Distributed-elite local search based on a genetic algorithm for bi-objective job-shop scheduling under time-of-use tariffs
Distributed-elite local search based on a genetic algorithm for bi-objective job-shop scheduling under time-of-use tariffs
复制标题
DOI：
10.1007/s12065-020-00426-4
复制
发表时间：
2020-05
影响因子：
2.6
通讯作者：
B. Kurniawan;Wen Song;W. Weng;S. Fujimura
中科院分区：
文献类型：
--
作者：
B. Kurniawan;Wen Song;W. Weng;S. Fujimura
电力需求的快速增长促使世界各国政府实施能源意识政策，例如使用时间关税。制造业可以通过实施创新的调度系统来实施这些政策，以减少能源消耗。因此，本研究针对分时电价下，具有总加权拖期与电力成本最小化之双目标作业车间排程问题。该问题可以分解为两个子问题，操作排序和启动时间的确定。为了解决这个问题，我们提出了一个分布式精英局部搜索的基础上，遗传算法，使用局部改进策略的基础上分布的精英。具体地，染色体编码使用对应于操作序列和开始时间的两行基因表示。我们提出了一种解码方法，以获得一个时间表，结合操作排序和开始时间。提出了一种降低电力成本的扰动方案。最后，基于精英分布的局部搜索框架被用来指导个体的选择和扰动的确定。利用文献中的基准数据进行的综合数值实验表明，所提出的方法比NSGA-II，MOEA/D和SPEA 2更有效。本文的研究结果对制造业采取分时电价政策具有一定的参考价值。
The rapid growth of electricity demand has led governments around the world to implement energy-conscious policies, such as time-of-use tariffs. The manufacturing sector can embrace these policies by implementing an innovative scheduling system to reduce its energy consumption. Therefore, this study addresses bi-objective job-shop scheduling with total weighted tardiness and electricity cost minimization under time-of-use tariffs. The problem can be decomposed into two sub-problems, operation sequencing and start time determination. To solve this problem, we propose a distributed-elite local search based on a genetic algorithm that uses local improvement strategies based on the distribution of elites. Specifically, chromosome encoding uses two lines of gene representation corresponding to the operation sequence and start time. We propose a decoding method to obtain a schedule that incorporates operation sequencing and start time. A perturbation scheme to reduce electricity costs was developed. Finally, a local search framework based on the distribution of elites is used to guide the selection of individuals and the determination of perturbation. Comprehensive numerical experiments using benchmark data from the literature demonstrate that the proposed method is more effective than NSGA-II, MOEA/D, and SPEA2. The results presented in this work may be useful for the manufacturing sector to adopt the time-of-use tariffs policy.