Abstract
An important element in the integration of the fourth industrial revolution is the development of efficient algorithms to deal with dynamic scheduling problems. In dynamic scheduling, jobs can be admitted during the execution of a given schedule, which necessitates appropriately planned rescheduling decisions for maintaining a high level of performance. In this paper, a dynamic case of the multiprocessor open shop scheduling problem is addressed. This problem appears in different contexts, particularly those involving diagnostic operations in maintenance and health care industries. Two objectives are considered simultaneously—the minimization of the makespan and the minimization of the mean weighted flow time. The former objective aims to sustain efficient utilization of the available resources, while the latter objective helps in maintaining a high customer satisfaction level. An exact algorithm is presented for generating optimal Pareto front solutions. Despite the fact that the studied problem is NP-hard for both objectives, the presented algorithm can be used to solve small instances. This is demonstrated through computational experiments on a testbed of 30 randomly generated instances. The presented algorithm can also be used to generate approximate Pareto front solutions in case computational time needed to find proven optimal solutions for generated sub-problems is found to be excessive. Furthermore, computational results are used to investigate the characteristics of the optimal Pareto front of the studied problem. Accordingly, some insights for future metaheuristic developments are drawn.
Highlights
Scheduling problems have several different structures and exist in a wide variety of real-life applications [1]
The objective of this paper is to develop an exact algorithm for obtaining optimal Pareto front solutions for the DMOSP
In the DMOSP structure, there is no restriction that a job has to visit all workstations
Summary
Scheduling problems have several different structures and exist in a wide variety of real-life applications [1]. Despite the fact that most scheduling problems belong the NP-hard class, there is a demand for exact algorithms as they can be used for solving small to intermediate size instances. They provide a basis for developing and assessing approximate algorithms. The multiprocessor open shop scheduling problem (MOSP) is commonly encountered in maintenance and health care diagnostic systems. These systems are composed of a set of workstations that are available to process a set of jobs. These machines are not necessarily identical, meaning that the processing times
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
