Book

The Multi-Agent Transport Simulation MATSim

Authors:

Abstract

The MATSim (Multi-Agent Transport Simulation) software project was started around 2006 with the goal of generating traffic and congestion patterns by following individual synthetic travelers through their daily or weekly activity programme. It has since then evolved from a collection of stand-alone C++ programs to an integrated Java-based framework which is publicly hosted, open-source available, automatically regression tested. It is currently used by about 40 groups throughout the world. This book takes stock of the current status. The first part of the book gives an introduction to the most important concepts, with the intention of enabling a potential user to set up and run basic simulations.The second part of the book describes how the basic functionality can be extended, for example by adding schedule-based public transit, electric or autonomous cars, paratransit, or within-day replanning. For each extension, the text provides pointers to the additional documentation and to the code base. It is also discussed how people with appropriate Java programming skills can write their own extensions, and plug them into the MATSim core. The project has started from the basic idea that traffic is a consequence of human behavior, and thus humans and their behavior should be the starting point of all modelling, and with the intuition that when simulations with 100 million particles are possible in computational physics, then behavior-oriented simulations with 10 million travelers should be possible in travel behavior research. The initial implementations thus combined concepts from computational physics and complex adaptive systems with concepts from travel behavior research. The third part of the book looks at theoretical concepts that are able to describe important aspects of the simulation system; for example, under certain conditions the code becomes a Monte Carlo engine sampling from a discrete choice model. Another important aspect is the interpretation of the MATSim score as utility in the microeconomic sense, opening up a connection to benefit cost analysis. Finally, the book collects use cases as they have been undertaken with MATSim. All current users of MATSim were invited to submit their work, and many followed with sometimes crisp and short and sometimes longer contributions, always with pointers to additional references. We hope that the book will become an invitation to explore, to build and to extend agent-based modeling of travel behavior from the stable and well tested core of MATSim documented here.
... However, doing both simultaneously, i.e. modelling conflicts at intersections while also simulating and modelling on-link congestion as well as route choice and traffic assignment of both bicycle and car traffic of an entire metropolitan area is. The study does this by extending the open-source agent-based transport simulator MATSim (Horni et al. 2016). As MATSim is already capable of simulating a large geographical area in a feasible time, the specific objective of this study is to replace the existing, simplistic intersection (node) model of MATSim with a detailed one that obeys multi-modal right-of-way at intersections and apply it to a large-scale case study with a large proportion of bicycle traffic. ...
... In this section we will present the right-of-way node model developed in this study for implementation in MATSim (Horni, 2016). 6 The reason for implementing the methodology in MATSim is threefold: firstly, because the software is open source; ...
... The mobility simulation in MATSim (Horni et al. 2016) consists of two separate models: a link model and a node model (Flötteröd 2016). The link model determines at link entry at which time a vehicle will be ready to leave the link again, at which point the vehicle is moved onto the buffer of the link. ...
Article
Full-text available
Intersections typically account for a substantial part of the total travel time in urban areas, and an even higher share of the congested travel time, especially for bicycle traffic. Nevertheless, delays caused by yielding for cyclists or cars at intersections have previously not been modelled in large-scale bicycle traffic assignment models. This study proposes a computationally efficient large-scale applicable methodology for explicitly modelling yielding for conflicting moves at multi-modal intersections in an agent-based traffic assignment model. Nodes representing the intersections are classified into five node types that simulate potential moves across nodes differently while obeying right-of-way and preventing simultaneous conflicting moves. The methodology is implemented within a joint assignment model capable of modelling on-link congestion of both car and bicycle traffic and is applied to a large-scale case study of a Metropolitan Copenhagen network with 144,060 nodes and 572,935 links. The MATSim case study with 4,593,059 trips shows manageable computation times similar to when not modelling right-of-way at intersections. Especially for car traffic, yielding at intersections imposes considerable excess travel time. The effects are larger for trips going to the central part of the city where the inter-modal impact of conflicting bicycle traffic is identified as a major source of added travel time. The study finds that failing to model conflicting moves at intersections generally underestimates travel times and causes too much traffic to go through the urban core, highlighting the importance of joint modelling of intersections.
... To this end, we propose a simulation framework that facilitates analyzing the potential demand of UAM. This framework interleaves a version of the agent-based simulation platform MATSim (which stands for multi-agent transport simulation [2]) for trip realization, which is extended to include UAM [3], and the agent-based demand model called the microscopic transportation orchestrator (MITO) [4]. ...
... The project's approaches and first results have already been published and described in more detail [6,30,31]. Additionally, the used methods strongly build upon prior work, such as research by Horni et al. [2], Rothfeld et al. [3], Moeckel et al. [4], Moeckel [5], Rothfeld et al. [20], Balac et al. [28], and Moreno and Moeckel [32]. The interested reader is referred to the respective publications for more details on the different methods because the scope of this paper does not allow us to describe each of them in detail. ...
... To estimate the potential UAM travel demand of the greater Munich area in the year 2030, we set up a simulation framework that integrates the agent-based and trip-based models (MITO) [4] and the multi-agent travel-based model (MATSim) [2]. MITO simulates and generates travel demand individually for every household given by the synthetic population of the SILO [5]. ...
... MATSim, which is an open-source framework for implementing large-scale agent-based transport simulations, is the main simulator in research work in this thesis. In MATSim, flow is investigated based on a macroscopic traffic flow model in the link-level, but with a queuing model: a vehicle entering a link will join the tail of the waiting queue (Horni et al., 2016). It will remain there unless all of the following three conditions are satisfied: 1) The time for traveling the link with free flow has passed. ...
... Flow-density relationship in MATSim(Horni et al., 2016) ...
... Many applications in the area of transportation system design rely on accurate modeling of human mobility demand. So-called mobility plans constitute a prominent representation of such demand, which is needed to design new mobility systems for growing and modern cities using, for example, state-of-the-art traffic simulators such as Multi-Agent Transport Simulation (MATSim) [1] or Simulation of Urban Mobility (SUMO) [2]. A set of mobility plans includes information on the coordinates of origins and destinations, as well as departure and arrival times for every individual trip occurring within a population, timeframe, and spatial environment. ...
... Three output requirements (Req 1 ) are imposed on the methodology to ensure that the generated mobility plans are qualitatively similar to those generated by full-scale activity-based models. 1 The generated methodology is implemented to accomplish the following: • Req 1a : generate microscopically consistent and feasible activity-based mobility plans; • Req 1b : achieve realistic macroscopic mobility behavior with regard to the order, frequency, location, time, and duration of activities as well as the induced travel distances; • Req 1c : differentiate by sociodemographic groups. ...
Article
Full-text available
Manifold applications in transportation system engineering rely on accurate modeling of human mobility demand. This demand is often represented by so-called mobility plans. Distinguished by their levels of aggregation, activity-based and trip-based models are the most prominent types of demand models in the literature. Macroscopic trip-based models are widely available but do not model mobility at the person level. In contrast, activity-based approaches simulate mobility microscopically but are complex and thus rarely available. The goal of this article is to present, apply, and validate an approach to generate activity-based mobility plans which microscopically reproduce real-world mobility demand but circumvent the complexity of activity-based approaches. To achieve this, existing trip-based models and mobility surveys are employed. Application results for car mobility in the city of Munich show that the obtained mobility plans are realistic on both a microscopic and a macroscopic level with regard to time, space, and activities. The presented approach can thus be considered appropriate for generating activity-based mobility plans whenever the development of a full-scale activity-based demand model is infeasible.
... In this study, the activity-based transport simulation framework MATSim is used. It requires a transport travel demand represented by a synthetic population which includes daily plans (each plan is a sequence of activities and legs), a transportation network, and a public transport scheduling [3]. ...
... The MATSim [3] process consists of three modules: mobsim, scoring and replanning. The mobility simulation module executes the selected plans of the whole population in parallel. ...
