A Multiobjective Evolutionary Algorithm Based on Decomposition for Hybrid Flowshop Green Scheduling Problem

Biao Zhang,Quan-ke Pan,Liang Gao,Xin-yu Li,Lei-lei Meng,Kun-kun Peng

DOI: https://doi.org/10.1016/j.cie.2019.07.036

IF: 7.18

2019-01-01

Computers & Industrial Engineering

Abstract:Energy saving has attracted growing attention due to the advent of sustainable manufacturing. By this motivation, this paper studies a hybrid flowshop green scheduling problem (HFGSP) with variable machine processing speeds. A multi-objective optimization model with the objectives of minimizing the makespan and total energy consumption is developed. To solve this complex problem, a multiobjective discrete artificial bee colony algorithm (MDABC) based on decomposition is suggested. In VND-based employed bee phase, the variable neighborhood descent (VND) with five designed neighborhood is employed to each subproblem to realize their self-evolution. In the collaborative onlooker bee phase, the promising subproblems selected by the order preference technique according to their similarity to an ideal solution (TOPSIS) is evolved by collaborating with the other neighboring subproblems. Particularly, a dynamic neighborhood strategy is developed to define the neighborhood relationship to retain the population diversity. In the solution exchange-based scout bee phase, a solution exchange strategy is developed to enhance the algorithm efficiency and enable the solutions to be exploited in different directions. Moreover, according to the problem-specific characteristics, encoding and decoding methodologies are developed to represent the solution space, and several definitions are proposed to implement objective normalization, and an energy saving procedure is designed to reduce the energy consumption. Through comprehensive computational comparisons and statistical analysis, the developed strategies and MDABC shows highly effective performance.