An Efficient Negative Correlation Gravitational Search Algorithm

Gravitational search algorithm (GSA) is known as an effective optimization algorithm based on population. To further improve the performance of GSA, taking the combination of diversified search mechanisms into consideration would be a constructive solution for increasing the possibility of obtaining global optimum. In the meantime, the negative correlation search (NCS) algorithm has proven its ability of maintaining diversity effectively to develop the population. Thus, with such inspiration, an improved gravitational search algorithm based on negative correlation learning is proposed in this paper. While gravitational search conducts exploitation in the search space, negative correlation fulfills exploration by encouraging discrepant search behaviors to increase the optimization accuracy. The superiority of the proposed algorithm is demonstrated with experimental results based on several benchmark functions in comparison with its component algorithms.