A Novel Search Biases Selection Strategy for Constrained Evolutionary Optimization

The issues of the search biases selection based on stochastic ranking are pointed out by an example with three possible outputs and are also demonstrated by an experiment designed here. In order to improve the explicit search biases ability in feasible regions, three conditions for explicit search biases are presented and a novel search biases selection strategy with stochastic ranking is proposed in this paper. This strategy is applied to our new algorithm based on ES (evolution strategy). The new algorithm has been tested on 13 common benchmark functions and the experimental results have demonstrated that to some extent the convergence speed, the numerical accuracy and stability of best solutions are improved.