A parameter-less genetic algorithm with customized crossover and mutation operators

Abstract

Genetic algorithm is one of the well-known population based meta-heuristics. The reasonable performance of the algorithm on a wide variety of problems as well as its simplicity made this algorithm a first choice in lots of cases. However, the algorithm has some weaknesses such as the existence of some parameters that need to be carefully set before the run. The capability of the parameters to change the balance between exploration and exploitation make them crucial. Exploration and exploitation are the bases of every evolutionary algorithm. Conducting a balance between these elements is crucial for the success of any evolutionary algorithm. In this research a GA is proposed on which the crossover and mutation rates are removed. A probability vector holds the probability of the alleles for every locus within the individual. The probability is with regards to the contribution of the allele on either increasing or decreasing the fitness of the chromosome. The probability of an allele will increase if the fitness of the chromosome increases by a change or vise versa. The experiments conducted on a wide range of multi-modal and epistatic problems show good performance of the proposed method in comparison to other algorithms in literature.

An Adaptive GA in Partitioned Search Space

“…The AGAPSS is mainly inspired from PGA [23] however, these two methods differ in some aspects including the variation operators. Besides that, the AGAPSS is an adaptive method due to utilization of parameter for intensification and diversification.…”

“…Selection of the compared algorithms was based on their relevance to the proposed method. The test functions are the same as those that have been used by the chosen algorithms for comparison [1,23,[30][31][32]. This facilitate, the comparison of the results with other methods in literature.…”

“…To this end, the literature could be divided into two main branches [23], which are parameter calibration methods and parameter-reduced methods.…”

“…Depending on the number of changes (τ ) that have occurred on a given individual, its respective probability vector (Z) will get updated [23] using,…”

Adaptation of Parametric Uniform Crossover in Genetic Algorithm

“…The proposed algorithm will be compared with some of the algorithms from literature including canonical GA with a randomly chosen constant mutation rate [7], PGA [8], SSRGA [7], SSRGA-II [9], self-adaptive (SAGA) [10], adaptive (AGA) [11], and the algorithms in [12] and [13].…”