Cellular manufacturing is a practical application of group technology in which functionally dissimilar machines are grouped together to produce a family of parts. The fundamental problem of cellular manufacturing system design is the identification of part families and machine cell compositions. This problem is commonly referred to as the Part Family / Machine Cell (PF/MC) formation problem. Given a set of parts, processing requirements, and available resources, the objective of the PF/MC formation problem is to obtain a satisfactory partition of parts into families and machines into cells. The effectiveness of a cellular manufacturing system is sensitive to fluctuations in the demand for products, the product mix, and the availability of resources. This research offers a multi-period formulation of the PF/MC formation problem. It addresses the dynamic nature of the production environment by considering a multi-period forecast of product mix and resource availability during the formation of part families and machine cells. The goal of the multi-period formulation is to obtain a cellular design that performs well with respect to the design objectives over the entire planning horizon. Design objectives of the multi-period formulation of the PF/MC formation problem are the minimization of intercell material handling costs, the minimization of investment in additional machines, and the minimization of the cost of system reconfiguration over the planning horizon. A mathematical model of the problem is developed and a solution procedure is presented based on a genetic algorithm. The advantages of using a genetic algorithm to solve the multi-period PF/MC formation problem include the ease with which alternate design objectives can be incorporated and the ability to generate alternative system designs. The output of the multi-period PF/MC formation methodology developed in this dissertation is a period by period description of the part families and machine cell compositions. Results are presented of a preliminary investigation of the benefits of using a multi-period model versus one that assumes that the product demand, product mix, and available resources remain constant. In the design problems considered, the multi-period approach to solving the PF/MC formation problem resulted in a cellular design that performed the best overall with respect to the design objectives. / Ph. D.