Abstract: Unbalanced distribution of resident and employment is a very predominant issue in the rapid metropolis development. It can cause the unbalance of traffic demand and heavy traffic congestion. We construct a network dynamics model for the relationship between residentemployment distribution and commuting with urban development macro data as influential variables to investigate the relationship in city development. We also establish parameter relationship based wavelet neural networks to solve the model. We further employ Beijing macroeconomic data and commuting data from nine typical residential areas to calibrate and validate the model. Results show that the model can better denote the influence of residentemployment distribution structure on commuting feature and predict the trend of commuting feature variance.

Key words: residentemployment distribution, commuting, wavelet neural networks, networks dynamic