| This dissertation develops a new global optimization approach for solving the integrated facility design problem. The optimization of the integrated facility design involves both the optimization of the facility layout and material handling system. Both optimization problems have been studied in the literature. The facility layout problem has been approached using a number of different methods: activity interrelationship approaches, weighing systems, and proximity matrices, for example, and commercial software products such as CORELAP, ALDEP, and CRAFT. Although the use of these methods and products may generate optimal solutions to the facility layout problem, they create solutions that are locally optimal and independent of the material handling optimization problem.; In optimizing material handling systems, many times dynamic simulation, transportation algorithms and costing/space minimization approaches are used as typical methods. The solutions to the optimal facility layout and optimal material handling system are said to be independent because both problems are most often solved as two separate sequential problems and not as an integrated global design problem.; The present dissertation proposes a new global optimization approach, LAMA (LAyout and MAterial handling), for solving the integrated facility design problem. This is accomplished by incorporating a static simulation program for generating alternative facility design layouts, with a dynamic simulation program testing alternative material handling systems under realistic, stochastically characterized operating conditions. Global optimization of the integrated problem is achieved with the implementation of an evolutionary program. LAMA's performance is evaluated by comparing its results against those of three previously published optimization methods: a heuristic developed by Khare et al. (1988a), FLOAT and HOT developed by Imam and Mir in 1993 and 2001 respectively. The results show that LAMA found the minimum total facility design cost. In addition, LAMA was the only method that reported the optimal number of transporters required by the system. Another advantage that LAMA has over the other methods is that it has the capability to search for the optimal solution at a given user-defined percentage of improvement. |
