【摘要】 软件开发项目调度需考虑员工薪水、技能等多个因素,是一种资源有限型项目调度问题,复杂度较高,对此提出一种基于大小蚁群与超立方体的调度算法,优化了项目的工期与总成本。根据项目子任务的不同优先级建立有向图结构,全局最优蚂蚁与单轮迭代最优蚂蚁交替更新路径信息素,提高样本多样性;利用超立方体框架简化大小蚁群算法,并为蚁群设置两个启发信息,提高收敛速度。实验结果表明,本算法的优化效果与求解成功率均优于同类型算法。更多还原

【Abstract】 Software project scheduling needs to consider the salary of employees, skills etc. It is a very complex resource constraint problem, an algorithm based on max-min ant colony and hyper cube is proposed for solving this problem, and the project duration and total cost are optimized by the proposed algorithm. The directional graph is constructed based on the sub-tasks which have different priority, the pheromone trails of the path are updated by the global optimal ant and iterated optimal ant alternately to enhance the diversity of the individuals. The hyper cube is used to simplify the max-min ant colony algorithm, and two heuristic information are set for ant colony to speed up the converge. Experiment results show that the optimal results and hit rates of the proposed algorithm are both better than the other algorithms.更多还原