Author:
Ghodrat Moghadampour
Affiliation:
VAMK, University of Applied Sciences and Technology and Communication, Finland
Keyword(s):
Evolutionary algorithm, Genetic algorithm, Function optimization, Mutation operator, Multipoint mutation operator, Random building block operator, Fitness evaluation and analysis.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence and Decision Support Systems
;
Enterprise Information Systems
;
Evolutionary Programming
Abstract:
Genetic algorithms work on randomly generated populations, which are refined toward the desired optima. The refinement process is carried out mainly by genetic operators. Most typical genetic operators are crossover and mutation. However, experience has proved that these operators in their classical form are not capable of refining the population efficiently enough. Moreover, due to lack of sufficient variation in the population, the genetic algorithm might stagnate at local optimum points. In this work a new dynamic mutation operator with variable mutation rate is proposed. This operator does not require any pre-fixed parameter. It dynamically takes into account the size (number of bits) of the individual during runtime and replaces a randomly selected section of the individual by a randomly generated bit string of the same size. All the bits of the randomly generated string are not necessarily different from bits of the selected section from the individual. Experimentation with 17
test functions, 34 test cases and 1020 test runs proved the superiority of the proposed dynamic mutation operator over single-point mutation operator with 1%, 5% and 8% mutation rates and the multipoint mutation operator with 5%, 8% and 15% mutation rates.
(More)