This thesis is concerned with the cell formation problem which requires the identification of machine-cells and part-families. One of the well known mathematical programming approach for forming machine cells is the p-median model which utilizes similarity coefficients.
In this thesis, a new similarity measure, adherence coefficient, is proposed. With the adherence coefficient, an extended p-median formulation is presented introducing constraints on the number of machine cells and the number of machines in each cell. A user-controlled algorithm for the cell formation problem is developed based on the extended model.
The performance of the adherence coefficient is examined through a comparative study with some existing solution methods in terms of grouping efficiency. Also, the proposed algorithm is illustrated with several example problems.