System-Level Optimization Model for Green-Wave Coordination Control of Urban Road Networks with Mixed Intersection Release

Abstract

Relying solely on NEMA phases for urban green-wave control often restricts the feasible regions of network optimization, resulting in narrow bandwidths or unsolvable models. To address this limitation, this paper proposes a mixed-integer linear programming model for regional signal coordination based on a mixed-phase release strategy that integrates NEMA dual-ring phase and split phase. By utilizing shared lanes under split phasing, the model maximizes lane resource efficiency and extends coordination benefits to left-turn traffic. Introducing 0–1 decision variables establishes a unified formulation for internal phase offsets, enabling flexible, intersection-specific release selection. To balance network efficiency and fairness, the optimization objective minimizes the weighted sum of the red-wave bandwidth-to-cycle ratio, subject to spatiotemporal and clockwise closed-loop constraints. A real-world case study in Suzhou, solved via the branch-and-bound method, demonstrates that the optimal design deploys split phase at seven intersections and NEMA phases at two. VISSIM simulations confirm that compared to the NEMA-only approach, the proposed mixed model reduces red-wave bandwidth by 49.52%, average delays by 26.36%, and stops by 17.5%. The proposed model provides a system-level signal coordination framework for improving the adaptability and reliability of urban traffic control systems.