Abstract: Marine equipment has the characteristics of extremely large number of workpieces,long development cycle and manufacturing while designing,resulting in uncertain workpiece processing time and low correlation of parallel equipment,which leads to a low coincidence between the actual completion time of the scheduling plan and the planned completion time.To solve the problem,an improved genetic-harmony search algorithm for solving the problem of marine equipment workshop scheduling problem was proposed.A hybrid flow shop scheduling model with workpiece transfer time,specific equipment constraints and fuzzy time was established first,and a matrix coding method based on workpieces and equipment was adopted,then a new Slope algorithm was proposed to generate the initial population.The crossover operator was improved based on the linear sorting and tabu search strategy,and the adaptive mutation rate and two mutation operators were used to improve the global optimization ability.Further,the harmony Search algorithm was used to enhance the local search ability.The performance comparison experiments and example verification results showed that the proposed method was better than other methods in recent literature,and had good practical application value.

Key words: hybrid flow workshop, marine equipment, fuzzy time, genetic algorithm, harmony search

摘要： 海工装备具有工件量极大、研制周期长、边设计边制造等特点,造成工件加工时间不确定、并行设备关联度低等现象,最终导致调度方案的实际完工时间与计划完工时间吻合度偏低。针对该问题,建立带有工件转移时间、特定设备约束和模糊时间的混合流水车间调度模型,提出一种用于求解海工装备制造车间调度问题的改进遗传—和声搜索算法。采用基于工件和设备的矩阵编码方式,提出新的Slope算法用于生成初始种群;基于线性排序和禁忌搜索策略对交叉算子进行改进,使用两种变异算子提高算法全局寻优能力;嵌入和声搜索算法用于增强局部搜索能力。通过性能测试及实例验证表明,相较于已有方法,所提算法具有更好的解和较好的实际应用价值。

关键词: 混合流水车间, 海工装备, 模糊时间, 遗传算法, 和声搜索