Article
Full-text available
For too long, many refined transportation models have focused solely on private and public transportation, assuming that bicycles only require simple models, such as bird flight distance or trips on horizontal tracks at a constant speed. This paper aims to study the impact of the road characteristics, such as road gradient, type of road and pavement surface of the road, on cyclists’ behavior using dedicated modules of MATSim. For that, we compare two approaches: a standard approach which does not consider the road characteristics, and a second approach that uses MATSim bicycle extension of Ziemke et al. The two approaches are analyzed over a sub-regional area around a district, focusing on a suburban city with an undulating relief made of average-to-steep hills. The focus is on the bicycle transportation model because the catchment area has a particularly challenging altitude profile and a large variety of roads, whether in type—from residential to national highway—or in pavement surface due to the number of green areas, such as parks and forests. This area is defined as a rather large 7 × 12 km, including five suburban cities in the South of Paris, France. A synthetic population of 126,000 agents was generated at a regional scale, with chains of activity made of work, education, shopping, leisure, restaurant and kindergarten, with activity-time choice, location choice and modal choice. We wanted to know how accurately a standard model of bicycle travels can be made with a 2D flat Earth assumption by comparing it to an algorithm extension that explicitly considers road characteristics in cyclists’ route choices. Our finding is that the MATSim bicycle extension model impacts mainly the long trips. Otherwise, the differences are minimal between the two models in terms of travel time and travel distance.
... Plans.xml provides information about agents and their daily mode-based travel plans between various activities. Network.xml describes the road network as a graph (nodes and links). The modular MATSim structure enable to additionally integrate information about the population, households, facilities, vehicles or public transit schedules and routes (Horni, et. al, 2016). ...
... In this paper, we use output files from a MATSim transport demand simulation in the city-state of Singapore (Horni et al., 2016). This MATSim simulation generates typical mobility patterns based on collected data from different sources between 2008 and 2014, like car ownership, public transport schedule info, the transportation network and agent activity types. ...
Conference Paper
Full-text available
Simulation models generate an abundance of rich raw data that remains difficult to access for non-experts. However, such data could be unlocked and utilised with a Semantic City Planning System that improves data accessibility and transparency. This paper describes a process of ontologically representing mobility simulation output data using Semantic Web technologies and storing it in a dynamic geospatial knowledge graph. Our work presents two benefits: 1) formally representing simulation output data increases the accessibility and transparency of urban simulation models, and 2) access to under-utilised rich data unlocks novel cross-domain knowledge explorations and research possibilities. We demonstrate these benefits by means of cross-domain queries related to typical city planning questions.
... MATSim is a popular open-source transportation simulation framework that focuses on travelers' adaptation to everyday traffic conditions via travel mode, route, and time (Horni et al., 2016). MATSim is suited for the task as it affects urban traffic at a resolution of individual travel within the explicit GIS-based urban environment. ...
... MATSim is suited for the task as it affects urban traffic at a resolution of individual travel within the explicit GIS-based urban environment. MATSim agents adapt to the existing traffic conditions through learning and self-correction (Horni et al., 2016). MATSim's advantage is in its ability to simulate traffic and transportation in realistic urban settings and analyze a vast urban system based on the simulation of a fraction of its population (Ben-Dor et al., 2020). ...
Article
Full-text available
Transportation Network Companies (TNC), like Uber, Lyft, and VIA, started their activities a decade ago with a far-reaching hope that Mobility-On-Demand (MOD) transportation services would decelerate or even stop the ever-growing congestion. However, it didn't happen; the negative incentives, like congestion charges and higher parking prices, seem to be the only policy tools for influencing congestion and associated negative externalities like pollution and noise. The question is whether we can establish socially acceptable congestion charges and parking prices that will effectively reduce the arrivals and traffic in highly congested areas and become the background for the future MOD arrangement? We employ the MATSim agent-based simulation model (Horni et al., 2016) of multi-modal traffic in Jerusalem Metropolitan Area (JMA) to address this problem. We investigate whether the combination of congestion and parking prices can force drivers to use Public Transport (PT), thus reducing arrivals with the private cars into the center of the city. The model study demonstrates that a reasonable charge of 7–12€ for entering the city center could decrease arrivals by 25%. From the transport policy point of view, the effects of congestion charges and parking prices are different – the increase in the congestion charges decreases arrivals. In contrast, the increase in parking prices decreases the dwell time. We discuss the policy consequences of employing each of the two mechanisms.
... There are various tools for simulating people flow. For example, the open-source MatSim [30] is a multi-agent framework designed to simulate transport networks, supporting several transportation modes. Furthermore, SUMO [31] is an opensource traffic simulation package for demand modelling which considers various traffic management topics. ...
... The discussed tools, while focus on a single mobility operator (except [33]), do not support what-if analysis when creating alternative scenarios to be compared with Actual Scenario. Additionally, the user cannot access the analysis results using a web-based interface that supports GTFS data feed (except [30,33]) and REST APIs. ...
Article
Full-text available
A main key success for public transportation networks is their tuning by the analysis of mobility demand with respect to the offer in terms of public transportation means. Most of the solutions at the state of the art have strong limitations in taking into account: multiple contextual information as attractors/motivations for people movements, modalities of travel means, multiple operators, and a range of key performance indicators. For these reasons, a model for analyzing the demand with respect to the offer of mobility has been studied, and the corresponding tool DORAM developed. DORAM allows to perform the analysis of alternative scenarios, as what-if analyses, when the transport service offer and the mobility demand changed in the scenario, adopting a fast-computation strategy to compare scenarios with the aim of detecting/identifying motivations of crowded conditions on stops and on the vehicles. The analysis can exploit a wide range of data sources when computing a set of key performance indicators. The DORAM solution has been defined and developed in the MOSAIC research and development project with ALSTOM and other companies. The DORAM solution is validated by using real data and conditions in the Tuscany region.
... Agent-based models (ABMs) are a tool that allow the simulation of the dynamic processes of behaviour change at the population level (the model) by accounting for interactions between heterogeneous individuals (the agents) and their environment [13]. While there exist ABMs exploring transport systems at a low-level [14], there do not exist models that investigate the impact of social norms on transport behaviour. The existing work is focussed mainly on traffic flow and route decisions [15,16,17]; our work focusses on the decision for the method of commuting and is not focussed on modelling the flow of people through an environment. ...
... Our model differs from existing models of transport [14,15,16,17]; it is not a traffic simulation -we do not aim to model traffic flow through a city. The intended purpose is not to make accurate predictions of how people commute to work. ...
Conference Paper
Interventions to increase active commuting have been recommended as a method to increase population physical activity, but evidence is mixed. Social norms related to travel behaviour may influence the uptake of active commuting interventions but are rarely considered in their design and evaluation. In this study we develop an agent-based model that incorporates social norms related to travel behaviour and demonstrate the utility of this through implementing car-free Wednesdays. A synthetic population of Waltham Forest, London, UK was generated using a microsimulation approach with data from the UK Census 2011 and UK HLS datasets. An agent-based model was created using this synthetic population which modelled how the actions of peers and neighbours, subculture, habit, weather, bicycle ownership, car ownership, environmental supportiveness, and congestion (all configurable parameters) affect the decision to travel between four modes: walking, cycling, driving, and public transport. The developed model (MOTIVATE) is a configurable agent-based model where social norms related to travel behaviour are used to provide a more realistic representation of the socio-ecological systems in which active commuting interventions may be deployed. The utility of this model is demonstrated using car-free days as a hypothetical intervention. In the control scenario, the odds of active travel were plausible at 0.091 (89% HPDI: [0.091, 0.091]). Compared to the control scenario, the odds of active travel were increased by 70.3% (89% HPDI: [70.3%, 70.3%]), in the intervention scenario, on non-car-free days; the effect is sustained to non-car-free days. While these results demonstrate the utility of our agent-based model, rather than aim to make accurate predictions, they do suggest that by there being a ‘nudge’ of car-free days, there may be a sustained change in active commuting behaviour. The model is a useful tool for investigating the effect of how social networks and social norms influence the effectiveness of various interventions. If configured using real-world built environment data, it may be useful for investigating how social norms interact with the built environment to cause the emergence of commuting conventions
... • MATSim [6,7]; ...
Article
The preprint is devoted to the development of the interactive program for traffic flow modeling, which can be used in the educational process. A brief overview of currently existing software solutions for transport modeling, as well as products to support computer learning methods, is given. The original models of macro- and microscopic types, which form the basis of the educational platform, are described. Particular attention is paid to the development of tools for dynamic interpretation of simulation results using the Unity environment, which is integrated with Microsoft Visual Studio.
