基于混合遗传算法的虚拟装配作业过程规划

Hybrid genetic algorithm-based planning method for virutal assemblyoperation

提出了一种基于回报机制和逆向求解机制的装配作业规划方法.使用遗传算法与Q学习算法相结合,共同对作业过程规划问题进行求解.构建适合信息融合和进化计算的"维修元"模型,并将"维修元"进行排列组合,构成装配作业序列,解决装配作业过程中的信息融合问题.为保证作业序列的可行性和维修元的可操作性,将特征匹配机制作为启发机制,避免不可行序列参与进化计算,提高进化效率,解决了虚拟装配过程中多要素融合条件下的规划问题.

To aid in finding the optimal strategy,sequence constraint mechanism and characteristics matching mechanism are proposed,which are suitable for the condition of multi-information fusion.In order to make planning content be more abundant,the concept of maintenance unit is presented,which is suitable for information integration and evolutionary computation.Based on maintenance unit,the assembly sequence model is established.To solve this model,an assembly process planning method based on hybrid genetic algorithm is put forward.To aid in finding the optimal strategy,sequence constraint mechanism and characteristics matching mechanism are proposed,which are suitable for the condition of multi-information fusion.