The Journal of Science and Engineering Applications (JSEA)

Abstract

The scheduling of parallel machines has attracted significant attention due to its wide utilization in the fields of manufacturing and data processing. This article presents an in-depth analysis of Uniform Parallel Machine Scheduling problems (UPMS), considering sequence-dependent job characteristics with release dates are imposed. It focuses on efficiently scheduling N jobs across M machines to achieve minimal overall performance penalties that integrates total flow time, tardiness, earliness, and late work. From a practical's and point it is important since most real manufacturing settings involve multiple machines within each workstation, however, efficient solution methods are still scarce, and the optimal scheduling algorithms currently available for multiprocessor systems are limited in their ability to address large scale problems. As a result, practitioners tend to adopt methods based on dominance properties to guide solution selection. This leads to key questions: Are current schedules sufficiently efficient? Do dominance properties effectively support optimal decision-making in multi-objective UPMS problems? How close are the solutions to optimality? This paper focuses on the NP-hard problem of scheduling n jobs on uniform parallel machines and proposes a dominance-based approach.  The method is tested on a wide range of a large set of randomly generated instances.