Coke production scheduling problem: A parallel machine scheduling with batch preprocessings and location-dependent processing times. (April 2019)

• Record Type: Journal Article
• Authors: Liu, Ming; Chu, Feng; He, Junkai; Yang, Dapeng; Chu, Chengbin
• Is Part Of: Computers & operations research. Volume 104(2019)
• Page Start: 37
• Page End: 48
• Publication Date: 2019-04
• Subjects: Coke production -- Batch preprocessings -- Location-dependent processing times -- Scheduling -- Integer programming; Operations research -- Periodicals; Electronic digital computers -- Periodicals; 004.05
• DOI: 10.1016/j.cor.2018.12.002
• Languages: English
• ISSNs: 0305-0548

Abstract

Highlights: A novel coke production scheduling problem is studied, integrating batch preprocessings and location-dependent processing times. An integer programming model is constructed to minimize the makespan. An industrial-based heuristic and a problem-specific genetic algorithm are devised. Experimental results demonstrate the efficiency and effectiveness of the proposed solution methods. Abstract: In this paper, an integer programming model is developed for a newly addressed coke production scheduling problem, in which two typical characteristics are considered: (i) The transportation of raw coal by a vehicle causes a batch preprocessing; (ii) The heating of raw coal by closely located coke ovens may extend the processing times of cokes, under the temperature influence. To the best of our knowledge, such a geographically location-dependent processing time has not been studied. The purpose is to minimize the completion time of the last coke among all ovens, i.e., the makespan. Therefore, the problem of interest can be viewed as a parallel machine makespan minimization scheduling problem, featured with batch preprocessings and location-dependent processing times. For this NP-hard problem, a problem-specific genetic algorithm and a fast heuristic are devised to enhance the computational efficiency. Experimental results on 330 randomly generated instances show the effectiveness and efficiency of the proposed solution methods.