Integrated Scheduling of Tasks and Preventive Maintenance Periods in a Parallel Machine Environment With Single Robot Server

Lotfi Hidri,Khaled Alqahtani,Ahmed Badwelan,Achraf Gazdar

doi:10.1109/access.2021.3081495

Abstract

In this study, the objective of minimizing makespan has been considered for a scheduling problem of identical parallel machines with a single server and unavailability constraints. The unavailability constraints correspond to preventive maintenance periods. In this study, the jobs and the maintenance periods are scheduled simultaneously. This scheduling problem has a wide range of potential application areas in the manufacturing environment. In addition, the studied problem is a challenging one from theoretical point of view, due to its NP-Hardness. To conduct the study, a lower bound ( LB ) for the problem, and three metaheuristics namely Simulated Annealing ( SA ), Tabu Search ( TS ), and Genetic Algorithm ( GA ) have been proposed. The best parameters settings of the proposed algorithms were conducted using pilot runs with a Taguchi design. The algorithms performance has been assessed by using a set of test problems generated randomly. These test problems are based on a literature benchmark. The size of the instances, or number of jobs, were up to 500. Along with the performance analysis of the proposed algorithms, the effect of varying processing times and unavailability periods on the performance of the proposed algorithms is studied. The present work provides strong evidence of the efficiency and the performance of the proposed algorithms.

Highlights

Customer demands are constantly changing and the need to deliver better quality products are among the most important driving factors to improve production system performance
The second scenario corresponds to t0k = ((a + b)∗n/2)/3 = (a + b)∗n/6, which involves that the preventive maintenance begins while the processing of jobs is not yet finished
AND DISCUSSIONS performance of the proposed Simulated Annealing (SA), Tabu Search (TS) and Genetic Algorithm (GA) algorithms is evaluated by conducting computational experiments with 40 problem types of which have been randomly

Summary

INTRODUCTION

Customer demands are constantly changing and the need to deliver better quality products are among the most important driving factors to improve production system performance. This study considers identical parallel machine scheduling problem with a single server and availability constraints on each machine with an aim in minimizing makespan. Approximation methods like Iterative Search ITS, GA, SA, TS and Particle Swarm Optimization, and Harmony Search Algorithm, are generally found to be robust and produce good results Their performance is satisfactory as long as the objectives specified in the Identical Parallel Machine Scheduling environment (IPMS) and unavailability period of the machines. This encourages researchers to apply metaheuristic techniques that support in providing near optimal solutions and reduce computational burden. This study considers identical parallel machine scheduling problem with single server and availability constraints on each machine with the objectives of minimizing makespan. A conclusion summarizing the main findings and presenting future research directions, is presented

LITERATURE REVIEW

PROPOSED METAHEURISTICS

Replacement strategy

PARAMETERS TUNING

RESULTS AND DISCUSSIONS

CONCLUSION