A New Approach for Solving the Flow Shop Scheduling Problem Through Neural Network Technique With Known Breakdown Time and Weights of Jobs

1. Introduction

The flow-shop problem is one of the numerous scheduling problems which is highly used in both manufacturing systems and service industry. It is one of the most popular machine scheduling problems with extensive engineering relevance, representing nearly a quarter of manufacturing systems, assembly lines, and information service facilities in use nowadays. The flow shop the scheduling problems are also classified into two categories namely with and without waiting time in operation intervals. In a flow shop with waiting time the jobs are processed from one machine to the next allowing waiting time in between, whereas in a no wait flow shop system the jobs are processed from one machine to the next without waiting time. Every production system should have a kind of production scheduling, no matter whether it is managed traditionally or have a systematic and scientific approach to the planning in the production system, because it can be guaranteed that production scheduling yields better usage of resources especially the machinery and the manpower giving an edge to stay alive in the fierce market competition among other manufacturing firms. Due to complexity and uncertainty of the machining processes, soft computing techniques are being preferred to physics-based models for predicting the performance of the machining processes and optimizing them. Major soft computing tools applied for this purpose are neural networks, fuzzy sets and genetic algorithms. Any scheduling problem mainly depends upon three important factors namely job transportation time, relative importance of a job over another job and breakdown machine time. In the scheduling problem one of the task in high level synthesis is the problem of determining the order in which the operations in the behavioural description will execute. The number of possible schedules of the flow shop scheduling problem involving n jobs and m machines is. Machine breakdowns in a production schedule may occur on a random basis that makes the well known hard combinatorial problem of flow shop scheduling problems. It is essentially important for the manufacturing firms to improve the performance of production scheduling systems that can address internal uncertainties such as machine breakdown, tool failure and change in processing times.

This paper considers a flow shop scheduling of n-jobs on m-machines with breakdown times. Each machine is capable of processing at most one job at a time, and once a job is started it must proceed to completion. A neural network technique is used to solve the flowshop scheduling problem in this study. The objective of this present paper is to find a sequence or schedule of jobs on machines of a given flow shop problem to minimize the makespan i.e. time required to complete all the jobs. In the present work an artificial neural network model is trained to sort jobs of a flow shop scheduling problem into sequences that will eventually result in minimization of makespan. Presented neural network model consist of two phases for solving the flow shop scheduling problems of n jobs on m machines. In phase-I, a multi layer neural network technique is used to obtain the optimal job sequence. Phases II determine makespan and weighted mean flow time by identifying the breakdown interval (a, b) in which the jobs and processing times are getting affected. Here a modified flow shop scheduling problem is constructed for further calculation by modifying the processing times for affected intervals.