Urban rail transit network planning based on dynamic spatiotemporal graph convolution

To expand the scale and enhance the comprehensiveness of urban rail transit networks, a strategic framework for urban rail transit network planning has been developed by leveraging the power of dynamic spatiotemporal graph convolution. This innovative approach utilizes a dynamic spatiotemporal graph convolutional network to effectively identify spatiotemporal interdependencies within rail transit sensor data. Through the integration of dynamic spatiotemporal graph convolution and a convolution interaction module, it performs spatiotemporal feature extraction and interactive learning on rail transit data. A physical continuous environment model, which incorporates slope and terrain details, is constructed for urban rail transit network planning. With the aim of minimizing the overall passenger travel time and considering spatiotemporal attributes and supply-demand balance, a comprehensive planning model for urban rail transit networks is established. Experimental results verify the effectiveness of the proposed method. The use of dynamic spatiotemporal graph convolution and related comprehensive planning models provides an efficient method for optimizing urban rail transit network planning. It not only improves network coverage and connectivity, but also significantly reduces transfer time, thereby enhancing the overall efficiency and quality of urban rail transit services.