Abstract: In flexible job-shop scheduling with single piece and small batch production mode, the optimized objective is to reduce production costs, improve production efficiency and avoid bottleneck. This paper investigates an optimization model of Flexible Job-shop Scheduling Problems（FJSP）, which aims at a comprehensive objective combined with minimized makespan, machine total load and single maximum load. It designs a unidimensional-encoded Particle Swarm Optimization（PSO） taking probability as continuous particle component. Combined with completion-time-earliest heuristic rules, these components are discretized by probability interval to solve operation sequence scheduling and machine tools selecting. After comparing and analyzing different sizes of examples, the proposed algorithm is found a distinct advantage in solving large scale problems.

Key words: Flexible Job-shop Scheduling（FJS）, Particle Swarm Optimization（PSO）, unidimensional-encoded particle, heuristic rules

摘要： 以单件小批量生产方式为主的柔性车间调度中，快速得到满足低生产成本、高生产效率，避免瓶颈发生的调度方案，是调度优化算法的设计目标。就此建立了以制造期、机床总负荷和单机最大负荷为综合目标的柔性车间调度问题（Flexible Job-shop Scheduling Problems，FJSP）优化模型；设计了一种以概率值为分量的一维粒子群优化算法，通过概率区间划分将连续粒子分量离散化，结合完工时间最早启发式规则，实现工序的排序与加工机床的选取。通过不同规模算例的比较，分析结果表明该方法在求解较大规模问题时具有一定的优势。

关键词: 柔性车间调度, 粒子群算法, 一维粒子编码, 启发式规则