柔性设计任务协同调度算法

Collaborative Scheduling Algorithm for Flexible Design Job

指出柔性多任务协同调度是一个NP难题,并分析了协同任务调度在协同设计系统中的重要性,提出一种基于遗传算法和模拟退火算法的混合算法,利用该算法实现设计任务的选择。设计二维结构的矩阵编码,并基于这种编码方式,提出行算子与列算子,融入约束条件,采用列交叉算子与列变异算子;为了加快群体的收敛性,采用精英保留策略;此外引入灾变算子,以保证群体的多样性;在个体生成过程中,考虑能力等相关因素对设计效果的影响,在解码过程中实现任务的时间调度与优化,并设计解码算法。通过实例仿真分析,所提出的混合遗传算法收敛速度快,寻优能力强。

Flexible job scheduling （FJS） is pointed out as a NP-problem, and the significance of FJS in collaborative design system is analyzed. A hybrid algorithm based on genetic algorithm （GA） and simulated annealing （SA） is proposed, which is used to schedule the tasks. A two dimensional matrix encoding is designed, and on this basis, row operator and column operator are put forward, column crossover operator and column mutation operator are adopted by considering the constraints. Elitism preservation strategy is employed for accelerating colony convergence. Moreover, catastrophic operator is imported to guarantee the diversity of colony. Capabilities and other factors that will influence the design results are considered in the generation process of individuals. Time scheduling and optimization are implemented in the decoding process, and decoding algorithm is contrived. A simulation experiment is carried out by using the proposed algorithm, the results shows fast convergence and strong optimization ability.