A flexible neural network approach for machine cell formation

The aim of this paper is to develop an artificial neural network approach for solving the generalized machine cell formation problem. The capability of spontaneous generation of the schema constraint satisfaction model is applied in the authors´ approach to find a near-optimal solution for the problem. A similarity coefficient method is used to compute the relationship between machines and between parts for the construction of the neural network. By modifying the relationship between machines, the factor of sequence of operations can be involved. By modifying the relationship between parts, the factor of multiple process plans can be involved. The result shows the authors´ approach is flexible and efficient to satisfy various requirements in machine cell formation