Optimization of spatial organization in architectural plan design using particle swarm optimization algorithm

Today, due to the wide range of variables affecting architectural design, the computer is used as a tool in interaction with the design process to find optimal and high-performance solutions. It is necessary to investigate these methods due to their limitations and the complexity of space planning regarding the effective number of parameters. Improving the design quality and construction of architectural works is a common concern in developing countries, a step that can promote contemporary Iranian architecture. Furthermore, one of the design and construction characteristics in these developing countries is the tendency to improve the quality of architecture by using new technologies.

Space planning in architecture is one of the most practical and complex issues in architectural design and is considered one of the most challenging issues in recent research. The use of new technology-dependent methods in design, especially with an emphasis on using evolutionary algorithms as a solution, has been considered in the present research. Therefore, this study investigates the application and use of these algorithms as a solution for design optimization. The research questions of this study are:- Concerning the application of optimization algorithms in space planning design, which algorithms have been used as basic or complementary algorithms? How frequently have the Swarm Intelligence Algorithms, especially particle swarm optimization algorithms, been used in this regard? - What are the possibilities and limitations in spatial organization design in architecture using genetic algorithms compared to the particle swarm optimization algorithm as the two main evolutionary algorithms?- What is the implementation process and application of the particle swarm optimization algorithm in spatial organization design in architecture?In order to answer the research questions, 35 types of research that have used optimization algorithms in architectural spatial planning design are collected. Then, the content analysis method was used to extract all the variables used in architectural plan optimization. In the same way, the basic algorithm and the complementary algorithms, if any, were extracted.Based on the study of specific sources regarding the evolutionary optimization algorithm from the available bibliographic resources and the analysis of planning requirements and architectural space design, the genetic algorithm and swarm particle optimization algorithm were explained in designing the architectural spatial organization. The speed and quality of these two algorithms in investigating the research problem have been scrutinized based on software capacities in algorithm implementation and possibilities and limitations in using analytical methods for designing architectural plans. After explaining the objectives and numerical criteria, the particle swarm optimization algorithm using Microsoft Visual Studio programming software and .NET programming platform in C #, with WinodowsForm graphical user interface, was used to monitor the algorithm developing process and its results further.

This research introduces the features of metaheuristic algorithms and presents various optimization algorithms, including deterministic, heuristic, and metaheuristic algorithms. Also, the application of optimization algorithms in architecture is explained. The background of the evolutionary optimization algorithms and swarm particle optimization in the architectural plan design were also examined. After comparing the performance of the genetic algorithm and particle swarm optimization algorithm in optimizing the spatial organization, the swarm optimization algorithm structure was introduced. Finally, the application of this algorithm in plan design was studied and explained. Accordingly, plan design algorithms were classified into three stages: In the first stage, the goals, criteria, and constraints affecting the architectural spatial organization were determined, and they were classified into (1) The primary criteria and constraints, (2) The designers’ criteria and constraints, and (3) The contacts’ (client and users) criteria and constraints. In the numeric stage, the criteria of land boundary, list of spaces, permissible aspect ratios, permissible dimensions, total area, space interference, space adjacency, daylighting, and verification of the spaces were quantified. Then, the algorithm is determined based on the particle swarm optimization algorithm in two steps. Finally, the implementation platform of the algorithm is determined.

The difficulty of working with programming languages, software skills, and the software complexity due to inadequacy to combine various scientific fields has made using programming languages uncommon for designers to control the design parameters. This study compared the performance of two genetic algorithms and particle swarm optimization as the representatives of the two main groups of evolutionary algorithms in a base problem. It was shown that the particle swarm optimization algorithm converges faster and has a higher quality to optimize the plan regarding the parameters affecting the plan design. Implementation of an operational solution to optimize the spatial organization of the plan with emphasis on the affecting parameters in the formation of architectural plans was proposed in a three-step process using this algorithm. The problem objectives were examined in three related areas. The quantification process and the final model implementation were completed based on the particle swarm optimization algorithm in the .NET programming platform, along with a graphical interface as user interfaces for architects to understand the implementation process better. In addition to achieving optimal plans, future research interests in this field were also introduced. Regarding the specific entity of architecture, the existing computer software is insufficient for implementing the algorithm, providing visual and operational efficiency, and needs to be developed and customized for broader application in various fields of design.