Research on the Robustness of Degree-Fitness-Distance Evolutionary Networks Based on Response Surface Methodology

Abstract

This paper introduces a Degree-Fitness-Distance (DFD) evolutionary network, utilizing Response Surface Methodology to investigate the impact of degree strength, fitness strength, distance strength, and their interactions on the robustness of the DFD network. The regression equation was determined, followed by a variance analysis of different factors affecting the target response. The results show that the influence of different factors on the size of the largest connected component and the overall efficiency of the network are in the following order: degree strength > fitness strength > distance strength. When degree strength is 1, fitness strength is 2, and distance strength is 3, the size of the largest connected component of the network and the overall efficiency reach their peak values, respectively at 56.57% and 10.15%. Multi-objective optimization was performed on the DFD network; when degree strength is 1, fitness strength is 1, and distance strength is 3, the predicted size of the largest connected component is 58.39%, and the overall efficiency is 10.37%. These figures deviate by approximately 5% from the actual values, which demonstrates that the predictive model possesses high accuracy and reliability.