Path-based adaptive traffic signal control with a mixed backpressure index

The network-wide urban traffic signal settings are generally optimized with the link-based overall indices, incapable of taking the traffic flow at the path level into account. This paper proposes a mixed backpressure index for a path considering both its throughput and the movement occupancies. With this index, we propose a one-step control to optimize the network-wide signal timing plans every time step embedded with a path-based store-and-forward model and extend it to multi-step by applying model predictive control methodology for the purpose of avoiding erratic switching among stages. Both the one-step and multi-step control are equivalently translated into mixed-integer linear programming problems for computational efficiency. The results of a numerical study on a calibrated network of Caohejing District in Shanghai indicate that the mixed backpressure index based methods provide better performance in terms of delay time as well as maximum queue length than other throughput-optimal methods. Additionally, they allow for relatively fair distribution of delay time associated with paths with appropriate parameter design.