... In the present paper, transportation equity is studied with a sophisticated simulation model system that incorporates hydrodynamic simulations using the USGS Coastal Storm Modeling System (CoSMoS) (Deltares, 2015) and the Multi-Agent Transport Simulation (MATSim) (Horni et al., 2016) calibrated for the San Francisco Bay Area. The transportation system in the Bay Area MATSim consists of freeways, arterials, secondary roads, and the Bay Area Rapid Transit (BART) with the workday schedule. ...
Article
Full-text available
Transportation equity is an essential aspect of urban transportation planning. With climate change becoming inevitable, coastal cities are considering the mitigation of the impact of sea level rise on infrastructure. Transportation equity and sea level rise adaptation are usually considered separately. However, research pointed out that these two challenges could have considerable overlap and interaction. The present paper discusses transportation equity issues resulting from the impact of sea level rise and associated protection strategies. A case study of the San Francisco Bay Area points out cases where transportation equity can be negatively impacted when the optimal protection strategy against sea level rise is implemented. An integrated hydrodynamic and transportation model system is used in the present paper to demonstrate several scenarios where the most efficient protection strategies for the whole region increase the inequity that exists between the disadvantaged communities and other communities. Nevertheless, this impact can be mitigated with a relatively small addition to the protection strategy. The paper suggests that transportation equity cannot be overlooked in planning climate adaptation, as even the protection plan that maximizes benefits for the region may negatively impact the most vulnerable communities.
... As shown in Figure 1, BEAM [27] integrates MATSim [13] toolkit and therefore possess an agent-based modeling approach. Each agent in BEAM employ reinforcement learning across successive simulated days to maximize their personal utility through plan mutation (exploration) and selecting between previously executed plans (exploitation). ...
Preprint
Full-text available
MATSim (Multi-Agent Transport Simulation Toolkit) is an open source large-scale agent-based transportation planning project applied to various areas like road transport, public transport, freight transport, regional evacuation, etc. BEAM (Behavior, Energy, Autonomy, and Mobility) framework extends MATSim to enable powerful and scalable analysis of urban transportation systems. The agents from the BEAM simulation exhibit 'mode choice' behavior based on multinomial logit model. In our study, we consider eight mode choices viz. bike, car, walk, ride hail, driving to transit, walking to transit, ride hail to transit, and ride hail pooling. The 'alternative specific constants' for each mode choice are critical hyperparameters in a configuration file related to a particular scenario under experimentation. We use the 'Urbansim-10k' BEAM scenario (with 10,000 population size) for all our experiments. Since these hyperparameters affect the simulation in complex ways, manual calibration methods are time consuming. We present a parallel Bayesian optimization method with early stopping rule to achieve fast convergence for the given multi-in-multi-out problem to its optimal configurations. Our model is based on an open source HpBandSter package. This approach combines hierarchy of several 1D Kernel Density Estimators (KDE) with a cheap evaluator (Hyperband, a single multidimensional KDE). Our model has also incorporated extrapolation based early stopping rule. With our model, we could achieve a 25% L1 norm for a large-scale BEAM simulation in fully autonomous manner. To the best of our knowledge, our work is the first of its kind applied to large-scale multi-agent transportation simulations. This work can be useful for surrogate modeling of scenarios with very large populations.
... To study the deployment of charging stations for EVs we integrate the Adaptive Epsilon Sampling Epsilon Hood ( ) MOEA [2,4] with electric vehicle mobility simulation by the Multi-Agent Traffic Simulator (MATSim) [1]. ...
Conference Paper
This article reports using a bi-objective evolutionary algorithm interacting with a traffic simulator and data exploration methods to analyze the optimal capacity and location of charging infrastructure for electric vehicles. In this work, the focus of the study is the city of Cuenca, Ecuador. We configure a scenario with 20 candidate charging stations and 500 electric vehicles driving according to the mobility distribution observed in this city. We optimize the vehicle's travel time that requires recharging and the number of charging stations distributed in the city. Quality of Service is defined as the ratio of charged vehicles to vehicles waiting for a charge and is considered a constraint. The approximate Pareto set of solutions produced in our experiments includes a number of trade-off solutions to the formulated problem and shows that the evolutionary approach is a practical tool to find and study different layouts related to the location and capacities of charging stations. In addition, we complement the analysis of results by considering Quality of Service, charging time, and energy to determine the city's best locations. The proposed framework that combines simulated scenarios with evolutionary algorithms is a powerful tool to analyze and understand different charging station infrastructure designs.
... ALBATROSS utilizes machine learning, including a series of decision trees, to represent the choice heuristics of travelers and then calculate these heuristics from activity-travel data. Activity-based models have been developed and utilized in combination with one of the leading micro-simulation tools, MATSim (Axhausen 2016), in order to study transportation demand. MATSim is an open-source transport simulation tool that can serve as a link between large-scale agent-based network simulations and activitybased models. ...
Article
Full-text available
In this paper, a heuristic method which contributes to the solution of the Daily Activity Chains Optimization problem with the use of Electric Vehicles (DACO-EV) is presented. The DACO-EV is a time-dependent activity-scheduling problem of individual travelers in urban environments. The heuristic method is comprised of a genetic algorithm that considers as its parameters a set of preferences of the travelers regarding their initial activity chains as well as parameters concerning the transportation network and the urban environment. The objective of the algorithm is to calculate the traveler’s optimized activity chains within a single day as they emerge from the improved combinations of the available options for each individual traveler based on their flexibility preferences. Special emphasis is laid on the underlying speed-up techniques of the GA and the mechanisms that account for specific characteristics of EVs, such as consumption according to the EV model and international standards, charging station locations, and the types of charging plugs. From the results of this study, it is proven that the method is suitable for efficiently aiding travelers in the meaningful planning of their daily activity schedules and that the algorithm can serve as a tool for the analysis and derivation of the insights into the transportation network itself.
... This analysis is part of an approach to determine system-wide and long-term effects of AVs from local microscopic observations. For this purpose, the microscopic simulation "SUMO" [18] is integrated in a framework with the mesoscopic agent-based mobility-simulation "MATSim" [19] to derive realistic flow capacities for traffic signal-controlled intersections for different shares of CV and AV. This framework is applied to systematically adjust and parameterize the effect of future infrastructure and vehicle technologies in the network of the greater Düsseldorf area in Germany, which is part of the German national test field for automated and connected driving 3 . ...
Article
In this paper, we assess the effects of different shares of autonomous vehicles (AVs) on the traffic flow and, in particular, on the maximum possible capacity at signal-controlled intersections. For this purpose, all signal-controlled nodes in the traffic network of the Düsseldorf metropolitan area were systematically simulated and evaluated using the microscopic traffic simulation tool SUMO.The analysis shows that defensively parameterized AVs – as envisaged in the umbrella project of this research – may decrease the maximum possible traffic at signal-controlled intersections. Moreover, the simulation runs indicate that capacity at these intersections decreases almost linearly with a growing share of AV. In a second part of this analysis, a freeway section was simulated with the same varying shares of CV and AV to investigate free-flow traffic. In this case, the simulation results of the maximum traffic flow can be approximated by a third-order polynomial fit. The minimum capacity is found for the uniform share of both vehicle types (i.e. 50 % AV and 50 % CV).The overall intent of this project is to provide an approach to determine system-wide and long-term effects of AVs from local microscopic observations. To this end, the SUMO microscopic traffic simulation will be utilized to derive realistic flow capacities for signal-controlled intersections. In a next step, these capacities will be transferred to a mesoscopic traffic simulation. Subsequently, flow capacities can be systematically adjusted in this network-wide mobility simulation to parameterize the influence of future infrastructure and vehicle technologies.
... To do so, Microsimulation Analysis for Network Traffic Assignment (MANTA), an ultra-fast, highly-parallelized GPU-based microsimulation platform, is developed [7]. Existing simulators have typically revealed a tradeoff among accuracy, computational speed, and geographic scale of simulation [7,[9][10][11][12]. However, MANTA exhibits performance benefits in all three of these areas, enabling it to be used for agile scenario planning, particularly with an emergent mode such as UAM, whose deployment is still in its inchoate stages. ...
