Accelerating the solving process of optimal layout design using genetic algorithm

Abstract

The layout design optimization is a complicated process of an architectural design which is concerned with finding feasible locations and size of rooms that meet design requirement and design preference. This paper formulates the optimal layout design as multi-objective mixed integer programming model using the binary variables and branch & bound technique to determine the best location and size of a group of interrelated rectangular rooms by placing a representative point at the center of the room. Although solving the layout problems using MIP model is easy to formulate and adapt for meeting architectural requirements, the number of iterations to find the optimal solution is still influenced by the number of rooms. For this reason, we decrease the number of iterations by accelerating branch and bound process. The genetic algorithm has been adopted to find a candidate sequence of branching variables which helps reducing the search tree. From the empirical test, we found that the iterations can be reduced significantly.