车间调度问题是典型的NP难题,在生产调度领域具有很高的研究价值.一种更为符合实际的作业车间调度问题是加工机器具有学习退化效应,它能够为生产者安排生产计划提供借鉴.为了可以更好的解决具有学习退化效应的作业车间调度问题,本文提出了改进的萤火虫算法(IFA),即在基本的萤火虫算法基础上增加了局部寻优的过程,并融合了布谷鸟算法中生物移动的莱维分布特点.通过MATLAB模拟分析,IFA能够更快速的收敛到JSP的最优解.最后,本文分析了不同学习率与退化效应因子组合对目标函数求解的影响.

The Job- shop Scheduling Problem( JSP),a typical NP- hard problem,is of great value in theoretical research fields. A more practical situation on JSP is that processing machines have learning and deterioration effects,which can offer reference to make production design. In order to solve the JSP with learning and deterioration effects,an improved firefly algorithm( IFA) is proposed by introducing local optimization researching and making use of the Lévy distribution used in the cuckoo search( CS) algorithm. Simulation with MATLAB shows that the IFA is more effective in standard probes and then can be applied to solve the JSP with learning and deterioration effects.