Fuzzy agents versus intelligent swarms
Dimitar Lakov · 2003
The paper considers development of soft computing agents (SCA) via a fuzzy paradigm into its narrower class: fuzzy intelligent agents (FA). As a synergy between soft computing technology and intelligent agents SCA take advantages of flexible interpretation of the first and autonomous functioning of the second. Although the fuzzy intelligent agent narrows its tools to the fuzzy paradigm it still makes clear the basis of the fuzzy seeking strategy. The investigation is focused on information FA using such a strategy. A promising advance in optimal routing marks ant-based intelligent agents. The practice of swarm oriented agents SOA investigates a great majority of perspective decisions that are clarified using a pheromone-like strategy. A drawback of this approach is the rather intensive information exchange that forms optimal decisions. We propose a similar swarm-oriented strategy based on the fuzzy paradigm. In comparison with SOA it has two advantages: (i) significant decrease of information exchange due to single swarm flooding after service activation; (ii) customer oriented preferences by service implementation. The strategy is applied for creation of information FA in optimal routing tasks. These agents provide timely, consistent, and qualitative information to the node soliciting services. In such a way the bottleneck in the chain of information decision-state refreshing is decreased.