Abstract: The path optimization for flame cutting has two primary objectives: shortening the cutting path and reducing the thermal error caused by improper path planning. Thermal deformation constraints are quantified by dynamically defining the pierce points set, which can be determined through position relations between parts. Dummy nodes are added in the distance matrix, the cutting path planning problem can be transformed to dynamic TSP. The method for expanding solution space under thermal deformation constraint and the strategy about bidirectional updating of trail pheromones is presented based on ant colony algorithm. Experimental results show that the scale of TSP is restricted efficiently and better solutions can be obtained. It is feasible to optimize the flame cutting path under thermal deformation constraint.

Key words: thermal deformation, pierce points, Traveling Salesman Problem（TSP）, ant colony optimization

摘要： 火焰切割路径优化的主要目的是控制切割路径不当引起的热变形误差并对路径长度寻优。通过零件位置关系动态定义切割过程中的可选打孔点集合，将热变形约束条件量化；引入虚拟结点并定义距离矩阵，将路径规划转化为动态描述的TSP问题；基于蚁群算法提出约束条件下增大解空间的方法和信息素更新策略。实验结果表明，改进后的蚁群算法能够有效控制问题的规模并且得到更高质量的解，对热变形约束条件下的数控火焰切割路径优化有较好的效果和实用性。

关键词: 热变形, 打孔点, 旅行商问题, 蚁群优化