Architectural form exploration by soft computing: The case of post-disaster shelter

Soft Computing is composed of several computing paradigms that allows the incorporation of human knowledge and judgement into the digital environment. These methods are known to be permissive to imprecise concepts and partial truth, thus becoming quite applicable to architectural design problems. The major strengths of Fuzzy Logic and Neural Networks result in a learning system called Fuzzy Neural Tree which is transparent yet capable of dealing with medium complexity problems. In this thesis, the potentials of applying this newly developing method for an architectural design problem is searched and it is utilized as the main tool for a multi-optimization problem within a shelter settlement. Sheltering is one of the most crucial needs of humanity and every year many people are compelled to leave their homes due to disasters. Post-disaster shelters become a more critical subject of architectural design as the increasing number of natural disasters taking place each year, result in a larger number of people in need of urgent sheltering. When designing temporary shelters, it is a necessity to deal with the provision of materials, low-cost production and the time limit in the emergency as well as the needs of the users and the qualities of the space. Although computational approaches might lead much more efficient and optimal design solutions, they have been utilized in very few examples. For that reason in this thesis, computational tools are used throughout the conceptual design process. The study aims to create different configurations that satisfy the requirements of a post-disaster shelter. Towards this aim, Modularity, Flexibility, Upgrade Potential, Digital Fabrication, Construction without External Help, Small Carbon Footprint and Computational Design Process are determined as the aspects to look for in a shelter design through the literature review. With the guidance of these aspects, previous shelter examples are analyzed and an algorithm that defines the growth pattern is constructed accordingly. A digital model is produced aiming to alter this algorithm as the pattern grows. These alterations are applied by Multi-Objective Genetic Algorithms and they perform towards satisfying the evaluation criteria within the Fuzzy Neural Tree. The struggle to define intangible architectural concepts for the evaluation of possible settlement scenarios holds potential for the further use of this particular method for architectural problems. Different from the previous studies, this thesis aims to focus on the design of a temporary shelter that can respond to different user types and disaster scenarios through mass customization, using an alternative methodology. While serving as a temporary space, the design outcomes are expected to create a more neighborhood-like pattern with a stronger sense of community for the users compared to the previous examples.