淮安大学学报

针对多焊接机器人系统协同任务分配与路径规划的焊接路径长度均衡与生产周期缩短的要求，将焊接问题模型简化为多焊点问题，在多焊接机器人系统协同规划建模基础上，提出了多目标改进单亲遗传算法（PGA）。改进PGA算法采用两段式染色体编码，融合kmeans算法与最优子结构策略进行种群初始化提供初始解；采用简化的选择算子，避免复杂变异概率的设置；利用个体染色体间的差异，控制不同个体染色体的变异位数。通过Matlab使用TSPLIB数据集，在不同权衡系数下进行实验得到较优的多目标权衡系数，通过多焊接机器人系统进行对比仿真，实验结果显示，改进PGA在迭代效率和求解精度上均有提升。

An improved multi-objective Partheno-Genetic Algorithm(PGA) is proposed to address the collaborative task allocation and path planning problem in multi-robot welding systems, with the specific objectives of balancing welding path lengths and minimizing production cycles. The welding process is modeled as a multi-point problem within a collaborative planning framework. The algorithm employs a dual-segment chromosome encoding scheme and integrates k-means clustering with an optimal substructure strategy to generate high-quality initial populations. A simplified selection operator eliminates the need for complex mutation probability settings, while an adaptive mechanism controls mutation magnitudes based on inter-chromosomal differences. Extensive simulations using TSPLIB datasets in MATLAB identified optimal weighting coefficients for the multi-objective function. Comparative results demonstrate that the improved PGA achieves significant enhancements in both convergence efficiency and solution accuracy compared to conventional approaches.