In this paper a novel discrete multi-objective cross-entropy optimization (CrMOCEO) algorithm is proposed to solve the path planning problem of dual-robot cooperative arc welding. We strive to find a low-cost, fast and more efficient solution for robotic welding of large complex components. Firstly, an optimization model of dual-robot welding path planning is established by considering various variables and constraints in the actual welding process. Then, three strategies are introduced to improve the multi-objective cross-entropy optimization (MOCEO) algorithm to better solve the discrete path planning problem. Finally, in order to verify feasibility and effectiveness of the proposed algorithm, it is used to solve the 2-, 3-, 5- and 7-objective WFG2–9 problems and plan some typical welding seams of a large complex component, the MOCEO, NSGA-Ⅱ, MOPSO and MOGWO are used for comparison. The simulation demonstrates that the CrMOCEO can obtain better solutions for multiple objectives than the other four algorithms, and the path solved by the CrMOCEO is tested in the Gazebo physical model and workshop site, the results further verified the effectiveness of the CrMOCEO algorithm. Particularly, a series of experiments provide more solutions for actual production.

中文翻译：

---

基于多目标交叉熵优化的双机器人协作弧焊路径规划算法

本文提出了一种新颖的离散多目标交叉熵优化（CrMOCEO）算法来解决双机器人协作弧焊的路径规划问题。我们努力寻找低成本、快速且更高效的大型复杂部件机器人焊接解决方案。首先，考虑实际焊接过程中的各种变量和约束，建立双机器人焊接路径规划的优化模型。然后，引入三种策略来改进多目标交叉熵优化（MOCEO）算法，以更好地解决离散路径规划问题。最后，为了验证所提算法的可行性和有效性，分别求解了2、3、5和7目标WFG2-9问题，并对大型复杂构件MOCEO的一些典型焊缝进行了规划。 、NSGA-Ⅱ、MOPSO和MOGWO进行比较。仿真结果表明CrMOCEO能够获得比其他四种算法更好的多目标解，并且CrMOCEO求解的路径在Gazebo物理模型和车间现场进行了测试，结果进一步验证了CrMOCEO算法的有效性。特别是一系列的实验为实际生产提供了更多的解决方案。