蚁群-粒子群算法求解多模式资源受限项目调度问题

通过分析多模式项目调度问题的特点,提出一种主、从递阶结构的蚁群粒子群求解算法。算法中,主级为蚁群算法,完成任务模式选择;从级为粒子群算法,完成主级约束下的任务调度。然后,以工期最小和资源均衡分配为目标设计蚂蚁转移概率、模式优选概率和任务优选概率。最后,针对PSPLIB中的测试集对算法主要参数进行优化,并通过与其他算法比较验证了算法的有效性。

Ant colony and particle swarm optimization algorithm-based solution to multi-mode resource-constrained project scheduling problem

A hybrid of ant colony and particle swarm optimization algorithms is proposed to solve the multi-mode resource-constrained project scheduling problems.The hybrid is formulated in a form of hierarchical structure.The ant colony algorithm is performed at the master level to select activity mode,while the particle swarm algorithm is carried out at the slave level to schedule activities without violating the result from the master level.Then,the transfer probabilities of ant,the selective probability for modes and the selective probability for activities are designed in order to distribute resource equably and minimize the makespan of project.Furthermore,the main parameters in the algorithm are optimized with the benchmark problems from PSPLIB.The simulation results and comparison with others' validate the effectiveness of the proposed algorithm.