Facility layout optimization using tabu search

The search for high quality facility layouts can be viewed as combinatorial optimization problems arising in a wide variety of spatial planning contexts. We examine here the problem of optimizing a layout of facilities in a supplied space that can include unallowed areas. The layout space and facility areas are rectangular and limited by aspect ratios. Considering the problem complexity, we propose a tabu search heuristic for the problem. First, an initial layout is generated using a clustering procedure. Considering n facilities, the representation is given by a binary tree with solution space size n!. The tabu search is then applied twofold, using the initial layout and using a random generated layout that maintains only the tree structure of the first layout. Quality results for both cases are obtained for problems of the literature.