基于集群优化蚁群算法的加工空行程优化方法

Cluster optimization ant colony algorithm based non-productive tool path optimization method

针对航天壁板结构件空行程优化的大规模非对称性,提出一种将最大最小蚁群系统与节点集优化算法和3-opt局部搜索算子结合的集群优化蚁群算法。在该算法中,以腔体特征为单位进行编码,通过节点集优化算法在多项式时间内计算当前腔体序列的最佳进/退刀点的选择方案;腔体间的启发式信息跟随当前蚂蚁已访问过的路径而变化,其值为当前腔体的所有进/退刀点组合相对于前一腔体的空走时间的增量期望值;提出一种非对称问题的3-opt局部搜索算子,在Hamilton回路的邻域重组中不倒转路径方向,进而通过复用相邻节点集间的最短路径计算结果降低算法复杂度。该算法将两阶段问题作为一个整体进行求解,使得每次迭代都能找到当前最优解的最佳邻域,保证了求解效率和精度。

For the large-scale and asymmetric non-productive tool path problem in aerospace panel part machining,a cluster optimization ant colony algorithm by combining the Max-Min Ant System(MMAS)with cluster optimization algorithm and 3-opt local search heuristic was proposed.In this algorithm,cavities were encoded as units,and the cluster optimization algorithm was used to calculate the optimal feed/retreat points selection for the current cavity sequence in polynomial time.The heuristic information between cavities varied with the visited path of the ant,and its value were the expected idle time increment of all the feed/retreat point combinations in current cavity relative to that of the previous cavity.A 3-opt heuristic for asymmetric problems was given,which did not reverse the path direction in the neighborhood reconstruction of the Hamiltonian loop.It reduced algorithm complexity by reusing the shortest path calculation results between adjacent node clusters.This algorithm solved the two-stage problem as a whole,enabling itself to find the best neighborhood of the current optimal solution in each iteration,which ensured its efficiency and accuracy in solving problems.