Towards sustainable smart cities: An empirical comparative assessment and development pattern optimization in China

Abstract

Dramatically increased societal demands on the municipal services that contradict environmental protection and information processing capability oriented to resource utilization efficiency suffer from opposing simultaneous requirements. The smart city provides better solutions for urban areas which are increasing at an unprecedented speed. This paper presents an empirical study carried out to assess and analyze the development pattern of 35 smart cities in China using the principal component analysis (PCA) and back propagation (BP) neural network techniques. The proposed PCA-BP neural network assessment processing model is applied with six dimensional factors and twenty-two operating indices. With the feature extraction and performance score calculated via PCA, BP neural network is adopted for city classification to recognize the development differences in smart cities. The results indicate that the factor-driven impetus evolves into innovation-driven impetus, narrowing the gap from the holistic perspective between the first and middle-ranking groups, while two middle-ranking groups show a similar upward trend in terms of developing a smart economy through sustainable productivity in innovative enterprises and high-tech industry. To some extent, in response to a similar improving trend in the application of smart services, a distinct advantage of an individual index can be a complementary offset to unapparent holistic highlighting the reception of the lowest average points. Unbalanced development exists in two subaverage groups that are deficient in the initial inventory of smart infrastructure and demands. A relatively large difference exists in the smart mobility index among cities, whereas the opposite case is found concerning the smart environment index. Finally, corresponding optimized development pattern are recommended for building a sustainable smart city.