Home Agents & MAS Swarm Intelligence
Swarm Intelligence and Stigmergy Print option in slimbox / lytebox? (info) E-mail

Swarm Intelligence is a property of systems of behaviorally unsophisticated entities exhibiting collectively complex and intelligent behavior. An appealing feature of swarm intelligence is that, despite the fact that the basic components are simple only, the output/outcome of the system can be very sophisticated. And this is accompanied by a highly robust, flexible, and adaptable system architecture. It permits to provide solutions that are even in the case of NP-hard problems astonishing good even so, of course, not necessarily optimal. What is also quite astonishing is the fact that the architecture and the functionality of the system is very simple. The individual in such an environment counts nothing; it is the interaction between and the common efforts of the society that makes the difference. Ants and ant colonies are closest to what is realized within this project. Thus, this idea will be discussed in more detail.

Especially, the examination of the foraging process permits some insights which are very valuable for the proposed system. Foraging relies on a few very simple behavioral and collaboration patterns. Communication is not direct but done through the use of chemical substances known as pheromones. They have a scent that decays over time through the process of evaporation (cf. [35], [36], [37], [38]). In order to solve their need for food as efficiently as possible ants actually have to solve the shortest path problems in order to find food sources closest to their colonies. Altogether, the foraging process of ants exhibits the following characteristics:

  1. Random search
    In the absence of any signs (expressed by scents) for food sources in the environment, ants set out in search of a food source by randomly walking around.
  2. Reinforcement learning and self-organization
    Once an ant finds a food source, it drops pheromones on its food-packed way back to its nest. By this it creates a scent trail between its colony and the food source. Whether this trail will be populated or not depends on the behavior of the other ants. Each ant that uses the trail deposes fresh pheromones and, by this, reinforces the information related to this food source. If the trail is not used for a while, e.g., since there a closer food sources or after the food source is exhausted, it will eventually evaporate. The described behavior of ants is based on negative (abandonment of trails due to evaporation of scent if the trails are no longer used) and positive feedback (reinforcement of pheromone trails by attracted ants) which is one of the basic requirements for self-organization to occur. It causes the emergence of order without any global coordination (cf. [35]).
  3. Stigmergy
    The above described use of asynchronous interaction and information exchange between agents (by pheromones), mediated by an active environment that changes constantly due to new and diminishing trails, is called stigmergy (cf. [39]). Stigmergic agent architectures are reactive and environment-driven. Principles of this approach are (cf. [36]):
    • Agents are simple, reactive, and unaware of other agents or of the emerging complex activities of the agent society
    • The environment is an important mechanism to guide activities of these agents and to serve as an up to date information storage and retrieval system for the agent society.
  4. Emergent behavior
    The foraging process of ants illustrates the emergence of coordinated behavior at the level of the ant society, although for the individuals themselves it is neither relevant nor visible whether they are taking part in a concerted activity. Ants follow their own simple agenda. Their behavior is only influenced by the permanently changing environment (cf. [35], [36]). If the shortest path is obstructed or the food source is exhausted, then at some later point in time the second best path will have the strongest pheromone. Thus, it will induce ants to traverse it thereby strengthening its scent. Thus, the decay in the pheromone level leads to redundancy, robustness and adaptivity, characteristics that describe emergent behavior (cf. [37]).
  5. Probabilistic behavior
    The behavior of an ant is always probabilistic: When it finds a trail due to some pheromones there is a high probability that it follows this trail, however no certainty. This is due to the fact that the probability depends on the strength of the pheromones.

 

 
Joomla templates by a4joomla