Mixed integer linear programming model and an effective algorithm for the bi-objective double-floor corridor allocation problem. (August 2021)

• Record Type: Journal Article
• Authors: Guan, Chao; Zhang, Zeqiang; Gong, Juhua; Liu, Silu
• Is Part Of: Computers & operations research. Volume 132(2021)
• Publication Date: 2021-08
• Subjects: Facility layout design -- Multi-objective optimisation -- Genetic algorithm -- Mixed integer linear programming -- Variable neighbourhood search; Operations research -- Periodicals; Electronic digital computers -- Periodicals; 004.05
• DOI: 10.1016/j.cor.2021.105283
• Languages: English
• ISSNs: 0305-0548

Abstract

Highlights: Mixed integer linear programming model is formulated for double-floor corridor allocation problem. Two objectives of minimizing the overall flow cost and the corridor length objectives are both considered. Comparison of proposed mathematical model versus previous model is executed. A genetic algorithm with a variable neighbourhood search technique is designed. Comparison with other methods indicates that the proposed algorithm achieves highly preferable results. Abstract: The bi-objective double-floor corridor allocation problem (bDFCAP) investigated here explores the effective placement of given departments in a double-floor space to minimise the overall flow cost and the corridor length objectives. Within each floor, departments are arranged in two parallel rows on opposite sides along a central corridor without overlapping. In this study, the bDFCAP is formulated as a mixed integer linear programming model, which has improved the performance over the previous one. Thereafter, a genetic algorithm with a variable neighbourhood search technique is designed and employed to solve the bDFCAP in a more effective manner. This technique is utilized to improve the local search capability by adaptively transforming between a deep-searching strategy and broad-searching strategy, and the superior performance of the proposed method is proven through comparisons with two other algorithms in current literature. Besides, the state-of-the-art lower bounds of several benchmark instances are updated.