Abstract :
Complex systems can be defined as systems with behavior that is poorly understood. These systems appear in a large variety of fields, such as natural phenomena. Modeling natural phenomena is an interesting challenge because natural phenomena, such as earthquakes, hurricanes and volcanoes, can cause severe damage. What´s more, in this area, most classical models have failed both to understand the underlying physical processes and to predict future behavior. Our approach is to consider macro-behavior (i.e., the result of the program execution) as the result of a set of interactions among smaller, independent entities. The whole system is then distributed among these entities, called agents. Our research project, named GEOMAS (GEOphysics and MultiAgent Systems), was started in 1994. As a test case, a complex system dealing with predicting volcano eruptions has been investigated. In particular, we studied the volcano of La Fournaise in La Reunion island
Keywords :
artificial intelligence; geophysics; geophysics computing; knowledge based systems; large-scale systems; software agents; volcanology; GEOMAS; GEOphysics and MultiAgent System; La Fournaise; La Reunion island; agents; complex systems; macro-behavior; multiagent approach; natural phenomena; program execution; volcano eruptions; Autonomous agents; Computational modeling; Distributed computing; Embedded computing; Lenses; Message passing; Multiagent systems; Programming; Symbiosis; Volcanoes;