Conference Paper
View Video Presentation: https://doi.org/10.2514/6.2022-3837.vid Over the past several years, Urban Air Mobility (UAM) has galvanized enthusiasm from investors and researchers, marrying expertise in aircraft design, transportation, logistics, artificial intelligence, battery chemistry, and broader policymaking. However, two significant questions remain unexplored: (1) What is the value of UAM in a region’s transportation network? and (2) How can UAM be effectively deployed to realize and maximize this value to all stakeholders, including riders and local economies? To adequately understand the value proposition of UAM for metropolitan areas, the authors develop a holistic multi-modal toolchain, SimUAM, to model and simulate UAM and its impacts on travel behavior. This toolchain has several components: (1) Microsimulation Analysis for Network Traffic Assignment (MANTA): A fast, high-fidelity regional-scale traffic microsimulator, (2) VertiSim: A granular, discrete-event vertiport and pedestrian simulator, (3) Flexible Engine for Fast-time Evaluation of Flight Environments (Fe$^3$): A high-fidelity, trajectory-based aerial microsimulation. SimUAM, rooted in granular, GPU-based microsimulation, models millions of trips and their movements in the street network and in the air, producing interpretable and actionable performance metrics for UAM designs and deployments. Once the ground-air interface is modeled, the authors find that the market for UAM decreases across all network designs relative to models with static assumptions about transfer times. However, significant improvements can be made to balance the demand and optimize the networks for transfer time, likely increasing the number of benefited trips. The modularity, extensibility, and speed of the platform will allow for rapid scenario planning and sensitivity analysis, effectively acting as a detailed performance assessment tool.
... MATSim is an open-source framework 6 for large-scale agent-based transport simulations (Horni et al., 2016). It is capable of simulating very large networks and populations while maintaining a relatively high level of detail. ...
Article
Full-text available
In rural areas with a low population density, demand-responsive transport (DRT) is considered a promising alternative to conventional public transport. With a fleet of smaller vehicles, DRT provides a much more flexible and convenient service. This characteristic makes the DRT also a potential mode of transport to serve the school children in rural areas. If the DRT vehicles are used to serve the school children, then the funding for conventional school buses (or adapted public transport schedules) can be reinvested in the DRT system. This may help to relieve the financial burden experienced by the DRT operators and enable the operation of a large-scale DRT service in rural areas. In this study, a demand model for school commutes based on real-world, open-source data for Landkreis Vulkaneifel, a rural region in Germany, is built. Then a feasibility study is carried out using an agent-based transport simulation. In the feasibility study, various setups and operational schemes are explored, which is followed by a systematic cost analysis. Results from the simulations show that an annual budget of 1617 Euro per student is sufficient to maintain and operate a fleet of DRT vehicles that can transport all the students in the region from home to school on time in the morning. During the remaining time of the day and on school holidays, the vehicles can be used for conventional DRT service for the public.
... Notably, MaaSSim is not intended for the complete modelling of transport systems for which there is an abundance of mature and developed frameworks, both commercial (like PTV Visum, CUBE, Emme) and open source (like MatSim [45], SUMO [46], DynaMIT [47], SimMobility [48] etc.). Instead, the explicit objective of MaaSSim is to support researchers with modelling and reproducing the emerging novel phenomena taking place in the context of two-sided mobility platforms and analyse their disruptive potential for urban transport systems. ...
Article
Full-text available
Two-sided mobility platforms, such as Uber and Lyft, widely emerged in the urban mobility landscape. Distributed supply of individual drivers, matched with travellers via intermediate platform yields a new class of phenomena not present in urban mobility before. Such disruptive changes to transportation systems call for a simulation framework where researchers from various and across disciplines may introduce models aimed at representing the complex dynamics of platform-driven urban mobility. In this work, we present MaaSSim, a lightweight agent-based simulator reproducing the transport system used by two kinds of agents: (i) travellers, requesting to travel from their origin to destination at a given time, and (ii) drivers supplying their travel needs by offering them rides. An intermediate agent, the platform, matches demand with supply. Agents are individual decision-makers. Specifically, travellers may decide which mode they use or reject an incoming offer; drivers may opt-out from the system or reject incoming requests. All of the above behaviours are modelled through user-defined modules, allowing to represent agents’ taste variations (heterogeneity), their previous experiences (learning) and available information (system control). MaaSSim is a flexible open-source python library capable of realistically reproducing complex interactions between agents of a two-sided mobility platform. MaaSSim is available from a public repository, along with a set of tutorials and reproducible use-case scenarios, as demonstrated with a series of illustrative examples and a comprehensive case study.
... However, most of the simulators are for a business purpose, and therefore provide very few design details, such as Aimsun (Casas et al., 2010a), PTV Vissim (Fellendorf and Vortisch, 2010a), Paramics (Cameron and Duncan, 1996). Some open source alternatives exist also, such as SUMO (Lopez et al., 2018), MATSim (Horni, Nagel, and Axhausen, 2016) and MITSIMLab (M. Ben-Akiva et al., 2010). ...
Thesis
There is a growing interest in autonomous driving as it is expected that fully autonomous vehicles can reduce car accidents and improve overall traffic safety. However, autonomous driving is a complex process combining sensing, perception, prediction, computation, and decision. In addition, the traffic environment is dynamic and involves interactions among road users. Therefore, driving tests are essential to validate the autonomous vehicle's functionalities. Real-world driving tests seem to be a great challenge as fatal accidents cannot be prevented yet. Alternatively, performing driving tests by simulation can reduce time and cost, and avoid potentially dangerous situations. The increasing use of traffic simulation for many studies highlights the importance of a good understanding and modeling of human driving behavior.This thesis mainly focuses on microscopic traffic modelling for human driving models, with the aim of creating, with numerical simulation, a realistic vehicular traffic, which is useful for the validation of autonomous vehicle's features.The main contributions of this thesis consist in :1. Car-collision generation in numerical traffic simulation: I proposed an approach of car-collision generation in numerical traffic simulation considering different car-following behaviors. After the investigation of different driver profiles in a real traffic data-set, I classified three driving profiles, where I distinguished aggressive and inattentive driver profiles from the normal profile. I then proposed to increase the proportion of the two ‘extreme’ driver profiles (aggressive and inattentive) in the whole traffic population by replacing the normal drivers, to simulate in a traffic simulator, SUMO (Simulation of Urban Mobility), and observe eventually the occurrence of car-collisions. I was able to formulate a relationship between the ratios of these two driver profiles over the entire driver population, and the number of car collisions. This analysis used part of the NGSIM 101 data-set and was validated on another part of the same data-set. I also studied the severity of the generated collisions. I found that collisions involved between an inattentive driver as the leader and an aggressive driver as the follower are the most frequent ones, while collisions between two inattentive drivers are the severest ones.2. Lane change modeling using reinforcement learning: The second work in my PHD is on the lane change modeling, where a reinforcement learning model has been developed. The model aims to imitate real lane change decisions, based on the NGSIM traffic data-set. I proposed a Q-learning model for the human lane change decisions. The model shows good performances in mimicking human decisions with up to 95% of success. Moreover, the model uses numerical traffic simulation (SUMO) to complete the unknown situations in the real data-set. We observed that 13% additional traffic conditions were created by the traffic simulation environment.3. LSTM neural network for human driving behavior: In the third work of my PHD, I proposed an LSTM neural network model for car-following and lane-changing behaviors modeling on road networks. In this work, I proposed different models with different input designs and compared them. The selected model shows good performances on both predicting the longitudinal speed and the lateral position of cars. Moreover, the obtained results show that the selected model outperforms the classical IDM (Intelligent Driver Model) in the accuracy of replicating car-following behavior. The models were implemented on the NGSIM 101 and the HighD traffic data-sets
... A time-dependent dynamic traffic assignment router (Verbas et al., 2018) routes vehicles whose experienced travel time is an outcome of a mesoscopic traffic flow model based on the link transmission model (de Souza et al., 2019). POLARIS exhibits finer link-level traffic behavior than queue-based algorithm approaches (Horni et al., 2016). The region's population is not downsampled, and time-dependent background traffic, such as freight and other external travel, is added to links to add increased realism. ...
