Random Matrix Theory Analysis on the Spatial Correlation of Section Speed

It is of great practical significance to understand the structure and operation condition of urban road network. In the Random Matrix Theory (RMT) analysis, we calculate cross-correlation matrix C of road sections in Shanghai, and analyze the distribution of eigenvalues and their corresponding eigenvectors of matrix C. A remarkable agreement between theoretical prediction and empirical data concerning the distribution of eigenvalues is shown, while there are still a few eigenvalues of matrix C fall outside the bounds predicted by RMT, and the distribution of eigenvectors corresponding to these eigenvalues display deviations from RMT results. In addition, the deviating eigenvectors show stability in time. At last, some applications of RMT analysis in transportation are discussed.