This paper describes a new evolutionary algorithm for the automatic generation of the knowledge base for fuzzy logic systems. In common with other evolutionary approaches, the approach adopted is to treat the problem of knowledge base generation as that of searching for a solution of an acceptable quality by applying genetic operators to a population of potential solutions. The algorithm presented dynamically adjusts the focus of the genetic search by dividing the population into three subgroups, each concerned with a different level of knowledge base optimization. The algorithm also includes a new adaptive selection routine that aims to keep the selection pressure constant throughout the learning phase.
|Number of pages||14|
|Journal||Institution of Mechanical Engineers. Proceedings. Part C: Journal of Mechanical Engineering Science|
|Publication status||Published - 1 Jan 2002|
- Evolutionary algorithms
- Fuzzy learning classifier systems
- Fuzzy logic