Article
Full-text available
The emergence of on-demand shared autonomous electric vehicle (SAEV) service requires careful charging station planning and a joint charging and repositioning strategy to mitigate empty travel. This study couples charging and repositioning events as a means of improving service quality (rider wait times), reducing empty travel due to repositioning or charging, and improving fleet utilization (average daily trips per vehicle and charging queues). This synergy is explored for the Austin, Texas region using POLARIS, an agent-based model. On average, wait times were 39% lower, and average daily trips served per SAEV increased up to 6.4 (or 28%) compared to SAEV repositioning with heuristic charging. Coupling repositioning with charging decreased the fleet's percent empty travel on average by 1.6%, relative to the scenario treating them as independent events (which varies by charging station design). Sparser charging stations reduce investment costs, and operators can leverage this framework to keep traveler wait times low. Free 50-day access link (expires July 17, 2022): https://authors.elsevier.com/c/1f92w4rgZinA0C
... Melbourne activity-based model (MABM) (Victoria, 2017) is developed by Infrastructure Victoria in the partnership with KPMG and Arup to simulate the activities and travel patterns in the large-scale Melbourne metropolitan network. The MABM utilizes the open source platform known as Multi-agent transport simulation (MATSim) which was developed incrementally since in the early 1990's (Horni et al., 2016). ...
Article
Coronavirus 2019 (COVID-19) and its variants are still spreading rapidly with deadly consequences and profound impacts on the global health and world economy. Without a suitable vaccine, mobility restriction has been the most effective method so far to prevent its spreading and avoid overwhelming the heath system of the affected country. The compartmental model SIR (or Susceptible, Infected, and Recovered) is the most popular mathematical model used to predict the course of the COVID-19 pandemic in order to plan the control actions and mobility restrictions against its spreading. A major limitation of this model in relation to modeling the spreading of COVID-19, and the mobility limitation strategy, is that the SIR model does not include mobility or take into account changes in mobility within its structure. This paper develops and tests a new hybrid SIR model; SIR-M which is integrated with an urban activity travel model to explore how it might improve the prediction of pandemic course and the testing of mobility limitation strategies in managing virus spread. The paper describes the enhanced methodology and tests a range of mobility limitation strategies on virus spread outcomes. Implications for policy and research futures are suggested.
... Several works have been developed to model BSS using GAMA (Kaziyeva et al., 2021;Lu et al., 2019;Veldhuis, 2018). MATSim (Horni et al., 2016) is an activity-based, extendable, multi-agent simulation framework implemented in Java that performs an integral microscopic simulation of resulting traffic flows and the congestion they cause. MATSim employs a microscopic description of demand by tracing the daily schedule and travelers' decisions. ...
Article
Full-text available
Fast urbanization and climate change require innovative systems for an efficient movement of people and goods in cities. As trends towards vehicle-sharing, autonomous vehicles, and the use of micro-mobility systems gain strength, the intersection of these fields appears as an area of great opportunity. Autonomy could potentially bring the convenience of on-demand mobility into already prevalent shared micro-mobility systems (SMMS), increasing their efficiency and incentivizing more people to use active mobility modes. The novelty of introducing autonomous driving technology into SMMS and their inherent complexity requires tools to assess and quantify the potential impact of autonomy on fleet performance and user experience. This paper presents an ad-hoc agent-based simulator for the assessment of the fleet behavior of autonomous SMMS in realistic scenarios, including a rebalancing system based on demand prediction. It also allows comparing its performance to station-based and dockless schemes. The proposed simulation framework is highly configurable and flexible and works with high resolution and precision geospatial data. The results of studies carried out with this simulation tool could provide valuable insights for many stakeholders, including vehicle design engineers, fleet operators, city planners, and governments.
... For example, Chow (2017a, 2017b) modeled the ridesharing system as an agent-based equilibrium problem with a day-to-day learning process. Wang et al. (2017) implemented an agent-based dynamic ridesharing model in the MATSim platform (Horni et al., 2016), with explicit consideration that drivers can switch their roles as drive alone if they would be late for their next activities. Beojone and Geroliminis (2021) proposed a simulation framework that accounts for passenger cancelation before assigning a driver. ...
... Activity-based models (ABMs) simulate and forecast daily activity tours of the population at an urban scale, which comprise multiple hierarchical dimensions of individual-level preferences in continuous time and spacewhen, where, for how long, in what sequence, and by which travel modes activities are performed (1,2). The synthetic population is the key input to ABM, which aims to mimic the actual population's joint distribution of individual-and household-level attributes. ...
Preprint
Full-text available
An ideal synthetic population, a key input to activity-based models, mimics the distribution of the individual- and household-level attributes in the actual population. Since the entire population's attributes are generally unavailable, household travel survey (HTS) samples are used for population synthesis. Synthesizing population by directly sampling from HTS ignores the attribute combinations that are unobserved in the HTS samples but exist in the population, called 'sampling zeros'. A deep generative model (DGM) can potentially synthesize the sampling zeros but at the expense of generating 'structural zeros' (i.e., the infeasible attribute combinations that do not exist in the population). This study proposes a novel method to minimize structural zeros while preserving sampling zeros. Two regularizations are devised to customize the training of the DGM and applied to a generative adversarial network (GAN) and a variational autoencoder (VAE). The adopted metrics for feasibility and diversity of the synthetic population indicate the capability of generating sampling and structural zeros -- lower structural zeros and lower sampling zeros indicate the higher feasibility and the lower diversity, respectively. Results show that the proposed regularizations achieve considerable performance improvement in feasibility and diversity of the synthesized population over traditional models. The proposed VAE additionally generated 23.5% of the population ignored by the sample with 79.2% precision (i.e., 20.8% structural zeros rates), while the proposed GAN generated 18.3% of the ignored population with 89.0% precision. The proposed improvement in DGM generates a more feasible and diverse synthetic population, which is critical for the accuracy of an activity-based model.
... The mentioned approaches consequently help to identify what level of intervention is required to move the system towards its set goals. Once suitable parameters have been identified, they can be tested in various scenarios in an agent-based and activity-orientated mobility simulator, e.g., MATSim (Horni et al., 2016). These scenarios can also test different pricing schemes and initial budget allocation scenarios. ...
Article
Full-text available
Economic instruments are considered promising to achieve the urgently needed reduction of greenhouse gas emissions in the transportation sector. In this context, tradable credit schemes have received more and more attention in recent years. These cap-and-trade systems have the primary goal of limiting congestion, but they can also incorporate emission reduction goals. In this paper, we present the conceptual extension of a tradable credit scheme from a congestion and emission-oriented to a holistic, full-trip, multi-modal mobility traffic management system. In addition to the demand it also includes the management of the supply side. The integration of all existing modes into one holistic scheme ensures that the overall system goals are reached as all behavioral responses remain within the system boundaries. The system comprises two key innovations. First, a central agency has the possibility to provide full-trip incentives across modes to support the overall system’s goal. Second, users of the system can spend parts of their allocated mobility budget for transport infrastructure upgrades as an addition to paying for mobility or monetizing it on the market. Those innovations are a distinct extension to the idea of tradable credits. Commonly used smartphones would serve as the enabling technology of the proposed system. They offer all technical requirements and almost every citizen has access to one. Smartphones are affordable compared to dedicated traffic management infrastructure and they are flexible to accommodate system changes, e.g., new modes, through software updates. Besides the potential technical implementation, overall design questions, social aspects as well as general implications of the concept are covered.
... The high-fidelity transport simulator AMoDeus [24] was used for contrasting and benchmarking of the AMoD algorithm (see Algorithm 1). AMoDeus is an open-source agentbased transport simulator based on Multi-Agent Transport Simulator (MATSim) [32]. It was intentionally developed to simulate AMoD systems and to test new algorithms for fleet control. ...
