Computation in Space Planning

What is automation in building programming? Its what happens after the building is done. Generative Design has been used to automate all kinds of the physical aspects of buildings before its made: doorways, facades, beams layouts etc. Building form is adjusted to optimize for daylighting, material costs, structure, and other factors. But how are designers optimizing buildings after they are built?

Automation is also being used to change existing buildings by automating how program is laid out within the building floorplates. We’ll start with how space planning problems are generally setup and then look at a case study of space planning. Space Planning is the process of organizing furniture and office functions to work effectively together while using space efficiently. Consider building codes and regulations, lighting, teaming requirements, inter-communication, and storage to make the best use of available space.

Space planning computational models start by breaking down a set space into a grid of voxels or individual cells. They are then serialized with different space types. The cells are filled in different configurations and then tested for how well they behave based on a numerical factors. This could be optimal adjacencies, program sizes being sufficient, programs being aligned in a proper sequence etc.

Another process works by sub sectioning spaces sequentially and filling them with program as they are made.

Softwares like Protofit or Testfit automate floorplan layouts in existing buildings replacing architects during the test-fitting process. Given a text list of what type of program a tenant needs — 30 workstations, 22 conference seats, etc, protofit will populate an empty floor plan with the desired program (optimizing for building codes, adjacencies and cost). This is possible by taking advantage of a few simple geometric rules.

Left: Protofit — test fitting furniture, Testfit — test fitting apartment units
  1. Simple fit search — starts from predefined furniture layout sets and then starts from the largest layout, tries to fit it into a space and pairs down the furniture until it finds a layout that fits
  2. This algorithm arrays objects based on spacing and setback rules and then removes objects that are obstructed by immovable objects like columns and walls. A series of layouts are generated this way but the layout that yields the greatest number of desks and some aesthetic rules about the preferred layout wins
  3. Packing attempts to pack a number of objects into a space by continually nudging their positions. The algorithm starts with an initial guess such as a hex grid and then works from their while avoiding obstacles
  4. Constraint/ cost search optimizes the placement of doors and furniture based on constraints around the objects and circulation and then weighs the least-cost solution.
  5. After an initial layout is made, the layout can be further adjusted as the layouts are broken down into spatial chunks which can be changed.
  6. This simultaneously generates a 3d model which can be navigated like a boring video game. It’s much easier for tenants, landlords and brokers to navigate this 3d environment than a floorplan.

Sequence in Space Planning

Industrial designers and space planners have been keen to optimize spaces around the functions, routines and lives of humans. The ergonomics of objects and space are sculpted to work around the size and reaches of the human body. Lillian Evelyn Moller Gilbreth was an American psychologist, industrial engineer, consultant, and educator who was an early pioneer in understanding the relation of the body, the sequence of tasks performed and the space a body interacts with. In her time-and-motion studies she charted the sequence of everyday activities like cooking. She worked to create optimal kitchen layouts that are still a staple of kitchens today.

The kitchen work triangle is a concept used to determine efficient kitchen layouts that are functional. The primary tasks in a kitchen are carried out between the stove, the sink and the refrigerator. These three points and the distances between them, make up what is called the work triangle. The idea is that when placed in a triangle of a particular distance conducive to the reach of the body, the use of the kitchen is more efficient.

These same principles of optimizing distances between spaces can be considered in more complex sets of routine relationships. Take for example hospitals. Hospitals are highly templated because patients must be processed through structured procedures. Equipment is highly specialized. Some equipment is mobile while others are not, however most are shared between patients and so their position in relation to procedures is highly important. Additionally, reducing fatigue among doctors is a crucial part of ensuring the safety of patients.

Hospital templates require very particular sets of spatial adjacencies

Today the ability to sync calendars and appointments coupled with the ability to track where assets or equipment is in space has immense potential to change how designers think about the built environment. Take for example the difficulty of a broker trying to meet a client at a new apartment. The paths that each person takes on that given day is probably quite different. However it isn’t too difficult to imagine that their calendars could sync their meetings when there was potential for the two parties to be in close proximity.

A broker (blue) and her client (red) take different paths but overlap for a moment (purple). Each person’s calendar (top right).

However the importance of sequencing does not stop at space planning. Consider the importance computation can play in optimizing assembly line sequences, supply chain logistics, routing or value chain mapping.

Companies like Zipcar and Citibike were able to create successful businesses around optimizing what is in use at any particular time depending on where it is in space. This will only become more true as climate change and growing populations put higher pressure on resources, demanding more efficient use of resources.

The design of a building is actually about the design of an experience. While architects have focused on phenomenological aspects of buildings and cities, their emphasis has been overly focused on material aspects of a building. However, human behavior is increasingly controlled by technology. Increasingly, it isn’t just walls or city grids that define the behavior of humans, but the architecture of our Google calendar, Yelp, and Amazon accounts. These recommendation engines influence where we go and what we do. To be most effective, architects of the future will have to wrestle with these technological circumstances that are evolving human behavior. They must also design the experiences of the city not only through zoning, policy, and physical means, but through choreographing recommended events, working with owners and technologists on new paradigms of use, and exploring dynamic materiality to enable spaces to adapt to individuals in real-time.