Signal coordination models for long arterials and grid networks

•A MaxBandLA model is proposed to coordinate signals along long arterials.•MaxBandLA divides a long arterial into small subsystems, and perform signal coordination for each subsystem.•A MaxBandGN model is proposed to coordinate signals in grid networks.•MaxBandGN explicitly optimizes the offsets of all signals so that no cycle constraints are required.•Both MaxBandLA and MaxBandGN models are small-sized mixed-integer linear programs.

This paper proposes two models to tackle traffic signal coordination problems for long arterials and grid networks. Both models, denoted as MaxBandLA and MaxBandGN, are built based on Little’s bandwidth maximization model, and the resulting formulations are both small-sized mixed-integer linear programs. Model MaxBandLA can optimize arterial partition plan and signal coordination plans of all the subsystems simultaneously. Model MaxBandGN directly optimizes the offsets for all the signals in a grid network, and as such, no ’cycle constraints’ need to be constructed. Numerical tests are presented to show that both models have the potential to produce coordination plans that are comparable to signal plans optimized by Synchro.