BIOLOGICAL INSPIRATION FOR COMPUTING 253
8.2 EXAMPLES OF BIOLOGY AS A SOURCE OF PRINCIPLES FOR COMPUTING
8.2.1 Swarm Intelligence and Particle Swarm Optimization
Swarm intelligence is a property of systems of nonintelligent, independently acting robots that
exhibit collectively intelligent behavior in an environment that the robots do sense and can alter.^8 One
form of swarm intelligence is particle swarm optimization, based on the flocking of birds.^9
The canonical example of flocking behavior is a flight of birds wheeling through the sky, or a school
of fish darting through a coral reef. Somehow, myriad not-very-bright individuals manage to move,
turn, and respond to their surroundings as if they were as a single, fluid organism. Moreover, they seem
to do so collectively, without a leader: biologists armed with high-speed video cameras have shown that
the natural assumption—that each flock or school has a single, dominant individual that always ini-
tiates each turn just a fraction of a second before the others follow—is simply not true.
The first known explanation of the leaderless, collective quality of flocking or schooling behavior
emerged in 1986. This explanation used swarms of simulated creatures—“boids”—that could form
surprisingly realistic flocks if each one simply sought to maintain an optimum distance from its neigh-
bors. The steering rules of the so-called Reynolds simulation were simple:^10
- Separation: steer to avoid crowding local flock mates.
- Alignment: steer toward the average heading of local flock mates.
- Cohesion: steer toward the average position of local flock mates.
These rules were entirely local, referring only to what an individual boid could see and do in its
immediate vicinity;^11 none of them said, “Form a flock.” Yet the flocks formed every time, regardless of
the starting positions of the boids. These flocks were able to fly around obstacles in a very fluid and
natural manner. Sometimes the flock would even break into subflocks that flowed around both sides of
an obstacle, rejoining on the other side as if the boids had planned it all along. In one run, a boid
accidentally hit a pole, fluttered around for a moment, and then darted forward to rejoin the flock as it
moved on.
Today, the Reynolds simulation is regarded as one of the best and most evocative demonstrations of
emergent behavior, in which complex global behavior arises from the interaction of simple local rules. The
approach embodied in the simple-rule/complex-behavior approach has become a widely used tech-
nique in computer animation—which was Reynolds’ primary interest in the first place.^12
(^8) T. White, “Swarm Intelligence: A Gentle Introduction with Applications,” PowerPoint presentation, available at http://
http://www.sce.carleton.ca/netmanage/tony/swarm-presentation/tsld001.htm.
(^9) Bird flocks are an example of complex, adaptive systems. Among the many other examples that scientists have studied are the
world economy, brains, rain forests, traffic jams, corporations, and the prehistoric Anasazi civilization of the Four Corners area.
Complex adaptive systems are similar in structure and behavior even if they differ in their superficial manifestations. For
example, complex adaptive systems are massively parallel and involve many quasi-independent “agents” interacting at once.
(An agent might be a single firm in an economy, a single driver on a crowded freeway, and so on.) Each of them is adaptive,
meaning that the agents that constitute them are constantly responding and adapting to each other. And each of them is
decentralized, meaning that no one agent is in charge. Instead, a complex system’s overall behavior tends to emerge spontane-
ously from myriad low-level interactions.
(^10) C.W. Reynolds, “Flocks, Herds, and Schools: A Distributed Behavioral Model,” Computer Graphics 21(4):25-34, 1987, available
at http://www.cs.toronto.edu/~dt/siggraph97-course/cwr87/ and http://www.red3d.com/cwr/papers/1987/SIGGRAPH87.
pdf. An updated discussion, with many pictures and references to modern applications, can be found in C.W. Reynolds, “Boids:
Background and Update,” 2001, available at http://www.red3d.com/cwr/boids/.
(^11) More precisely, each boid had global information about the physical layout of its environment, including any obstacles, but it
had no information about its flock mates, except for those that happened to come within a certain distance that defined its local
neighborhood.
(^12) The first Hollywood film to use a version of Reynolds’ boids software was Tim Burton’s Batman Returns (1992), which
featured swarms of animated bats and flocks of animated penguins. Since then it has been used in films such as The Lion King
(1994) and many others (see http://www.red3d.com/cwr/boids/)..)