Decision support system for machine re -layout planning
Abstract
Facility layout and re-layout activities have been observed as recurring activities in most manufacturing industries. Because of the uncertainty of product demands and unpredictable consumer's behavior, the re-layout process is almost unavoidable. In a survey we conducted, the frequency of major re-layout projects is about once every 1.7 years. The existing layout models do not embed the re-layout constraints, such as fixed machine locations or the percentage of movable machines. There is a need to develop a model specifically for the re-layout process. The change of layout usually affects the performance of a material handling system. On the other hand, the material handling system limits the throughput performance of a layout. The material handling system should be properly designed according to the layout and the interaction between two systems. We integrated both designs in a decision support system for machine re-layout planning. The total flow distance has been used in most layout models as their objective functions. We evaluate existing performance measures and propose new indices to measure the system throughput. A simple and efficient algorithm, insertion algorithm, is proposed to obtain an initial layout. Then, the layout is improved by a genetic algorithm. A temperature function of the mutation rate is used in the proposed genetic algorithm to avoid being trapped in a local optimum. The performances of proposed models in terms of the throughput rate and the stability are addressed. An analytical model using Markov Chain model is proposed to estimate the traffic congestion level and the number of vehicles required. The computation time in the proposed models is attractive. The decision support system proposed is a real-time model that integrates the Markov Chain model and optimization techniques to support the decision analysis in the facility design, design of material handling systems, throughput evaluation and cost analysis. A simulation model is developed and used to support the IF-THEN analysis. This model provides a structured method for making decisions in facility planning. It allows a mixed sequence of various types of improvement in facility planning. That is not allowed in the traditional facility planning. Moreover, it enhances the method of information exchange between sub-systems of a manufacturing system.
Degree
Ph.D.
Advisors
Tanchoco, Purdue University.
Subject Area
Industrial engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.