A property-based hybrid genetic algorithm and tabu search for solving order acceptance and scheduling problem with trapezoidal penalty membership function. (15th May 2023)

• Record Type: Journal Article
• Authors: Zhao, Ziye; Chen, Xiaohui; An, Youjun; Li, Yinghe; Gao, Kaizhou
• Is Part Of: Expert systems with applications. Volume 218(2023)
• Publication Date: 2023-05-15
• Subjects: Order acceptance and scheduling -- Trapezoidal earliness/tardiness penalty membership function -- Common due window -- Property-based hybrid algorithm; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2023.119598
• Languages: English
• ISSNs: 0957-4174

Abstract

Abstract: To improve the profitability and customer satisfaction, in make-to-order environment, manufacturers need to simultaneously consider which orders should be accepted and how to arrange these accepted orders for production, i.e., order acceptance and scheduling (OAS) problem. In practice, orders are often expected to be completed in a certain time interval, but the existing OAS-related studies mainly focus on the hard due date. Therefore, this study is dedicated to solving a single-machine OAS problem that considers the earliness / tardiness penalty under the common due window to maximize the total net profit (TNP), and the main innovative works are described as follows: (1) a trapezoidal earliness/tardiness penalty membership function under the common due window is designed, and a mathematical model is established to characterize the concerned problem; (2) six problem-specific properties are derived for determining which orders to be accepted or rejected, arranging the processing sequence and deciding the start processing time for these accepted orders; and (3) an effective property-based hybrid algorithm (GATS-SSRIR) is designed to deal with the concerned problem, which hybridizes the genetic algorithm, tabu search and six problem-specific properties. In numerical experiments, Taguchi method is first employed to optimize the parameter setting of GATS-SSRIR under different initialization methods. Second, the effectiveness of the proposed problem-specific properties is verified by comparing with other strategies, and the average improvement is 45.3%. Next, the superiority of the proposed GATS-SSRIR algorithm is demonstrated by algorithm comparison, and the average improvement is 67.3%. Finally, a sensitivity analysis on the common due window is performed. To sum up, the proposed problem properties and GATS-SSRIR algorithm are efficient and benefit, and the length of due window has a significant impact on the TNP. Highlights: An order acceptance and scheduling problem with common due window is studied. A trapezoidal earliness/tardiness penalty membership function is designed. Six problem-specific properties related to the features of the problem are proved. An effective property-based hybrid algorithm is designed. A greedy selection method based on the similarity of individuals is designed.