Genetic Algorithms represent a technique of Artificial Intelligence which has developed from the paradigm of biological evolution. They use a population of potential solutions which gradually evolve toward the best solution which satisfies an objective function. By their nature, Genetic Algorithms use random numbers. In a typical algorithm running, a random number generator is used in many occasions, like selection of the best individuals, choosing the parents for crossover and actually applying crossover, and in mutation. Relying on a standard algorithm for random numbers has the advantage of simplicity and easy implementation (for example in embedded applications), but the quality of the random numbers could influence the final results. In this paper we investigate the effect of the random number generator used by a genetic algorithm in finding the optimal solution for two test functions.
[1]
Takuji Nishimura,et al.
Mersenne twister: a 623-dimensionally equidistributed uniform pseudo-random number generator
,
1998,
TOMC.
[2]
Pierre L'Ecuyer,et al.
TestU01: A C library for empirical testing of random number generators
,
2006,
TOMS.
[3]
Donald Ervin Knuth,et al.
The Art of Computer Programming
,
1968
.
[4]
John H. Holland,et al.
Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence
,
1992
.
[5]
Melanie Mitchell,et al.
An introduction to genetic algorithms
,
1996
.
[6]
Heinz Mühlenbein,et al.
The parallel genetic algorithm as function optimizer
,
1991,
Parallel Comput..