Solving Process Planning, ATC Scheduling and Due-date Assignment Problems Concurrently Using Genetic Algorithm for Weighted Customers

Abstract

Traditionally process planning, scheduling and due-date assignment functions are applied sequentially and separately. Since these three functions effect each other and if we don’t integrate then they will become poor input for downstream and global performance will be poor. In these competitive era we have to be competitive also. Integrating these functions will improve global performance. In this study we investigated the benefit of integration. We tested different integration level. First we looked at unintegrated results and later we integrated due-date assignment with process planning and finally we integrated these two functions with ATC (Apparent Tardiness Cost) Dispatching. In this study we observed that as integration level increases the solution becomes better. Integrating due-date assignment with process planning improves global performance and if we integrate ATC dispatching with these two functions then we get the best performance. In addition to integration level we also compared benefit of search techniques, especially genetic (directed) search. We compared ordinary solutions (initial populations) with results of random search and genetic search (after some certain number of iterations). We tested three shop floors which are small, medium and large shop floor. We observed that search is always better than ordinary solutions and genetic search outperformed random (undirected) search. At this study every customer has weight and they are scheduled by considering weights to improve performance value.