The objective of most facility layout problems is to minimize material hand
ling cost, which is directly proportional to both the distance between the
machines and the mix, as well as the volume of products handled. The mix an
d volume of products are dependent on the demand patterns, and the distance
is dependent on the layout plan used for the facility. Because it is relat
ively difficult to change the demand patterns, and hence the mix and volume
of products, the primary focus of most designers has been to deal with the
distance attribute of the material handling costs.
The limitations of available horizontal space create a need to explore vert
ical expansion of facilities. This brings up new aspects of vertical materi
al handling and flow that need to be considered in the facility design prob
lem. In this paper, we present a genetic algorithm-based heuristic for gene
rating block layouts for multiple-floor layout problems. This approach prod
uces better solutions than existing simulated annealing-based heuristics fo
r all but one of five multiple-floor test problems available in the literat
ure.