A Fusion Algorithm of Gravitational Search and Tabu Search
Ao Lou · 2019
In order to overcome the shortcomings of gravitational search algorithm (GSA), such as easily falling into local optimum traps and needing to be improved in accuracy, a fusion algorithm of gravitational search and tabu search (TS-GSA) is proposed. Firstly, the optimal solution of each iteration operation of GSA is regarded as a tabu search object. Secondly, a neighborhood search method is designed to search in depth between two historical extremes to avoid missing global optimum solution. Thirdly, tabu list is used to reduce the frequency of circuitous search and improve search efficiency. The simulation results show that TS-GSA has stronger global optimization ability than GSA, new hybrid populational-based algorithm with combination of particle swarm optimization and GSA (PSOGSA), adaptive gbest-guided GSA (GGSA), and can effectively overcome local optimum problem. Results also show that TS-GSA can find the theoretical optimum solutions of some high-dimensional multi-modal functions, and has stronger robustness for low-dimensional multi-modal functions.