Biologically inspired computing came about 20 years ago with the development of algorithms that simulate various aspects of natural processes to calculate useful results. For example, neural networks mimic some aspects of learning in mammalian brains to learn complex patterns; Simulated annealing simulates how metals are cooled in low energy crystalline states to solve difficult minimization problems; And genetic algorithms use abstractions of evolution mechanisms (selection, crossing, mutation) to traverse large spaces of search. All have found their way into mainstream computing, and all are regularly used in a wide range of real-world problems.
In this article, I examine a related technique that in many cases is equal to or better than existing optimization algorithms for a wide range of problems. Ant colony optimisers (ACOs) form sets of virtual insects that cooperate in various tasks. Surprisingly, such sets can be used to produce responses to a range of complex problems, although the simulated insects and the means they use to communicate are extremely simple. For example, ACOs are currently being used to simulate complex routing problems in telecommunication networks, where the network topology may vary over time.
Ant colony algorithms are closely associated with Marco Dorigo, who described the concept in his doctorate. Thesis in 1992.
Ant colony optimization is an example of a swarm algorithm. If you've read the Michael Crichton Prey thriller (Harper Collins, 2002), which describes lugubriously swarms of semi-intelligent nanobots in competition with humans, you're familiar with some of the ideas behind this newly developed area. In a swarm algorithm, a large number of agents cooperate to achieve a global goal without requiring any central control point.
Swarm-based systems are highly fault-tolerant because failure of a component in a swarm does not significantly degrade overall system performance. This makes them particularly suitable for hazardous or remote environments, and the US Army and NASA are currently investigating their use.