Toward simulating entire cities with behavioral models of traffic

Abstract

Resilient transportation systems enable quick evacuation, rescue, distribution of relief supplies, and other activities for reducing the impact of natural disasters and for accelerating the recovery from them. The resilience of a transportation system largely relies on the decisions made during a natural disaster. We developed an agent-based traffic simulator for predicting the results of potential actions taken with respect to the transportation system to quickly make appropriate decisions. For realistic simulation, we govern the behavior of individual drivers of vehicles with foundational principles learned from probe-car data. For example, we used the probe-car data to estimate the personality of individual drivers of vehicles in selecting their routes, taking into account various metrics of routes such as travel time, travel distance, and the number of turns. This behavioral model, which was constructed from actual data, constitutes a special feature of our simulator. We built this simulator using the X10 language, which enables massively parallel execution for simulating traffic in a large metropolitan area. We report the use cases of the simulator in three major cities in the context of disaster recovery and resilient transportation.

Exact-Differential Large-Scale Traffic Simulation

“…Second, the traffic simulation is required to be modeled on the top of optimistic PDES since the exact-differential simulation is totally based on the optimistic PDES as we discussed in Section 2. To meet such requirement, the model of our traffic simulation is based on IBM Megaffic [13,14,17] and SCATTER [15,18].…”

“…In Tokyo traffic simulation [13,14,16,6], for example, we need to execute 770K times simulation when we simulate what happens if one of the roads is blocked because there are 770K junctions in Tokyo. When we simulate multiple blocks of the roads, we need to execute 2 770K times (the sum of combination from 770K choosing 0 to 770K).…”

Towards large-scale what-if traffic simulation with exact-differential simulation

“…The model of the traffic simulation is based on IBM Megaffic (Osogami, Imamichi, Mizuta, Morimura, Raymond, Suzumura, Takahashi, and Ide 2012, Osogami, Imamichi, Mizuta, Suzumura, and Ide 2013, Suzumura, Kato, Imamichi, Takeuchi, Kanezashi, Ide, and Onodera 2012 and SCATTER (Perumalla 2006, Yoginath andPerumalla 2008), where traffic systems are represented as discrete event systems. The individual vehicles' behavior is based on Megaffic, where a vehicle's track of junctions from origin to destination is all fixed before execution by estimation with some defined behavior model, for example, shortest path or minimum hops of junctions.…”

“…For example in simulation of Tokyo traffic (Osogami, Imamichi, Mizuta, Morimura, Raymond, Suzumura, Takahashi, and Ide 2012, Osogami, Imamichi, Mizuta, Suzumura, and Ide 2013, Suzumura and Kanezashi 2013, Hanai, Suzumura, Ventresque, and Shudo 2014, which has 770K junctions, we need to repeat simulation tasks 770K times when we pick up one junction from the 770K junctions and simulate what happens if the junction is blocked. When we simulate multiple blocks of the junctions, we need to execute 2 770K times (the sum of combination from 770K junctions choosing 0 to 770K).…”

An Adaptive VM Provisioning Method for Large-Scale Agent-Based Traffic Simulations on the Cloud

“…In our study we use Megaffic, a large-scale distributed agent-based traffic simulator developed by IBM [22], [23], [27]. The advantage of using Megaffic for traffic simulations is twofold.…”