Skip to main content
Skip main navigation
No Access

Neighbourhood-based small-world network differential evolution with novelty search strategy

Published Online:pp 65-76https://doi.org/10.1504/IJBIC.2023.134971

Differential evolution (DE) algorithms that focus too much on local search ability will lead to premature convergence, but excessive reliance on global search ability will cause poor search ability. Therefore, balancing global search ability and local search ability is a controversial research topic in DE. In this paper, we propose a DE algorithm based on small-world network topology and novelty search (NSWDE), which uses small-world network to construct dynamic neighbourhoods to change the dynamics of information exchange within populations. The novelty search strategy will continuously search for new behaviours in the search space. Additionally, we set different scaling factors for the two different drivers based on fitness and novelty to regulate the population diversity effectively. To verify the effectiveness of NSWDE, it is compared with several advanced DE algorithms on 18 benchmark test functions; the results show that NSWDE improves both robustness and effectiveness for complex optimisation test functions.

Keywords

differential evolution algorithm, small-world network, novelty search, topology structure, parameter adaptation