Preprint
This paper considers the problem of supply-demand imbalances in Autonomous Mobility-on-Demand systems (AMoD) where demand uncertainty compromises both the service provider's and the customer objectives. The key idea is to include estimated stochastic travel demand patterns into receding horizon AMoD optimization problems. More precisely, we first estimate passenger demand using Gaussian Process Regression (GPR). GPR provides demand uncertainty bounds for time pattern prediction. Second, we integrate demand predictions with uncertainty bounds into a receding horizon AMoD optimization. In order to guarantee constraint satisfaction in the above optimization under estimated stochastic demand prediction, we employ a probabilistic constraining method with user defined confidence interval. Receding horizon AMoD optimization with probabilistic constraints thereby calls for Chance Constrained Model Predictive Control (CCMPC). The benefit of the proposed method is twofold. First, travel demand uncertainty prediction from data can naturally be embedded into AMoD optimization. Second, CCMPC can further be relaxed into a Mixed-Integer-Linear-Program (MILP) that can efficiently be solved. We show, through high-fidelity transportation simulation, that by tuning the confidence bound on the chance constraint close to "optimal" oracle performance can be achieved. The median wait time is reduced by 4% compared to using only the mean prediction of the GP.
... However, DaySim does not specify the routes for trips; hence, other traffic simulation models are needed to model these. A popular choice is the Multi-Agent Transport Simulation (MATSim) developed by a group of European scholars (Axhausen et al., 2016). The individual travel itineraries from DaySim can be fed into MATSim, along with the transportation network characteristics, to assign routes for each trip from individual travel itineraries. ...
Preprint
Full-text available
Big data can help government agencies better make equity-related decisions in transportation systems. However, the lack of disaggregated big data is an obstacle to inform policy actions. In this paper, we review data sources that can be used to investigate transportation equity. Following the general three-step framework for transportation equity outcomes analysis, we categorized the identified data sources into a taxonomy with four broad data categories: population data, transportation infrastructure data, mobility data, and other data (facility data, traffic accident data, traffic-related air pollution data). Sources to collect these four types of data for the United States were identified. Representative studies from the literature were reviewed to reveal how data collected from the identified sources have been or could be used to characterize transportation outcomes equity. Here we call for efforts to construct a transportation equity big data library, specifying the associated technological, epistemological, methodological, and political challenges. This paper will offer an important reference where government agencies and transportation researchers can seek data to improve transportation equity.
... A pivotal example of the former one is the traditional 4-step model, while in the case of the latter are the agent-based models (e.g. Horni et al. (2016)). Obviously, a microscopic demand model requires much more detailed data on a person level and the development of many statistical sub-models, increasing considerably the computational effort to reach the equilibrium point. ...
Thesis
Full-text available
Travel demand models have increased their data demands massively both in scope and scale as they have become more complex over the course of years. Against that trend, the current dissertation pursues the development of a direct demand modelling approach tailored for speed and volume prediction purposes. In this regard, the main objective of this dissertation is twofold. First, to investigate how the predictive accuracy of a direct demand model can be enhanced if viewed through the lens of a spatial perspective, and second to identify and resolve the statistical shortcomings that arise due to the spatial nature of data. Methodologically, the family of spatial regression models is exploited while the issue of endogeneity governing the relationship between the two modelled transport phenomena is taken into consideration. On the spatial interaction side, a new framework that revisits the definition of the distance decay function is introduced. Subsequently, this revision is translated into a series of modified accessibility measures. Furthermore, a new indicator that combines the concepts of centrality and gravity-based accessibility in a unified measure is introduced. This indicator provides a richer picture of the ways a transport system generates connectivity and how accessibility is jointly generated by the network and the landscape of opportunities. In addition, the new centrality indicator is thoroughly tested for its ability to improve the predictive accuracy of direct demand models. Finally, a comparison of the results of the developed modelling approach against the output of a traditional four-step model showcases that direct demand models can provide a trustworthy alternative to more advanced, but definitely more data demanding and computationally burdensome approaches. Especially in cases where the development of more advanced models is not possible, either due to data availability issues, or due to various limitations in place, a direct demand model can constitute a viable alternative.
... org/). A general reference for this simulation framework isHorni et al. (2016). 2 https:// geose rvices. ...
Article
Full-text available
We extend a multimodal transport model to simulate an increase of the market share of electric vehicles. The model, which is described in detail in Kilani et al. (Sustainability 14(3):1535, 2022), covers the north of France and includes both urban and intercity trips. It is a multi-agents simulation based on the MATsim framework and calibrated on observed traffic flows. We find that the emissions of pollutant gases decrease in comparable proportion to the market share of the electric vehicles. When only users with shorter trips switch to electric vehicles, the impact is limited and demand for charging stations is small since most users will charge by night at home. When the government is able to target users with longer trips, the impact can be higher by more than a factor of two. But, in this case, our model shows that it is important to increase the number of charging stations with an optimized deployment for their accessibility.
... This entire chapter cites the book The Multi-Agent Transport Simulation MATSim(Horni, Nagel, and Axhausen, 2016) ...
Technical Report
Full-text available
Addressing congestion in large cities is a major concern for traffic management centres, especially when new transportation modes emerge rapidly. Lately, connected and autonomous vehicles have gain a lot of popularity due to the promise of providing a new flexible transportation mode, which could be shared, real-time, electric and without the need of using any parking slots due to the continuous movement of cars. Future predictions foresee a considerable increase of the effective road capacity through intelligent and inter-connected vehicles. On the flip-side, more vehicles might be observed on the road if the autonomous vehicles get so comfortable, cheap and available that aggregated transit forms such as buses or trains become obsolete. But evaluating the impact of injecting a new fleet of autonomous vehicles in the existing road network is quite challenging as many unknown factors need to be taken into consideration, especially if they will run in a taxi-mode environment. Multi-agent simulation seems like an adapted tool to be able to evaluate the behaviour of such vehicles, the variation of traffic demand, the network response to incidents, etc. This report reviews the state-of-the-art multi-agent transport simulation frameworks along with presenting the methods, results and limitations of simulating a real-world urban dynamics using Victoria Road, Sydney as the case study.
... Various approaches exist for transport modelling and assessment. These comprise four-step models (Ortúzar and Willumsen, 2011), activity-based approaches (Horni et al., 2016), landuse and transport interaction (LUTI) models (Waddell, 2002), urban spatial general equilibrium models (Anas and Kim, 1996), and hybrids of these. These models have different aims and purposes. ...
Thesis
Full-text available
This dissertation makes use of urban spatial computable general equilibrium (USCGE) models tailored towards the application to urban transport in general and urban air transport specifically. The thesis assesses the effects of urban air mobility (UAM) introduction and hopes to contribute to the young field of research by providing evidence on possible long-run effects of transport drones. In recent years there has been a strong increase in research output in the field of UAM. The studies on the topic are wide-spread discipline-wise, reaching from vehicle design over UTM, regulation and certification to acceptance and adoption. Not only research, but also business activities are strongly increasing. Yet, UAM still faces technical, infrastructural and societal hurdles on the way to introduction. One of the main hurdles is the support of authorities, policy makers and the public. To enable a supporting environment early on, it is important to provide tools and methods that enable an assessment of the long-run effects of transport drones. Making use of USCGE models, this thesis broadens the discussion on UAM impacts to include also welfare effects, environmental aspects, and differentiate between the impacts on different parts of society. Applying a USCGE model to an existent transportation issue, namely parking, gives confidence in the chosen method. Tailoring the model to UAM and incorporating both high- and low-skilled households enables us to derive several interesting findings. Using agglomeration effects and amenities the model allows to differentiate between cities where high-skilled locate close to the city-centre and cities where high-skilled rather move to the suburbs. Differentiating between both initial spatial structures shows that the impact that the city structure has on the impact of UAM introduction, is minor. UAM system characteristics, like, land demand, prices, marginal cost or travel speed, in contrast significantly impact direction and magnitude of welfare effects. We find that the welfare effects for households with different income levels strongly differ and hence want to emphasise the relevance of understanding the differential impacts of UAM on user and non-users. Expanding the assessment to also include electric ground mobility and explicitly considering the environmental effects of UAM introduction shows that differences in taxation between gasoline and electricity lead to welfare losses when a forced transition from gasoline cars to electric cars takes place, while CO2 emissions go down. The higher tax on gasoline compared to electricity, as it is currently in place in Germany, results in a better internalization of otherwise untackled congestion externalities and hence explains this somewhat unexpected effect. The model also provides evidence, that introducing UAM as a substitute for gasoline cars has the potential to reduce CO2 emissions, whereas serving as an alternative to electric cars UAM usage increases CO2 emissions due to higher energy demand. Drones can also be used for cargo transport. In order to understand these effects as well the USCGE model is adapted to model different retail channels (local shops, online shopping and delivery via drone and online shopping with delivery via truck) and the logistic structures behind them. The assessment shows that additional retail channel choice options increase welfare and that the rise of e-commerce could significantly impact location choices in cities. This research shows that especially the long-run impact of passenger and cargo drones on users as well as non-users need consideration when assessing promising applications. From an environmental perspective, it is essential to identify applications that either allow to save energy due to shorter routes (e.g. due to geographical barriers), or justify the additional energy use due to the value added by the service (e.g. emergency applications or generating parity in living conditions).
