Traditional urban traffic management treats congestion as a reactive problem—an environmental state to respond to once queues have already formed at major intersections. However, physical, geographical, and financial constraints prevent cities from simply building more roads to meet peak demand (Optimizing lane reversals in transportation networks to reduce traffic congestion: A global optimization approach). As argued by Ben Liang in From "Passive Congestion" to "Active Scheduling" — Rebuilding Travel Experience and Resource Allocation for Intelligent Transportation Networks, mitigating the economic and environmental costs of vehicle idling requires a paradigm shift from "passive congestion" to "active scheduling."
To achieve this, modern frameworks leverage pervasive urban data streams—ranging from mobile phone location data to smart camera networks (Internet of smart-cameras for traffic lights optimization in smart cities; The Application of Big Data Technology in Smart City). By transforming these passive data sources into predictive models, traffic managers can anticipate bottlenecks and dynamically reallocate network capacity before gridlock occurs.
Maximizing Existing Infrastructure via Dynamic Capacity Allocation
One of the most effective ways to optimize a network without physical expansion is to exploit the spatial-temporal asymmetry of urban travel. Traffic demand is rarely uniform; highways frequently experience heavy inflows during morning commutes and matching outflows in the evening.
Tidal-flow lanes and dynamic lane reversals capitalize on this imbalance by changing the direction of specific lanes based on real-time demand. The authors of Adaptive Approaches for Tidal-Flow Lanes in Urban-Road Networks demonstrate that moving from static schedules to adaptive, occupancy-based algorithms can reduce total vehicle travel times by up to 81%.
However, scaling lane reversals across an entire urban grid introduces a massive combinatorial challenge: deciding which lanes to reverse across interconnected intersections is an NP-hard problem. The discrete choice of lane directions creates a non-convex relationship with continuous traffic routing. To solve this, Salomón Wollenstein-Betech et al. developed a convex approximation method in Optimizing lane reversals in transportation networks to reduce traffic congestion: A global optimization approach. By smoothing the discrete lane-allocation variables into a continuous optimization space, their framework allows urban planners to efficiently identify the most impactful lanes to reverse, yielding up to a 60% reduction in travel times during highly asymmetric scenarios.
Targeted and Predictive Interventions
Traditional traffic alerts often fail because they lack granularity. When a navigation app broadcasts a delay to every driver in a region, it frequently shifts the bottleneck to an adjacent route, creating secondary congestion. Proactive optimization solves this by targeting interventions based on the precise origins of the traffic.
By fusing mobile phone datasets with geographic information systems (GIS), Pu Wang et al. discovered in Understanding Road Usage Patterns in Urban Areas that traffic on any given road segment typically originates from a surprisingly small number of source neighborhoods. Instead of broad, non-targeted advisories, routing incentives or congestion management strategies can be surgically directed at these specific driver sources, increasing mitigation efficiency up to sixfold.
To execute these targeted strategies, networks must accurately predict where and when congestion will manifest. Wei Yu et al. utilize the XGBoost algorithm in Research on Traffic Congestion Prediction Based on XGBoost to show that a road's specific spatial orientation and coordinates are often more critical predictors of gridlock than the time of day. When these predictive models are paired with real-time infrastructure, as explored by Yaming Guo et al. in Proactive Coordination of Traffic Guidance and Signal Control for a Divergent Network, cities can synchronize variable message signs with traffic signals. By subtly adjusting the travel time information displayed to drivers, the system balances network-wide loads proactively rather than reacting after an incident occurs.
Balancing Multi-Modal and Three-Dimensional Networks
As urban mobility expands beyond traditional passenger cars to include ridesharing, autonomous vehicles, and Urban Air Mobility (UAM), the optimization problem grows in complexity. The challenge shifts from managing uniform flows to coordinating highly heterogeneous agents across multiple dimensions.
On the ground, the introduction of Connected and Autonomous Vehicles (CAVs) changes traffic dynamics at high-friction bottlenecks like airport curbsides. Xin Chang et al. demonstrate in Analysis on Characteristics of Mixed Traffic Flow with Intelligent Connected Vehicles at Airport Two-Lane Curbside Based on Traffic Characteristics that even partial CAV penetration significantly reduces vehicle delays by smoothing out the stop-and-go patterns typical of human drop-off behavior. Similarly, in ride-pooling networks, the authors of Putting ridesharing to the test: efficient and scalable solutions and the power of dynamic vehicle relocation show that simple, localized vehicle relocation schemes can dramatically minimize passenger wait times and prevent localized supply-demand imbalances.
When transportation moves into the third dimension via drones and air taxis, intersections within aerial corridors become the new critical bottlenecks. To manage these high-density conflict points without relying on ultra-reliable, low-latency communication networks, Sahand Khoshdel et al. propose a rhythmic control framework in SkyGrid: Energy-Flow Optimization at Harmonized Aerial Intersections. By organizing aerial vehicles into virtual platoons and optimizing their trajectories using a multi-objective Pareto approach, they ensure collision-free passage while minimizing energy consumption.
Crucially, these aerial networks cannot be optimized in isolation. If a vertiport is placed poorly, the efficiency gained in the air is instantly wiped out by the ground traffic generated around the hub. To prevent this intermodal friction, Yue Yu 0004 et al. developed a mathematical framework using Mathematical Programs with Equilibrium Constraints (MPEC) in Vertiport Selection in Hybrid Air-Ground Transportation Networks via Mathematical Programs With Equilibrium Constraints. By reformulating the non-convex coupling of air and ground routing into a Mixed-Integer Linear Program, they provide a method to strategically locate vertiports so that the overall hybrid network remains balanced and congestion-free.
Key Foundational Work
A Rolling Horizon Auction Mechanism and Virtual Pricing of Shipping Capacity for Urban Consolidation Centers (2015) — contributed 1 concept elements to this direction.
Multi-Agent Simulation of En-Route Human Air-Traffic Controller (2012) — contributed 1 concept elements to this direction.
Understanding Road Usage Patterns in Urban Areas (2012) — contributed 1 concept elements to this direction.