Abstract
The most research on facility layout problems ignored the actual shapes of activity spaces and the aisles between activities. In many cases, the research also ignored the actual shape of building where the activities are to be arranged. In this paper, We present a block based layout technique that applies a genetic algorithm to search for a very good facility layout with horizontal aisles. From the extensive experiments for two different cases with respect to the shape of activity space, it was found that the proposed method generated better layouts than the ones obtained by applying Tam's algorithm in all test problems. The proposed algorithm showed about 10% improvement of performance on the average. We determined the best combination of the reproduction rule and the genetic operators with their probabilities for each test problem through the experiment.