... (Zwick and Axhausen, 2020) claims it is difficult to compare different pooling strategies when simulation frameworks and assignment and pooling strategies differ. MATSim by (W Axhausen et al., 2016) was extended to include two of these frameworks. In his Autonomous Mobility on-Demand (AMoD) extension, (Ruch et al., 2020) implemented different pooling strategies and found that in urban environments, the (Alonso-Mora et al., 2017) strategy performs the most efficient in terms of shared mileage and time savings. ...
Thesis
Full-text available
Ride-pooling is an on-demand service that offers convenient and cheap mobility solutions to reduce traditional car traffic by pooling several trips together in a single-vehicle. Due to the involvement of humans, they are not as efficient. Researchers have conducted several simulation studies to examine and improve this current inefficient system with the help of MATSim. This open-source Java application is helping them simulate large-scale ride-pooling operations. However, the simulations did not consider operational challenges in the ride-pooling studies until transport specialists developed an extension that incorporated the inefficient human factors that form the driver shift (and break) schedule. It is shift (and break) schedules that determine drivers’ working hours. A driver’s active shift time is every working hour, while their non-working hours account for their breaks. The active shift serves as a Travel Supply to the system, serving ride-pooling requests from customers that impact the static Travel Demand. Due to the uncertainty of travel supply and demand at any given moment, ride-pooling is inefficient. In order to eliminate uncertainty, demand and supply must be balanced to maintain an optimal equilibrium. One can only manipulate the Travel Supply, not the Travel Demand to achieve such optimality. The proposed model aims to resolve this issue by optimizing the shift schedules of drivers to reduce excess demand of unserved rides by a heuristic algorithm. Furthermore, the model ensures no excess Travel Supply of driver schedules that could potentially increase operating costs. Following a comprehensive literature review, the Simulated Annealing algorithm was adopted as the heuristic algorithm in the model due to its various advantages, including its ability to provide a globally optimal solution, its guarantee of convergence, and the lack of complicated mathematical equations. Nevertheless, it raises the question of whether a heuristic algorithm like Simulated Annealing can optimize drivers’ shift (and break) schedules in ride-pooling services? Having analyzed the model results, this thesis model is discerned for its potential, strength, and weaknesses in answering the research question. The model seemed to produce promising results under certain parametric conditions, so it was concluded that the model and its algorithm have the potential to optimize driver shift (and break) schedules.
... Multi-agent reinforcement learning (MARL) has found various applications in the field of transportation and simulating [50,1], stock price analyzing and trading [32,31], wireless communication networks [12,11,13], and learning behaviors in social dilemmas [33,28,34]. MARL, however, becomes intractable due to the complex interactions among agents as the number of agents increases. ...
Preprint
The marriage between mean-field theory and reinforcement learning has shown a great capacity to solve large-scale control problems with homogeneous agents. To break the homogeneity restriction of mean-field theory, a recent interest is to introduce graphon theory to the mean-field paradigm. In this paper, we propose a graphon mean-field control (GMFC) framework to approximate cooperative multi-agent reinforcement learning (MARL) with nonuniform interactions and show that the approximate order is of $\mathcal{O}(\frac{1}{\sqrt{N}})$, with $N$ the number of agents. By discretizing the graphon index of GMFC, we further introduce a smaller class of GMFC called block GMFC, which is shown to well approximate cooperative MARL. Our empirical studies on several examples demonstrate that our GMFC approach is comparable with the state-of-art MARL algorithms while enjoying better scalability.
Article
Full-text available
Reaching population immunity against COVID‐19 is proving difficult even in countries with high vaccination levels. Thus, it is critical to identify limits of control and effective measures against future outbreaks. The effects of nonpharmaceutical interventions (NPIs) and vaccination strategies are analyzed with a detailed community‐specific agent‐based model (ABM). The authors demonstrate that the threshold for population immunity is not a unique number, but depends on the vaccination strategy. Prioritizing highly interactive people diminishes the risk for an infection wave, while prioritizing the elderly minimizes fatalities when vaccinations are low. Control over COVID‐19 outbreaks requires adaptive combination of NPIs and targeted vaccination, exemplified for Germany for January–September 2021. Bimodality emerges from the heterogeneity and stochasticity of community‐specific human–human interactions and infection networks, which can render the effects of limited NPIs uncertain. The authors' simulation platform can process and analyze dynamic COVID‐19 epidemiological situations in diverse communities worldwide to predict pathways to population immunity even with limited vaccination.
Article
Evacuation plans in seacoast areas are essential for conducting people to secure zones in a timely manner. Typically, evacuation plans are based on the experience of previous evacuation drills, which are expensive processes that require coordination, planning and the collaboration of different institutions and people. During evacuation drills it is difficult to obtain all the data required to analyze the situation and additionally, it is difficult to detect all possible threatening situations. Computer simulations can be used to run evacuation models for evaluating different evacuation scenarios. However, developing realistic simulations is a complex task. Moreover, large simulation models considering many thousands of people demand a high computational cost and thereby, the simulation of different evacuation plans can become a highly time-consuming task. In this work, we present an approach to model and simulate the behavior of people in mass evacuations of seacoast areas. Our proposal aims to improve the computational efficiency of the calculations performed without compromising the quality of results by means of parallel computing. The simulation model divides the geographic area in cells of fixed sizes. Then, to reduce the amount of calculations performed in each simulation timestep, for each simulated agent we compute a mobility model by considering only the agents placed in the closest neighboring cells. The proposed simulation model achieves realistic results by combining geographic data, public census data, the density of the population, the surrounding view of each person and disaggregation by age groups. This reduces the error in decision making and allows a proper estimation of the distance of groups of people that cannot arrive at safe areas. The respective simulator has been implemented using agent-based programming in C++ and OpenMP. The simulation model was evaluated by performing experimentation on actual data collected from the Chilean cities of Iquique and Viña del Mar, and the city of Kesennuma in Japan.
Article
Transportation is a backbone of modern globalized societies. It also causes approximately one third of all European Union and U.S. greenhouse gas emissions, represents a major health hazard for global populations, and poses significant economic costs. However, rapid innovation in vehicle technology, mobile connectivity, computing hardware, and artificial intelligence (AI)-powered information systems heralds a deep socio-technical transformation of the sector. The emergence of connected, autonomous, shared, and electric (CASE) vehicle technology has created a digital layer that resides on top of the traditional physical mobility system. This article contributes a framework to direct research and practice toward leveraging the opportunities afforded by CASE for a more efficient and less environmentally problematic mobility system. The authors propose seven overarching dimensions of action. These range from designing real-time digital coordination mechanisms for the management of mobility systems to developing AI-powered real-time decision support for mobility resource planning and operations. Per each dimension, concrete angles of attack are suggested which, we hope, will spur structured engagement from both researchers and practitioners in the field.
Article
Disruptions in transport networks have major adverse implications on passengers and service providers, as they can yield delays, decreased productivity, and inconvenience for travelers. Previous studies have considered the vulnerability of connections and infrastructures. Although such studies provide insights on general disruption management approaches, there is a lack of knowledge concerning integrated multi-level traffic management and its effects on travelers to reduce the impacts of disruptions. Integrated multi-level traffic management refers to coordinating individual network operations to create an interconnected mobility management system. This study sought to assess the management of road disruption utilizing multi-level disruption management. Multi-level disruption management is proposed that integrates an information dissemination strategy and allows changing the functionality of parking spaces to traffic lanes to facilitate the movement of travelers. The capacity/frequency of public transport vehicles is also increased to help travelers reach their destinations by changing to public transport mode. To achieve such goals, an extension to an agent-based simulation was developed. Numerical experiments are applied to a part of the city of Zürich. The results indicate that the proposed approach, multi-level disruption management in a multimodal network, can shorten travelers’ delays, especially comparing the effects of disruption management. Results show heterogeneity of behavior among agents. Adding lanes as a disruption management enhances the usage of car-mode by all agents, whereas it reduces the usage of car-mode by the directly affected agents, those who cannot pass the disrupted roads. In the presence of full information and increased capacity of transit vehicles, delay is reduced by 47%.
