Optimum Design of Cellular Manufacturing Systems Using Hopfield Neural Network Approach: the First Step

Cellular manufacturing system (CMS) involves processing a collection of similar parts on a dedicated cluster of dissimilar machines. It has been recognized as one of the solutions to batch production in order to obtain a trade off between the flexibility of a job shop and the efficiency of a flow shop. This paper presents a resolution outline of cells formation problem using an extended similarity coefficient approach introduced by Islam and Sarker [1]. In their paper, they used an optimization mathematical approach for small and medium sizes problems whereas for larger ones, they employed a heuristic resolution approach for cell problems. The goal of the extension that we propose is to allow the simultaneous regrouping of the machines in cells and the parts in families whatever the size of the initial problem. This objective perhaps achieved only by the use of the mathematical programming modelled starting from the similarity coefficients as defined by Islam and Sarker [1]. Moreover, a new approach using neural networks techniques to solve CMS design problem is also introduced.