探讨一种新颖的分布式多智能体优化算法及其在求解约束优化问题上的实现方式.通过分析该算法在有限采样数下寻优能力所受到的影响,提出在迭代优化过程中引入补偿采样机制和平滑算子的改进方案,在保留原算法优点的基础上提高了采样过程对决策空间的覆盖度,增强了方法的全局和局部邻域搜索能力.实验结果表明,引入补偿采样和平滑算子后的概率集群优化算法在收敛速度、解质量和稳定性等方面均得到了明显改善. This paper discusses a novel distributed multi-agent optimization algorithm and its implementation in solving constrained optimization problems.By analyzing the influence of this algorithm on the optimization ability under the limited number of samples, it is proposed to introduce compensation sampling in the iterative optimization process Mechanism and smoothing operator, the coverage of decision-making space is improved based on the advantages of the original algorithm and the global and local neighborhood search ability of the method is improved.The experimental results show that the introduction of the compensation sampling and smoothing algorithm Sub-cluster probability clustering algorithm has been significantly improved in terms of convergence speed, solution quality and stability.