Adaptive genetic algorithm for data center resource distribution

Telenik S.F., Rolik A.I., Savchenko P.S.

Аn adaptive version of the genetic algorithm is proposed. Algorithm is based on the simultaneous use of two genetic algorithms that solve the problem of parametric and algorithmic adaptation. This allows you quickly select and adjust the strategy selection parameters for genetic algorithms and the likelihood of their use. Sequential steps of the algorithm are given, introduced the modified parameters of mutation and crossover, introduced the concepts of rewards and performance parameters that are allowed to regulate the sequence of genetic operators, and to adjust the probability of their use.


Full text (pdf)