A fuzzy evolutionary framework for adaptive agents
Ferdinando Cicalese, Antonio Di Nola, Vincenzo Loia · 1999
Working on real size problems often requires to represent and treat the uncertainty that inevitably derives fmm one or more information sources.The recent explosion of agent-based technology, made nowadays so attmctiue thanks to the availability of global information infmstructure, stimulates new approaches wherr! the uncertainty of compiex problems is addressed with agent-based technology.In this work we describe a massive concurrent computational model in which the multi-agent platform adapts its global behaviour through a distributed fuzzy evolutionary model.Our fmmework is based on the notion of fuzzy reasoning actor, that is an autonomous entity that concurs its local behavior in a distributed environment; the environment (part of it composed by the agents themselves) is not static but submitted to an evolutionary law that improves the global ability of the system by genemting automatically new agents better skilled to perform their tasks.sources, by massive processing of very rich data sets, and by cooperative resolution of goals and conflicts; l the solution space of such domains is often too complex to be represented with a complete model that copes all the features of the problem.