Solving the Problem of Flow Shop Scheduling by Neural Network Approach

If there is a continuous flow of production jobs for some machines, the problem of flow shop scheduling arises. As mentioned in many researches, the complexities of this problem are of exponential kind; therefore it is necessary to design less complex methods or algorithms for solving it. In this paper, a new solution is presented for this kind of scheduling problem by using the idea of neural networks. In fact, this research is a response to the need for solving large and complex problems of this type by non-classical methods. The purpose of the paper is to create an artificial intelligence for doing this kind of scheduling via the neural network training process. Here, the neural network has been trained by using training data obtained from optimal sequence of solved problems of flow shop scheduling. The trained network can provide a priority which shows the sequence of the job and will be very close to the optimal sequence.