To address the problem that traditional traffic assignment models struggle to accurately depict the operational state of urban interchanges due to their complex topological characteristics of layered structures and interwoven paths, this study constructs a graph-theory-based traffic flow assignment and path optimization model. First, the interchange is abstracted as a multi-layered dynamic weighted graph incorporating real-time traffic parameters. Next, a flow equalization algorithm coupling inter-layer flow and weaving conflict constraints is designed. Then, an adaptive K-shortest path generation strategy incorporating a structural penalty mechanism is proposed. Finally, a game theory-based distributed path collaborative assignment framework is established. Experimental results show that the model improves interchange traffic efficiency by more than 30.0% during peak hours compared to traditional methods, and maintains a 36.4% speed advantage even in accident scenarios. Parameter optimization further improves performance by 6.0%. Extending to regional road networks, the collaborative strategy reduces total travel time by 15.2%. These findings validate the effectiveness and advancement of the proposed method in accurately describing the dynamic traffic behavior of interchanges and achieving intelligent collaborative management.
Liu et al. (Thu,) studied this question.