Real-coded genetic algorithm benchmarked on noiseless black-box optimization testbed

Genetic algorithms—a class of stochastic population-based optimization techniques—have been widely realized as the effective tools to solve complicated optimization problems arising from the diverse application domains. Originally developed version was a genetic algorithm with the binary representation of candidate solutions (i.e. chromosomes), the real-coded versions are, however, basically superior and frequently utilized in tackling complex real-valued optimization tasks. In this paper, a real-coded genetic algorithm (RCGA), which employs an adaptive-range variant of the well-known non-uniform mutation, is furnished with a multiple independent restarts mechanism to benchmark the noisefree black-box optimization testbed. The maximum number of function evaluations for each run is set to 50000 times the search space dimension. For low search space dimensions, the algorithm shows encouraging results on several functions. Although the algorithm is unable to solve all the functions ...

• Type: Conference
• Year: 2010
• Where: GECCO
• Authors: Thanh-Do Tran, Gang-Gyoo Jin
• Keywords: GECCO 2010, Genetic Algorithm, Optimization, Search Space Dimensions, Stochastic Population-based Optimization