Specifying Optimal Number of Cells Using a Modified Dissimilarity Measure Based on AVV Approach Rami H. Fouad 1 and Adnan M. Mukattash 1 1 Industrial Engineering Department, Hashemite University, Jordan ABSTRACT The first step in Cellular Manufacturing (CM) is manufacturing cell formation (CF), i.e., grouping parts with similar design features or processing requirements into families and allocating associated machines. Many similarity and dissimilarity measures have been published in the literature to solve CF problem. The main step of theses measures is to find the similarity and/or dissimilarity between the machines and/or the parts. Some of theses measures have poor performance. In this paper one of the well- known dissimilarity measures is analyzed and tested against a case study taken from the literature and the test results show that this measures has some deficiencies in finding the optimal number of cells( the cell which has minimum sum of voids and/or exceptions). To avoid these deficiencies a modified dissimilarity measure has been developed in this paper for solving cell formation problem in cellular manufacturing systems. The modified measure is based on AVV approach. The same case study has been tested using our modified measure, and the result showed that our modified measure is superior in finding the optimal number of cells. Keywords: Cellular Manufacturing Systems, Cell Formation, Dissimilarity coefficients, Optimal Solution, Optimal number of Cells, Average Voids Value (AVV). Mathematics Subject Classification: 90B30, 90B90 1. INTRODUCTION Group Technology (GT) is a management philosophy about efficient problem-solving based on the knowledge about groups [1]. Cellular manufacturing (CM) is becoming more used for organizing batch production systems. It is based on the philosophy of Group Technology (GT) that identifies and exploits the similarities among attributes of a set of objectives [2]. Part-family formation and machine-cell formation are considered two major problems associated with CM. The objective of CF is to identify the part families and machine groups by rearranging the initial incidence matrix in a block diagonal form, with minimum inter cell movements [3]. The ideal situation is to rearrange the initial incidence matrix in such a way that all the ones are in the diagonal blocks and all the zeros, off the diagonal blocks [4]. However, this situation is rarely accomplished in practice. Therefore, the most desirable solution of cellular manufacturing systems is International Journal of Applied Mathematics & Statistics, Int. J. Appl. Math. Stat.; Vol. 28; Issue No. 4; Year 2012, ISSN 0973-1377 (Print), ISSN 0973-7545 (Online) Copyright © 2011-12 by IJAMAS, CESER Publications www.ceser.in/ijamas.html www.ceserp.com/cp-jour www.ceserpublications.com