最小化总反馈长度的耦合活动排程研究

耦合活动的排程直接影响新产品开发的周期和成本,因而受到了学者和研发管理人员的普遍关注。本文针对最小化总反馈长度这一耦合活动排程常用目标,将遗传算法与局部搜索算法相结合,提出了一种新的混合优化算法,并系统分析了参数对算法性能的影响。然后将算法应用到实际案例和大量随机算例中,实验结果表明混合优化算法较大幅度提高了现有局部搜索算法解的质量;同等情形下,混合优化算法所获得解比单纯运用遗传算法所获得解更好。

A Study on Scheduling Coupled Activities with Minimum Total Feedback Lengths

The scheduling of coupled activities directly affects the time and cost of new product development, and thus has become a focal point of both scholars and design managers. Researchers have shown that finding an activity sequence with minimum total feedback length can significantly reduce development time. However, the feedback length minimization problem is NP-hard and hard to solve. To solve large problem instances in reasonable time, this study combines the genetic algorithm with an existing greedy heuristic, and designs a new hybrid optimization algorithm. We also systematically analyze the effect of parameters, such as crossover rate and mutation rate, on algorithm’s performance, and suggest appropriate parameter values. Experimental results indicate that the hybrid optimization algorithm greatly improves the existing greedy heuristic. Moreover, with same settings, the hybrid optimization algorithm often yields better solutions than the genetic algorithm without the greedy heuristic.