Toward simulating entire cities with behavioral models of traffic T. Osogami T. Imamichi H. Mizuta T. Suzumura T. Ide ´ 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. Introduction Transportation authorities can rely on traffic simulations [1, 2] to evaluate the effectiveness of a particular action for a particular traffic situation. Traffic simulation can be used to infer the consequences of an action by applying predefined rules that govern the behavior of traffic flows. The models of traffic simulation range from Bmicroscopic[ to Bmacroscopic,[ depending on the level of detail [3, 4]. A microscopic model tracks the location of individual vehicles, while a macroscopic model tracks some features of flows such as speed and density. The quality of the decisions that the transportation authority can make often depends on how realistic the model of traffic simulation is. Microscopic models allow more detailed study and more faithful modeling of transportation systems than their macroscopic counterparts [4]. As a result, there has been a significant amount of effort in building traffic simulators with microscopic models [5–8]. For example, Nishi et al. [9] report that a minute change in the configuration of merging lanes can significantly reduce congestion, but such an impact can be captured only by microscopic models. A macroscopic model does not allow us to directly study the impact of small changes that cannot be represented by that model. In addition, we cannot directly study the impact of traffic control in more detail than what the macroscopic model tracks. For example, if we want to evaluate the impact of traffic control on specific vehicles such as ambulances, a microscopic model would be more appropriate than a macroscopic model. However, a microscopic model can have considerably more parameters than a corresponding macroscopic model. The values of these parameters, for example, determine how the driver of a vehicle, an agent, chooses the speed, the lane, or the route, as well as the origin and the destination. We need to carefully set these values for individual agents, because they essentially determine the results of the simulation. Calibrating these values is time-consuming and often relies on intuitions and knowledge of an expert on transportation systems [5, 10, 11]. Because of the difficulty of calibration, some of the details are often omitted from microscopic models. For example, the driver of a vehicle is often assumed to take the route that minimizes ÓCopyright 2013 by International Business Machines Corporation. Copying in printed form for private use is permitted without payment of royalty provided that (1) each reproduction is done without alteration and (2) the Journal reference and IBM copyright notice are included on the first page. The title and abstract, but no other portions, of this paper may be copied by any means or distributed royalty free without further permission by computer-based and other information-service systems. Permission to republish any other portion of this paper must be obtained from the Editor. T. OSOGAMI ET AL. 6:1 IBM J. RES. & DEV. VOL. 57 NO. 5 PAPER 6 SEPTEMBER/OCTOBER 2013 0018-8646/13 B 2013 IBM Digital Object Identifier: 10.1147/JRD.2013.2264906