ABSTRACT
The optimum solution for these facility layout problems is not only controlled by numerical function, but more depends on the accepted baseline of the application of site and relevant requirements. Therefore, the solution for each single layout problem should not be a single solution with the optimum result based on the ratio of each function department and its weight value. Most of the research on facility layout utilizes the classical concept by either the Quadratic Assignment Problem (QAP) or a large-scale mixed-integer programming (MLP) problem (Montreuil 1990). Whereas nonlinear programming (NLP) formulations have been solved by numerical methods (Tam and Li 1991; van Camp et al. 1992), by simulated annealing (Tam 1992) or by genetic algorithm approaches (Tate and Smith 1995), mixed-integer programming (MIP) formulations have been solved by ad hoc interactive designer reasoning (Montreuil and Ratliff 1989) or by reducing the MIP to a linear programming optimization problem either by qualitative reasoning (Banerjee et al. 1992) or, once again, by ad hoc interactive designer reasoning (Montreuil et al. 1993) and by a genetic approach (Banerjee and Zhou 1995). Although integer and noninteger problems have solved complicated layout problems that are two-dimensional with ow and capital consideration, particular situations and single case problems may have to be evaluated in other ways.
