中国和巴西港口拥塞指数多重分形研究

Multifractal Research on Congestion Index of Chinese and Brazil Port

运用多重分形消除趋势波动分析法(MF-DFA)与多重分形消除趋势波动交叉相关分析法(MF-DCCA)对全球港口拥塞指数的中国港口ETA(Estimated Time of Arrival)海岬型船与巴西港口ETA海岬型船港口拥塞指数进行研究。研究表明:中国港口ETA海岬型船与巴西港口ETA海岬型船时间序列广义Hurst指数及MF-DCCA广义Hurst指数,均单调递减且值域在(0.5,1)之间,标度指数均表现凹向横轴的函数,分形谱密集钟形分布;两组时间序列均具有明显的多重分形特征,同时具有长期记忆性与多标度性,且二者多重分形交叉相关,表现长期相互影响。

The multifractal detrended fluctuation analysis method（ MF-DFA） and multifractal detrended fluctuation cross correlation analysis method（ MF-DCCA） were used to study the congestion index of ETA（ Etimated Ttime of Arrival） capesize in Chinese ports and in Brazil ports,which were components of GPCI（ the global port congestion index）. It is found that generalized Hurst index of ETA capesize in Chinese ports and in Brazil ports in time series and MF-DCCA generalized Hurst index are both monotonically decreasing and the range of their values is between 0. 5 and 1; scaling index shows a concave function and fractal spectrum shows a intensive bell shaped distribution. Research results show that the two groups of experimental time series both have obvious multifractal characteristics and also have a long memory and multi-marked degree; moreover,the two multifractals are cross-correlated,and their performances are mutually influenced in a long-term.