CS-DE: Differential Evolution with Cooperative Strategy for numerical optimization
Zhenyu Meng, Jeng‐Shyang Pan · 2019
Differential Evolution (DE) is arguably a very simple and efficient global optimization algorithm, nevertheless, there are still some shortcomings existing within its latest powerful variants. For example, the mutation strategy in the state-of-the-art DE algorithm performing good on some benchmarks may perform worse on the others as each mutation strategy has its own disadvantages. Therefore, in this paper, we proposed a novel Differential Evolution variant with Cooperative Strategy (CS-DE) which can tackle the weakness of DE variant employing a single mutation strategy. The novel algorithm proposed in the paper can be considered as a further improvement of a former proposed LPALMDE algorithm as they share the similar evolution framework and employ the same strategy in classifying individuals. Moreover, by incorporating a novel mutation strategy with depth information, the novel CS-DE algorithm proposed in this paper can obtain an overall better performance in comparison with several well-known DE variants. CEC2013 test suite for real-parameter single-objective optimization is employed in the validation of our algorithm and the experiment results demonstrate its competitiveness.