中文摘要：

为实现资源受限项目调度的多目标优化,通过改进传统蚁群算法,提出适用于多目标优化的多种群蚁群算法.该算法基于串行进度生成机制,每个蚁群具有各自的目标函数、与目标函数相匹配的不同搜索策略以及各自的信息素更新机制.各蚁群独立进行搜索决策,但各蚁群之间存在信息素的相互作用,从而实现加速搜索.针对多目标资源受限项目调度问题设计新的精英策略.在目标规划基础上构造一系列多目标项目调度算例,经系统测试表明,所提出的多种群蚁群算法能够有效优化资源受限项目的资源配置,实现多目标优化.

英文摘要：

A new ant algorithm was proposed to take advantage of multiple ant colonies in order to solve the multi-objective resource-constrained project scheduling problem.The proposed algorithm utilizes the serial schedule generation scheme to construct project schedules stage by stage.Each ant colony has its own objective function,a corresponding searching strategy and pheromone update mechanism designed for the specific function.The ant colony searches for better schedules individually,and meanwhile they share their pheromone information so as to improve the searching efficiency.A new elitist strategy was also designed for the multi-objective problem and integrated into the pheromone update mechanism.A series of multi-objective project scheduling instances were constructed using goal programming.Systematic computational tests showed that the proposed multi-colony ant algorithm can allocate constraint resources effectively to achieve the multi-objective optimization.