KB-GA-based hybrid system for layout planning of multistory buildings

Layout planning of multistory buildings requires multidisciplinary participation in the decision-making process, which is governed by intuition, creativity, and experience of the designers, and available information. Optimization techniques are rarely applied to the layout planning stage. The present study aims at a hybrid approach by combining knowledge based system and a genetic algorithm (GA)-based optimization technique by employing them for suitable tasks, in order to take specific advantages of respective techniques, i.e., the domain dependent part is isolated and developed as a knowledge-based (KB) system, while the knowledge lean part is handled by the adaptive search techniques. The main issue to be addressed while developing an integrated system is to identify the role of individual approaches and the interactions among them. This paper outlines this concept through an integrated hybrid system GALOP for optimal layout planning of multistory office buildings. A multiobjective GA-based optimization technique produces a set of optimal solutions, from which a designer can select the final design. © 2007 ASCE.