Architectural design thinking has undergone a profound transformation, evolving from traditional additive drawing techniques toward highly computational and algorithmic methodologies. Traditional additive logic builds complexity incrementally by layering and overlapping elements based on fixed conventions and typologies, which limits responsiveness and innovation due to the lack of component interrelation and inability to simulate real-world forces. Associative design logic, introduced with digital parametric tools such as Grasshopper, allows geometric components to dynamically respond to changes via parameters and constraints. However, the most advanced design paradigm today is algorithmic logic—where architecture is conceptualized as a system of instructions governed by algorithms, enabling the generation of multiple spatial outcomes, behavioral simulations, and optimization processes.

This project applies algorithmic design principles through abductive reasoning inspired by video game environments, where procedural generation, agent-based simulation, and immersive digital worlds converge. The design framework draws on the Wave Function Collapse (WFC) algorithm, commonly used in games like Minecraft and Cities: Skylines to generate complex, rule-based spatial forms procedurally. Spatial validation is conducted through AI agents modeled after human behavior using psychological Big Five personality traits, allowing realistic simulation of occupant interaction within generated spaces.

Responding to contemporary challenges—exacerbated by the COVID-19 pandemic and climate disruptions—this project proposes a dual-presence IT Institute physically situated in Purbachal, Bangladesh, and virtually realized in the metaverse. The metaverse serves as an immersive digital extension of the physical campus, designed with UX/UI spatial interface elements that foster meaningful social interaction, navigation, and learning experiences beyond traditional video conferencing tools.

The primary goals is to demonstrate the use of algorithmic design logic to generate contextually responsive and adaptable architectural environments. To validate spatial configurations through behavioral AI simulations for design optimization. Explore the architect’s emerging role as a designer of hybrid physical-digital environments, integrating metaverse spatial interfaces to enhance resilience, accessibility, and social engagement. Develop a prototype for educational infrastructure that transcends physical limitations and anticipates future disruptions by seamlessly bridging real and virtual spaces.

This project explores an innovative architectural design process that leverages advanced computational tools inspired by video game technologies to create a dynamic, user-responsive educational environment. It focuses on integrating generative design, AI simulation, and digital twin implementation to reimagine the conception, evaluation, and experience of an IT Institute with a dual physical and virtual presence.

At its core, the project employs a generative design approach using the Wave Function Collapse (WFC) algorithm, originally developed for game development and adapted here via the Monoceros/Grasshopper plugin. This algorithm facilitates the rule-based procedural generation of architectural forms, moving beyond traditional additive methods to explore a vast range of design possibilities. The primary functional unit—the classroom module of 25’ x 25’, designed ergonomically for 30 students—serves as the basic building block. Using predefined spatial rules categorized as typed-in, explicit, and indifferent, over 70 design iterations have been generated. These iterations include innovative spatial configurations such as dynamically shifting corridors that break monotony by opening onto green spaces and gathering areas, improving circulation and social interaction while integrating natural elements like sunlight and ventilation.

To rigorously evaluate these generated designs, the project integrates AI-driven behavioral simulation using the Unity game engine. Here, AI agents modeled after Non-Playable Characters (NPCs) and programmed with the Big Five personality traits—Extraversion, Openness, Conscientiousness, Agreeableness, and Neuroticism—simulate real-world user behavior. These agents reflect the diverse student body and provide detailed movement data, which is visualized through heatmaps and interaction graphs. This simulation identifies how spatial configurations affect circulation, engagement, and social dynamics, allowing data-driven selection of the most contextually appropriate design.

Building on these insights, the project realizes a digital twin of the physical campus in the Metaverse, comprising approximately 48% of the institute’s structure, including academic, administrative, and recreational blocks. This immersive virtual campus merges augmented reality (AR), virtual reality (VR), blockchain, and AI technologies to create an interconnected hybrid environment, enhancing accessibility and continuity of education amid disruptions such as pandemics or extreme weather events. Inspired by emerging architectural theories positioning architects as 3D UX/UI designers of virtual spaces, the digital twin incorporates familiar real-world spatial elements enriched with interactive navigation tools, information displays, and communication zones. This design approach aims to deepen user engagement by leveraging emotional connections tied to physical space memory.

The project’s methodology embodies a coherent algorithmic design framework, avoiding the pitfalls of traditional additive sketching followed by disjointed computational simulations. Instead, it abductively draws from the embedded algorithmic logic in video games specifically procedural geometry, AI agent-based validation, and metaverse spatial interfaces to establish a unified design process that meaningfully rethinks architectural practice.

The main structural system is constructed using cast-in-place reinforced concrete, chosen for its durability, load-bearing capacity, and thermal mass—ideal for the hot-humid climate of Bangladesh. Concrete allows flexibility in forming modular units while ensuring structural integrity over long spans. It also provides resistance against monsoon-driven weather and contributes to acoustic insulation in academic spaces.

The generative design enables early-stage visualization of multiple layout iterations, allowing construction planning to be integrated with design decisions. This supports phased construction and modular prefabrication schedules, reducing overall project timelines and site disruptions. Simulation of occupant flow and interactions informs placement of critical access points and emergency exits to comply with safety standards. Using the Wave Function Collapse (WFC) algorithm, modules are spatially arranged under predefined rules to optimize adjacency and circulation. This modular grid-based layout supports prefabrication methods and reduces construction waste by standardizing components and limiting ad hoc on-site adjustments. The algorithm also dynamically designs corridors that incorporate direction changes every 15–20 meters to create visual breaks, improve natural ventilation, and introduce green pockets—enhancing both construction feasibility and occupant comfort.

A virtual twin of the physical campus is developed for the metaverse, which can also serve as a construction monitoring tool. Real-time updates from the site can be integrated into the digital model to track progress, foresee clashes, and coordinate logistics remotely. This enhances communication among architects, engineers, and contractors, enabling proactive problem-solving and minimizing delays.

The modular design, combined with algorithmic spatial optimization, supports adaptability to future needs and reduces material wastage. Incorporating natural ventilation and daylighting decreases reliance on mechanical systems, promoting energy efficiency.