Development and Application of ANN Model for Worker Assignment into Virtual Cells of Large Sized Configurations

Abstract

This paper presents an extended version of study already undertaken on development of an artificial neural networks (ANNs) model for assigning workforce into virtual cells under virtual cellular manufacturing systems (VCMS) environments. Previously, the same authors have introduced this concept and applied it to virtual cells of two‐cell configuration and the results demonstrated that ANNs could be a worth applying tool for carrying out workforce assignments. In this attempt, three‐cell configurations problems are considered for worker assignment task. Virtual cells are formed under dual resource constraint (DRC) context in which the number of available workers is less than the total number of machines available. Since worker assignment tasks are quite non‐linear and highly dynamic in nature under varying inputs & conditions and, in parallel, ANNs have the ability to model complex relationships between inputs and outputs and find similar patterns effectively, an attempt was earlier made to employ ANNs into the above task. In this paper, the multilayered perceptron with feed forward (MLP‐FF) neural network model has been reused for worker assignment tasks of three‐cell configurations under DRC context and its performance at different time periods has been analyzed. The previously proposed worker assignment model has been reconfigured and cell formation solutions available for three‐cell configuration in the literature are used in combination to generate datasets for training ANNs framework. Finally, results of the study have been presented and discussed.