Why Melbourne needs a Dynamic Traffic Assignment (DTA) model?

​Melbourne, while being recognized as the most livable city in the world, suffers from tremendous economic loss due to urban traffic congestion. The avoidable cost of congestion for the Australian capital cities is estimated to be $16.5 billion for the 2015 financial year, having grown from $12.8 billion in 2010. Congestion can't be entirely eliminated, to be realistic. However, we can mitigate it by various means such as better use of existing infrastructure, reducing the mismatch between supply and demand, better land use, advocating active modes, applying travel demand management strategies, implementing congestion pricing, use of advanced technologies to improve efficiency of the system, and etc. To examine any of these measures before real-world implementation, transport planners and engineers require a model of the transportation system. Currently, most of the transportation modeling work in Melbourne is done either using the 4-step MITM/VITM and ZENITH models or through more detailed microsimulation of specific sites. Both are great but have their own limitations. MITM/VITM and ZENITH are classical 4-step models based on a static traffic assignment while microsimulation, although capable of modeling detailed traffic at a fine level, can't be applied to a large-scale network and often does not consider realistic route choice. Today, Melbourne requires to take the next few big and timely steps towards use of more advanced models such as activity-based models (ABM) and dynamic traffic assignment (DTA), a unified tool for transportation planning and traffic operations.

What is DTA?
A Dynamic Traffic Assignment model, known as DTA, estimates the evolution and propagation of congestion through detailed models that capture travel demand, network supply and their complex interactions. Unlike a static model, a DTA model can describe time-dependent dynamics of traffic and replicate the interactions between travelers and the transportation network.

DTA for Transportation Planning and Traffic Operations
Travel time and cost are key components of all travel demand models throughout the entire sequence of modeling steps (4-step or activity based). Travel time and cost measures obtained from static traffic models are time-invariant. DTA models seek to provide more detailed means to represent the interaction between travel choices, traffic flows, and time and cost measures in a temporally coherent manner. DTA models aim to describe time-varying network and demand interaction using a behaviorally sound approach.
​
Most traffic analysts rely on microscopic simulation modeling to assess the performance of transportation systems. Microscopic simulation models are not suitable for large-scale applications and do not often consider individual traveler’s route choice preference. The equilibrium-seeking DTA models are based on iterative algorithmic procedures that describes individual route choice adjustment. Mesoscopic DTA models are suitable for large-scale network applications.

Static vs. Dynamic
In a static model, inflow to a link is always equal to the outflow: the travel time simply increases as the inflow and outflow (volume) increases. The volume on a link may increase indefinitely and exceed the physical capacity (in vehicles per hour) of the link, as represented by a volume-to-capacity (V/C) ratio greater than 1. Since the link volume does not conform to the traffic flow limit that results from the physical characteristics of the roadway, the assigned link volume can be considered as demand— trips desired to traverse the link—instead of the actual flow. Following is an example graphic of V/C ratio in Sydney network from a study done by SGS using ZENITH to analyze the economic case for WestConnex.

​In dynamic models, as in reality, explicit modeling of traffic flow dynamics ensures direct linkage between travel time and congestion. If link outflow is lower than link inflow, link density will increase (congestion), and speed will decrease (fundamental speed–density relationship), and therefore link travel time will increase. This brings forth the issue of congestion spillback, which is not represented in a static model. A dynamic model can reproduce spillback and congestion propagation.

DynaMel: A Large-Scale Mesoscopic Simulation-based DTA Model of Melbourne
DynaMel is a simulation-based Dynamic Traffic Assignment (DTA) model of Melbourne. DynaMel is an analysis tool to address complex and dynamic transportation operations and planning issues of Melbourne road network. DynaMel overcomes the limitations of traditional static assignment models by using advanced traffic modeling techniques. DynaMel captures the dynamics of congestion formation and dissipation associated with time-dependent demand and network conditions.

​Following is a snapshot of 9:30 AM traffic patterns (assigned volumes) simulated by DynaMel.

DynaMel supports transportation network planning and traffic operations decisions, overcoming many of the known limitations of static traffic models used in current planning and operations practice. DynaMel can be used to assess the impacts of ITS and non-ITS technologies on the transportation network. DynaMel can be used to support decision-making for traffic control and management. DynaMel can be used to evaluate different congestion pricing schemes, to plan for special events, and emergency situations including evacuation scenarios.

To learn more about DynaMel, please visit: http://www.cityxlab.com/dynamel-dynamic-traffic-assignment-model-of-melbourne.html
​
References

• Dynamic Traffic Assignment: A Primer (2011) by Transportation Research Board of the U.S. National Research Council
• Deployment and Calibration of a Large-Scale Mesoscopic Dynamic Traffic Assignment Model of Melbourne, Australia (2017) by Shafiei, S., Gu Z., Sarvi M., Saberi M. presented at the Transportation Research Board (TRB) 96th Annual Meeting, Transportation Research Council of the U.S. National Academies, Washington, D.C., January 8–12, 2017. [pdf]
• A Sensitivity-Based Linear Approximation Method to Estimate Time-Dependent Origin-Destination Demand in Congested Networks (2017) by Shafiei, S., Saberi M., Zockaie, A., Sarvi, M. published in the Transportation Research Record: Journal of the Transportation Research Board. [pdf]
• Application of an Exact Gradient Method to Estimate Dynamic Origin-Destination Demand for Melbourne Network (2016) by Shafiei, S., Saberi, M., Sarvi M. presented at The 19th International IEEE Conference on Intelligent Transportation Systems (ITSC), Rio de Janeiro, November 1-4, 2016. [pdf]