Architecture Golf: Building Better

Architecture is often considered more of an art than solving an optimization problem. This results in buildings that are not optimized for minimum building cost, efficient use of material, ideal room layout, etc. Architecture Golf uses gamification to fuel a social process for iterative generation and improvement of solutions.

The Game

In software programming, Code Golf describes a competitive process used to find the shortest possible source code that solves a certain problem. Sometimes, minimizing the length of the source code is not the only goal of the competition. For example, memory consumption or processor time might be constrained. Multiple players compete in finding the best solution. Code Golf may also lead to the development of tools that help achieve the desired goals. Code Golf is a special type of competitive programming.

Architecture Golf applies the principle of Code Golf to building design and construction, aiming to improve the quality and affordability of buildings. It is both a competitive social process and an exploration of technology that can be employed to aid finding optimized solutions. The principle of Architecture Golf can be used for design and construction of whole towns, buildings, room layouts, as well as individual architectural details and construction techniques.

Architecture Golf differes from tenders and obtaining multiple offers for a certain project, as these approaches are lacking the crucial iterative competitive process and a clear definition of goals. However, certain patterns have proven to be suitable (i. e., optimized for specific goals) in a certain area and time, such as Victorian terraced houses. Solutions obtained by Architecture Golf competitions can be generalized to patterns (Alexander et al. 1977) that can be adapted and reused in other projects.

Defining the Competition

First, the project is split into problems that are suitable for an Architecture Golf competition. To do so, it is recommendable to determine independent entities, such as the type of fence of a garden or construction of windows. During the process, additional subproblems may be identified. For each problem an initial list of optimization goals is defined.

Architecture Golf is competitive process rather than a competition where submitted results are compared to determine the winner. Depending on the solutions developed during the process, the list of problems and goals may be updated or even altered. This prevents getting stuck at local optima that ignore additional important factors.

Architecture Golf can be either played as a publicly visible process or in a covert manner. Each of the variants has its advantages and disadvantages. Playing openly allows for cross-inspriation by another player’s good ideas, but it is more prone to narrow the range of presented solutions.

Participants may use or even develop tools during the competition. For example, floor plans might be generated or optimized using heuristics based on constraints such as room sizes, natural lighting, placement of supply lines, maximum distances, material prices, etc. This can potentially lead to the development of parameterized building generators. Even “artificial intelligence” may be used to improve the quality of solutions.

Sample Optimization Goals

Architecture Golf does not specify a certain set of goals, as they may vary depending on time, location, available materials, intended usage, social situation, etc. In order to generate sustainable buildings, the following goals should be considered:

Possible contraints are sufficient noise protection, protection against fire, maximum construction costs, maximum building height, etc.

Further Information