다중목적 최적화 설계 툴, 월라시의 사용자 개입방식의 연구 -최적 솔루션의 선정방식을 중심으로-

Abstract

(Background and Purpose) Recently, parametric-based design support tools have been actively used in architectural design, and optimization design methods using genetic algorithms linked with them have attracted attention. In architectural design, the use of optimization tools automatically derives multiple solutions and provides scientific grounds, supplementing the limitations of efficiency and objectivity. However, these technological advances have created another technological black box due to the lack of understanding of the process and method. It is especially because, the selection of optimal solution cannot be guaranteed due to the randomness of the final selection process, which is the user intervention process. Therefore, this study presents a new selection method based on performance analysis for the process of selecting an optimal solution, focusing on Wallacei, a multi-objective genetic algorithm-based optimization tool in the rhino grasshopper environment, to utilize optimization tools. (Method) This study suggests more sophisticated methods of selecting an optimal solution through a comparative analysis of the selectable solution group derived through K-means clustering linked to Pareto-based optimization and solution groups of various performance standards. The remainder of this study is structured as follows. First, a theoretical review of the optimization was performed. Second, various solutions were explored by setting examples for the optimization simulation. This example aims to derive the optimal arrangement and height of building groups within the site by setting the shading for external space, insolation of the building, and view of the site in the high-density downtown area. Third, the Pareto front group, which is the target group of optimal solution, was used to utilize the optimal solution selection process by comparing the fitness level from the performance point of view, and shape type from the genetic point of view. (Results) Based on the analysis of Pareto front group using K-means clustering, a more objective user selection method for the performance criteria was derived through the analysis of the relationship with the comparison group of the performance criteria. This was achieved by utilizing the intersection group based on the characteristics of each comparison group according to the various directions required for the selection of the optimal solution. (Conclusions) An objective method of performance criteria for the optimal solution selection process, which is the most important user intervention process in multi-objective optimization, is proposed, and improvement points and suggestions for the application of Wallacei are derived. This study contributes to the expansion of new design methods based on scientific objectivity in architectural design.