Date of Award

1990

Degree Type

Dissertation

Degree Name

Ph.D.

Department

Industrial and Manufacturing Systems Engineering

First Advisor

Monforton, G. R.,

Keywords

Engineering, Industrial.

Rights

CC BY-NC-ND 4.0

Abstract

One of the problems in the design of cellular manufacturing systems is the cell formation, which is essentially the identification of part families and machine groups. The objective of this research is to develop mathematical models to address the issues for cell formation encountered in cell design and suggest efficient solution methodologies to solve the models developed. It is assumed that a part can be produced through one or more process plans. Each operation in a process plan can be performed on alternate machines. Thus, for each process plan we have a number of production plans depending on the machines selected for each operation. It is also assumed that the demand for a part could be split and can be produced in more than one cell. The plans identified to produce the same part in different cells could be different. The cell formation problem, in addition to identifying part families and machine groups is to specify the plans selected for each part, quantity to be produced through the plans selected, machine type to perform each operation in the plans, total number of machines required, machines to be relocated, machines to be replaced and parts and machines to be selected for cellularization considering demand, time, material handling and resource constraints. Some pertinent objectives to be considered are minimization of investment, operating cost, machine relocation cost, material handling cost and maximization of output. Consideration of physical limitations such as upper bound on cell size, machine capacity, material handling capacity etc., should also be incorporated into the cell design process. Accordingly, a number of mathematical models are developed to provide a framework for discussing the issues related to design of cellular manufacturing systems. All the models developed are large scale linear and mixed integer programs. For the solution of the linear and relaxed mixed integer programming models, an efficient column generation scheme is presented. In the problems under consideration, column generation is achieved by solving simple assignment problems. A branch and bound scheme on the integer variables leads to an optimal solution for the mixed integer programs. Each node in the branch and bound tree represents a solution to an augmented continuous problem with additional constraints on the integer variables. These additional constraints are easily incorporated without increasing the size of the problem by the bounded variables procedure. A number of illustrative examples are solved to illustrate the application of the solution methodology. Computational experience is provided for a few test problems and statistics on number of nodes, number of plans generated, number of pivot operations, number of assignment problems solved and time for execution are included. (Abstract shortened by UMI.)Dept. of Industrial and Manufacturing Systems Engineering. Paper copy at Leddy Library: Theses & Major Papers - Basement, West Bldg. / Call Number: Thesis1990 .R347. Source: Dissertation Abstracts International, Volume: 52-11, Section: B, page: 6021. Chairperson: G. R. Monforton. Thesis (Ph.D.)--University of Windsor (Canada), 1990.

Share

COinS