A neural network-based algorithm for flow shop scheduling problems under fuzzy environment

Scheduling is a very complex but important problem in the real world environment applications. Production scheduling with the objective of minimising the makespan is an important task in manufacturing systems. For most scheduling problems studied so far, the processing time of each job on each machine has been assumed as a real number. However in real world applications the processing time is often imprecise which means the processing time may vary dynamically because of some human factor or operating faults. This paper considers an n jobs and m machines flow shop scheduling problem of minimising the makespan. In this work fuzzy numbers are used to represent the processing times in the flow shop scheduling. Fuzzy and neural network-based concepts are applied to the flow shop scheduling problems to determine an optimal job sequence with the objective of minimising the makespan. The performance of our proposed hybrid model is compared with the existing methods selected from different papers. Some problems are solved with the present method and it is found suitable and workable in all the cases. A comparison of our present method with the existing methods is also provided in this work.