Chaotic games driven grey wolf optimization: optimal performance design and comprehensive analysis

Abstract

Abstract Introduction : Meta-heuristics have attracted much attention due to their compatibility with other algorithms and excellent optimization ability. Gray wolf optimization (GWO) is also a meta-heuristic algorithm. GWO mainly tries to find the optimal solution by simulating the hierarchical structure and hunting behavior of gray wolves. GWO has the advantages of a relatively simple algorithm structure and fewer parameter Settings. Therefore, it is used in many fields, such as engineering and forecasting. Objectives GWO may have problems in harmonic convergence or be trapped into local optima for some complex tasks. An improved variant of basic GWO is proposed in this paper to efficiently alleviate this deficiency. Preferentially, chaos game optimization (CGO) is introduced into the conventional method to expand its neighborhood searching capabilities. Based on this strategy, we called the improved GWO as CGGWO. Methods To confirm the effectiveness and optimization ability of the CGGWO algorithm, CGGWO is compared with a set of meta-heuristics, including 7 basic meta-heuristics, 7 state-of-the-art meta-heuristics, and 5 enhanced GWO variants. The benchmark functions for comparison are IEEE CEC 2017. The dimensions(D) of the benchmark test function are 10, 30, 50, and 100. Moreover, CGGWO is applied to five practical engineering problems and two real-world benchmarks from IEEE CEC 2011. Non-parametric statistical Wilcoxon signed-rank and the Friedman tests are performed to monitor the performance of the proposed method. Results In benchmark function testing, CGGWO can find better solutions in most functions. In the Wilcoxon signed-rank and the Friedman tests, the P-value of CGGWO is mostly less than 5%. Among the five engineering problems, the feasible solution found by CGGWO is also the best compared with other methods. Conclusions In the benchmark function test, CGGWO has a better convergence effect than other methods and finds a better solution. From the results of the Wilcoxon signed-rank and the Friedman tests, we can see that the CGGWO results are statistically significant. In engineering problems, CGGWO can find feasible solutions.