The Evolutionary Optimizer (EVO)

The Evolutionary Optimizer (EVO) is a tool for optimizing any systems whose properties are determined by numerical parameters (fuzzy controllers, for example). The approach for optimizing the parameters is adapted from the biological evolution: A "population" of several parameter sets represents a "parents" generation, which generates "children" (new parameter sets). Generation for generation, the properties of the system will be improved by modifying the parameter sets using the following evolutionary operators:

Selection of the parents according to their quality ("survival of the fittest"),
Recombination (mixing) of properties from different parents (transmission),
Mutation (random changes) of properties.

The selection of parents and the generating of children with new properties will be repeated until the system has reached the required optimizing level.

Screen shot · Softcomputing · Contents