Discourse | Agents

An agent is an entity capable of action that promotes multiplicities, iteration, and adaptation.

In 1986, Craig Reynolds compiled an algorithm which simulated the aggregate behavior of birds and other flocking, herding, and schooling entities. These computer-simulated agents were termed ‘Boids‘ in his seminal essay, “Flocks, Herds, and Schools: A Distributive Behavioral Model,” published in the proceedings of the 1987 ACM SIGGRAPH conference. Each of Reynolds’ Boids  are conceived as independent actors that navigate according to its local, and thus limited  perception of its environment, the laws of simulated physics  that rule its motion, and a set of protocols encoded  by the “animator.” Reynolds points out that a natural flock seems to consist of two balanced, opposing behaviors: “a desire to stay close to the flock and a desire to avoid collisions within the flock.”¹ Therefore, the aggregate motion of the flock, and any other complex system that displays emergent behavior can be understood as the construct of this negotiation- a dense interaction of relatively simple protocols enacted at the local scale.

Three attributes to further articulate the interactions of agents can be defined  as multiplicity, iteration, and adaptation. Complex systems, like the example of the flock, are often highly parallel in their configuration allowing for multiple problem solutions to be attempted simultaneously. In this regard, the collective is given precedence over the individual and allowed to act collectively and as a multiplicity. While this parallelism involves a multiplicity in space, iteration involves persistence in time. At any given time the interaction of entities will result in changes in the space in which that negotiation has taken place- meaning that any future actions by an agent will be required to take these spatial changes into account, a property which Reynolds’ Boid simulations lacked. This property, referred to as adaptation, serves as the mechanism in which feedback occurs between an agent and its environment over time and represents an internal change in a system that mirrors an external event in its environment. ³

1. Reynolds, Craig. “Flocks, Herds, and Schools: A Distributed Behavioral Model.” Computer Graphics 21(4) (July 1987): 25-34.
2. Hensel, Michael, Heterogeneous Space in Architecture. West Sussex, UK: Wiley & Sons, 2009.
3. Flake, Gary William. The Computational Beauty of Nature Computer Explorations of Fractals, Chaos, Complex Systems, and Adaptation. New York: The MIT P, 2000.

Image redrawn from Albert Pope