Abstract: Confronted with the characters of the constraint systems of easy to trap into local optimization, a multi-agent constraint optimization algorithm（DCMACOA ）based on the decomposition and coordination method is proposed. Different from the traditional decomposition and coordination method, the algorithm selects association variables between subsystems as coordination variables, with the help of multi-agent and biological immune evolutionary ideas, each subsystem optimization and system coordination use multi-agent immune optimization method. The optimization search operators include: neighborhood clonal selection, neighborhood competition and neighborhood collaboration. Industrial process and optimization of heat exchanger simulation results show that compared with the traditional decomposition and coordination algorithm, DCMACOA can improve the global and local search performance, and has better global search results.

Key words: immune optimization, decomposition and coordination, multi-Agent immune constrained optimization, heat exchanger system, distillation system

摘要： 针对约束优化系统易陷局部优化的问题，提出了基于分解协调的多Agent约束优化算法（DCMACOA）。对可分系统，不同于传统的分解协调算法，DCMACOA选用各子系统间的关联变量为协调变量，借助于多Agent及生物免疫的进化思想，对各子系统优化及系统协调采用了多Agent免疫优化方法，优化搜索算子主要包括：邻域克隆选择、邻域竞争及邻域协作。工业流程和换热器面积优化仿真实例表明，相比传统的分解协调算法，DCMACOA能改善整体与局部的搜索性能，提高对可分系统的约束优化求解能力，具有较好的全局搜索性能。

关键词: 免疫优化, 分解协调, 多Agent约束优化, 换热器系统, 精馏系统