多目标无等待流水线调度的离散果蝇算法

Discrete Fruit Fly Algorithm for Multi-Objective No-Wait Flow Shop Scheduling Problem

本文提出了一种离散多目标果蝇优化算法,求解以最大完工时间和机床空闲时间最小化为目标的无等待流水线调度问题。与传统的果蝇算法不同,首先,该算法采用基于工序的编码方式,其次,利用GLOVE发生器进行初始化,提高初始解的分散度;最后,利用简单但有效的插入方法在邻域内进化精细搜索,增强算法的全局开发能力。仿真试验表明了所提果蝇算法的有效性和高效性。

This paper presents a Fruit Fly Optimization Algorithm (FOA) for solving the multi-objective no- wait flow shop scheduling problem (MNFSP) with makespan and idle time criteria. Firstly, unlike the traditional FOA, the proposed algorithm applies the job-permutation-based representation. Secondly, initialization method based on the Glove generator has a uniform distribution of the so-lutions. Finally, a simple but effective insert search algorithm is made to improve global exploration. Computational results show that the FOA presented in this paper is very effective and efficient for the MNFSP.