Optimal population size for genetic algorithm using fuzzy system

Abstract EN

A genetic algorithm (GA) uses the idea of biological evolution to seek good solutions to problems with very large search spaces.

It has the following parameters: population size, crossover rate, and mutation rate.

The selection of the initial parameters for a GA is very difficult.

Some attempts to find optimal combination of parameters used trial and error methods or combination approaches, while others used a GA to find optimal parameters for another GA.

The current work uses fuzzy system to determine optimal population size for any problem which is used the Genetic Algorithm.

Two combinatorial problems with a large search space are used to test the effectiveness of the current work.

The results are validated and GA is shown to be effective for the tested problems.