Diseño computacional generativo: pabellón en diseño paramétrico

Abstract

The incursion of the computer into architectural practice has gone from be- ing a mere tool for graphic representation to a relevant instrument for the

resolution of complex design problems. By applying a genetic algorithm, or

search algorithm, we try to emulate the genetic processes of nature (gene- ration, selection, crossover and mutation) to explore, efficiently, a vast de- sign space, which contains all the possible configurations of solutions to the

problem according to the type, limits and number of variables with which the

scheme of the object in question is designed. Since only what can be mea- sured can be optimized, this method does not consider aspects such as the

style or aesthetics of the pavilion and, in addition, being a multi-objective

design problem, it is not possible to find a single, optimal and global solu- tion that equally satisfies all the dimensions that constitute the architectural

project.

The purpose of the research was to explore the use of generative compu- tational design tools and to develop a parametric pavilion prototype to be

placed in the context of the university campus. Three design objectives were set: minimize cast shadow, minimize cutting waste and maximize contained

volume with the least possible weight; and three constraints: obstacle avoi- dance, keep part lengths under a limit and deflections within the allowable

range. A total of 6 optimization objectives and a set of 33 variables with which the

algorithm worked to simulate a group of 50 individuals through 100 genera- tions. As a result, a group of 10 valid and unique configurations were obtai- ned, with a certain level of similarity in their formal expression, to provide a

solution to the architectural design problem posed.