Simulating Complexity
Today, new technologies and methods are emerging for exploring and visualizing urban complexity through these many systems and invisible dimensions. Urban planning simulation, economic modeling, traffic simulation, operations research, and agent-based modeling comprise the new tools. Open data platforms and web based gaming applications give immense power to designers, researchers and engineers to create their own simulations.
Agent Based Models
An agent-based model (ABM) is a class of computational models for simulating the actions and interactions of autonomous agents (both individual or collective entities such as organizations or groups) with a view to assessing their effects on the system as a whole. [ref Wikipedia]
ABMs are an artificial representation of the world. They are a subjective view of how systems interact based on the encoded rules, interactions and systems an author chooses to represent.
Get Violet Whitney’s stories in your inbox
Join Medium for free to get updates from this writer.
While not a realistic representation of reality, agent based models can help us understand our world by creating simplistic models to test assumptions. Even from simple rules we can test how many complex behaviors emerge when these rules are applied to many agents. For example a simple rule to have a bird follow another bird within a short proximity to one another, creates complex clustering and swooping behavior that mimics a flocks of birds.
Components of an Agent Based Model
individuals (agents) — could be humans, parcels of land, vehicles, animals, etc
environment — framework in which the interactions occur, i.e. a city, a neighborhood, a floor plan, a room, etc
behavior — the procedural rules that define how the individuals behave — i.e. a person move away from other people if a space gets crowded, or a person may be attracted to other people with particular characteristics, etc
parameters — characteristics of the agent — i.e. a agent that is a person might have speed, size, distance the agent will stand next to other agents, etc
input — parameters that globally drive the model — i.e. population size of your agents, climate or location for your environment, or other data that is put into the model
output — global parameters that are the outcome of running your model
While we can recognize that these simulations are replicas of reality, the aim of simulation should also be to make our assumptions and biases explicit and testable, moving assumptions out of the black box of hidden algorithms. By clearly defining the elements of a simulation, assumptions can be vetted.
Open simulation tools can extend agency to people outside of a disciplinary expertise. Gaming platforms can combine simulations across disciplines and have intuitive user interfaces that allow anyone to participate in rulesets usually encapsulated in an industry’s specific modeling technique and software.
