Integration of Dispatch Rules for JSSP: A Learning Approach

Abstract

Dispatching rules have been studied from past decades, and it has concluded that these rules are composing vertebrae for various industrial scheduling applications. Generally, it is a very tedious process to develop incipient dispatching rules in a given atmosphere by indulging and implementing dissimilar models under consideration and also to appraise them through extensive research. For determining effectual dispatching rules, automatically, a pioneering approach is presented. Proposed work addresses job shop scheduling problem (JSSP) NP hard in nature with the objective of minimizing mean flow time. For achieving this, two latest dispatching rules have been introduced. These latest rules will combine the process time and total work content of a job in queue under subsequent process. Additive and alternative approaches have been taken for combining the rules. For evaluating the performance of proposed dispatching rules, a rigorous study has been carried out against S.P.T rule, WINQ rule, S.P.T and WINQ rule, and the other best existing rules in common practice.