Conference Paper
Full-text available
Large-scale agent-based transport models of whole territories have become an important tool in research and planning of new services and policies. Yet, studies based on those tools are rarely reproducible due to the complexity of data sources and modeling processes. One important element towards fully replicable simulations is automatic calibration of behavioral and infrastructural model parameters. The present paper contributes to standardizing the calibration process by describing a consistent framework for benchmarking calibration objectives and optimization algorithms. Furthermore, the paper advances the current state of the art by exploring the integration of a search acceleration method for iterative simulators (opdyts) with sample-based evolutionary search algorithms. In a use case for Paris and the MATSim simulator, we demonstrate the applicability of the framework. We show that opdyts accelerates the parameter search process, although its comparative runtime benefits decrease with higher availability of computational resources.
Preprint
Full-text available
The rapid development of digital twin technology has significantly changed the way virtual cities are used in smart cities and the transportation. In particular, digital twins provide a playground where various mobility systems, algorithms, and policies can be developed and tested. In this study, DTUMOS, a digital twin framework for urban mobility operating systems, is proposed. We construct open-source framework that can easily and flexibly apply to any city and mobility system worldwide. A novel architecture that combines an AI-based estimated time of arrival model and vehicle router algorithm enables DTUMOS to achieve high-speed performance while maintaining accuracy when implementing large-scale mobility systems. The proposed DTUMOS has distinct strengths in scalability, speed, and visualization compared to the existing state-of-the-art mobility digital twins. The performance and scalability are verified by using actual data in large metropolitan cities, such as Seoul, New York City, and Chicago. A lightweight and open-source environment of DTUMOS opens a new era for developing various simulation-based algorithms and quantitatively evaluating the effectiveness of policies for future mobility systems.
Article
Evaluating ridesharing potential is a trend in current research efforts because ridesharing provides additional mobility alternatives without extra putting vehicles on the road. Nevertheless, in most studied scenarios, the demand revealed by surveys and demographic information does not include multi-day characteristics of a trip such as frequencies on weekdays. Yet this is important for estimating the supply of rides, as the recurrence or regularity of a trip may affect the likelihood of a driver making the effort of registering the trip as being available for sharing. Likewise, if automated apps are used to recognize patterns in one’s trips and pro-actively offer them for sharing, the successful anticipation of such apps may again depend on the regularity of the trip. However, since multi-day data are complex to produce, in this paper, a data fusion procedure is proposed to generate an enriched synthetic demand for more realistic assessments. This can be achieved by combining standard single-day data sets with travel behavior patterns, which can be extracted from lifelogging data collected by most existing mobile apps. The resulting data sets after transferring information from the travel patterns to a recipient data set via statistical matching, will constrain matching trips by multi-day characteristics allowing complex scenarios. This approach enhances the evaluation of ridesharing and other shared-mobility systems and thus their ability to plan better strategies.
Chapter
To reduce the pollution and noise in the cities, the authorities encourage intermodality, notably through private cars and public transport combinations. The application of dissuasive measures such as urban tolls is an increasingly investigated solution. This paper proposes an agent-based simulation to assess the impact of an urban toll on intermodal trip behaviors (private car + public transport) in a city. The impact of the urban toll is modeled through a multinomial logit (MNL) model, which is used to estimate the modal choice for each agent. To avoid paying the toll tax, people (agents in our simulation) prefer to combine different modes of transportation, e.g., their private cars and public transport, by parking their vehicles in the park and ride facilities at the entrance to the city. Our experiments based on the MATSim platform infer that with 20 euros of toll tax, it is possible to reduce by \(20\%\) the use of the private car in the European Metropolis of Lille (MEL, France).KeywordsAgent-based modelingIntermodalityUrban tollMATSim
Article
Background More details about human movement patterns are needed to evaluate relationships between daily travel and malaria risk at finer scales. A multi-agent mobility simulation model was built to simulate the movements of villagers between home and their workplaces in two townships in Myanmar. Methods An agent-based model (ABM) was built to simulate daily travel to and from work based on responses to a travel survey. Key elements for the ABM were landcover, travel time, travel mode, occupation, malaria prevalence, and a detailed road network. Most visited network segments (MVS) for different occupations and for malaria-positive cases were extracted and compared. Data from a separate survey was used to validate the simulation. Results Mobility characteristics for different occupation groups showed that while certain patterns were shared among some groups, there were also patterns that were unique to an occupation group. Forest workers were estimated to be the most mobile occupation group, and also had the highest potential malaria exposure associated with their daily travel in Ann Township. In Singu Township, forest workers were not the most mobile group; however, they were estimated to visit regions that had higher prevalence of malaria infection over other occupation groups. Conclusions Using an ABM to simulate daily travel generated mobility patterns for different occupation groups. These spatial patterns varied by occupation. Our simulation identified occupations at a higher risk of being exposed to malaria and where these exposures were more likely to occur.
Chapter
Given a coarse satellite image and a fine satellite image of a particular location taken at the same time, the high-resolution spatiotemporal image fusion technique involves understanding the spatial correlation between the pixels of both images and using it to generate a finer image for a given coarse (or test) image taken at a later time. This technique is extensively used for monitoring agricultural land cover, forest cover, etc. The two key issues in this technique are: (i) handling missing pixel data and (ii) improving the prediction accuracy of the fine image generated from the given test coarse image. This paper tackles these two issues by proposing an efficient method consisting of the following three basic steps: (i) imputation of missing pixels using neighborhood information, (ii) cross-scale matching to adjust both the Point Spread Functions Effect (PSF) and geo-registration errors between the course and high-resolution images, and (iii) error-based modulation, which uses pixel-based multiplicative factors and residuals to fix the error caused due to modulation of temporal changes. The experimental results on the real-world satellite imagery datasets demonstrate that the proposed model outperforms the state-of-art by accurately producing the high-resolution satellite images closer to the ground truth.
Article
Full-text available
Integrated land-use transport models are often accused of being too complex, too coarse or too slow. We tightly couple the microscopic land use model SILO (Simple Integrated Land Use Orchestrator) with the agent-based transport simulation model MATSim (Multi-Agent Transport Simulation). The integration of the two models is person-centric. It means, firstly, that travel demand is generated microscopically. Secondly, SILO agents can query individualized travel information to search for housing or jobs (and to choose among available modes). Consequently, travel time matrices (skim matrices) are not needed anymore. Travel time queries can be done for any time of the day (instead of for one or few time periods), any x/y coordinate (instead of a limited number of zones) and take into account properties of the individual. This way, we avoid aggregation issues (e.g., large zones that disguise local differences) and we can account for individual constraints (e.g., nighttime workers who cannot commute by public transport for lack of service). Therefore, the behavior of agents is represented realistically, which allows us to simulate their reaction to novel policies (e.g., emission-class-based vehicle restrictions) and to extract system-wide effects. The model is applied in two study areas: a toy scenario and the metropolitan region of Munich. We simulate various transport and land use policies to test the model capabilities, including public transport extensions, zones restricted for private cars and land use development regulations. The results demonstrate that the increase of the model resolution and model expressiveness facilitates the simulation of such policies and the interpretation of the results.
Article
Large-scale agent-based simulations require higher computing resources than are usually available. Consequently, many applications rely on downscaling, that is, simulating with smaller population samples in which the results are then scaled. Existing studies have shown a need to investigate the impact of downscaling on the output statistics of such simulations. Downscaling is a common practice in transport modeling. In this study, we investigate the impacts of population downscaling on a ride-sharing service with a focus on vehicle occupancy and wait time, travel time and detour time. Our findings reveal that if transport modelers want to model on-demand services with ride sharing, it is strongly recommended to use a 100% population, or when using a smaller population sample, to estimate the relative biases of their desired metrics compared to the results of a 100% population in order for their results to be applicable for real-world situations.
ResearchGate has not been able to resolve any references for this publication.