Swarm Intelligence From Natural to Artificial Systems

Social insects--ants, bees, termites, and wasps--provide us with apowerful metaphor for creating decentralized problem-solving systems composed ofsimple interacting, and often mobile, agents. The emergent collectiveintelligence of social insects, swarm intelligence, lies not in complexindividual capabilities but rather in networks of interactions among individualsand between individuals and their environment. The daily problems solved by asocial insect colony--finding food, dividing labor among nestmates, buildingnests, responding to external challenges--have important counterparts inengineering and computer science.The collective behavior of social insects is not only decentralized, it is alsoflexible and robust: flexibility allows adaptation to changing environments,while robustness endows the colony with the ability to function even though someindividuals may fail to perform their tasks. Swarm-intelligent artificialsystems exhibit the same features and provide a useful approach to thetremendous increases in the amount of information and in the complexity ofcomputer software. These systems replace an emphasis on control, preprogramming,and centralization with designs featuring autonomy, emergence, and distributedfunctioning. This unique book surveys several examples of swarm intelligence insocial insects and describes how to design distributed algorithms, multiagentsystems, and groups of robots modeled on social insects.

Eric Bonabeau is currently the Interval Research Fellow at the Santa Fe Institute.