Multi-agent Transport Simulation Research Articles

Assessing Equity in Carsharing Systems: the case of Munich, Germany

This paper shows an application of a multi-agent transport simulation to evaluate equity effects of the introduction of a carsharing system. Using vehicles, members, and planning data of Oply, a carsharing service that operated in Munich until March 2020, we analyze the evolution of the distribution of costs and benefits among the inhabitants of this city. By explicitly introducing the income as an active part of the utility calculation we evaluate how offering a new mode of transport impacts the score of the agents. Two scenarios are employed to assess equity in economic terms and accessibility terms. Two different outcomes are expected: firstly, as a high pricing service, carsharing will favor high-income agents, thus skewing benefits towards them; secondly, we show that the granularity of this agent-based simulator makes it a handy tool when conducting policy evaluations on the introduction of a carsharing system.

Travel Behavior of Car Travelers with the Presence of Park-and-Ride Facilities and Autonomous Vehicles

Travelers' behavior is predicted based on their individual preferences. People search for alternatives to maximize their benefit from doing activities, such as increasing the activity time by minimizing the travel time. Traffic congestion and the scarcity of parking spaces in the city center motivate the decision-makers to encourage travelers to use the park-and-ride (P&R) system. An evaluation concerning the impact of using the P&R system on the travel behavior of car users is conducted. Some of the existing P&R facilities are incorporated into the daily activity plans of car travelers to produce new daily activity plans (i.e., P&R facility is considered an activity). By using the Multi-Agent Transport Simulation (MATSim) open-source tool, simulations of the daily activity plans including the P&R system and autonomous vehicles (AVs) are conducted. The study examines three scenarios: (1) a simulation of the existing condition, (2) a simulation of the daily activity plans of the travelers with the P&R system, and (3) a simulation of the daily activity plans of the travelers with the P&R system and AVs. The results show that using the P&R system increases the overall travel time compared with the existing conditions, and the use of AVs as a transport mode impacts the existing modal share as follows: 64 % of the car users switch to AVs, while 15 % of the car users switch to public transport. The output of this study might be used by policy-makers in parking pricing and the location of the P&R facilities.

Assessing road pricing effects on a multimodal network based on macroscopic fundamental diagram hysteresis

PurposeRoad pricing is an efficient strategy for managing urban traffic to relieve congestion. The macroscopic fundamental diagram (MFD), which relates the average network density and flow, is a simple tool for assessing road pricing effects on transportation network performance. However, recent research indicates that it may have complexity (an MFD hysteresis loop), especially for city-scale networks. Although ignoring MFD hysteresis may provide inaccurate results, pricing models that consider this hysteresis are scarce. This paper aims to assess road pricing effects on network performance considering MFD hysteresis characteristics.Design/methodology/approachThis paper evaluated different pricing strategies spatially and temporally and compared network performance based on MFD shape in the presence of MFD hysteresis loops. These strategies were developed on a multimodal (cars and buses) network using a multi-agent transport simulation (MATSim).FindingsThis study found that pricing some links for a short duration with an optimum charge calculated based on the MFD provides higher travel time savings than the previous relevant studies.Originality/valueThese findings may facilitate assessing road pricing effects on multimodal network performance considering MFD hysteresis.

The Influence of Introducing Autonomous Vehicles on Conventional Transport Modes and Travel Time

Introducing autonomous vehicles (AVs) on the market is likely to bring changes in the mobility of travelers. In this work, extensive research is conducted to study the impact of different levels of automation on the mobility of people, and full driving automation needs further study because it is still under development. The impacts of AVs on travel behavior can be studied by integrating AVs into activity-based models. The contribution of this study is the estimation of AVs’ impacts on travelers’ mobility when different travel demands are provided, and also the estimation of AVs’ impact on the modal share considering the different willingness of pay to travel by AVs. This study analyses the potential impacts of AVs on travel behavior by investigating a sample of 8500 travelers who recorded their daily activity plans in Budapest, Hungary. Three scenarios are derived to study travel behavior and to find the impacts of the AVs on the conventional transport modes. The scenarios include (1) a simulation of the existing condition, (2) a simulation of AVs as a full replacement for conventional transport modes, and (3) a simulation of the AVs with conventional transport modes concerning different marginal utilities of travel time in AVs. The simulations are done by using the Multi-Agent Transport Simulation (MATSim) open-source software, which applies a co-evolutionary optimization algorithm. Using the scenarios in the study, we develop a base model, determine the required fleet size of AVs needed to fulfill the demand of the different groups of travelers, and predict the new modal shares of the transport modes when AVs appear on the market. The results demonstrate that the travelers are exposed to a reduction in travel time once conventional transport modes are replaced by AVs. The impact of the value of travel time (VOT) on the usage of AVs and the modal share is demonstrated. The decrease in the VOT of AVs increases the usage of AVs, and it particularly decreases the usage of cars even more than other transport modes. AVs strongly affect the public transport when the VOT of AVs gets close to the VOT of public transport. Finally, the result shows that 1 AV can replace 7.85 conventional vehicles with acceptable waiting time.

Analysis of Electric Moped Scooter Sharing in Berlin: A Technical, Economic and Environmental Perspective

Electric moped scooter sharing services have recently experienced strong growth rates, particularly in Europe. Due to their compactness, environmental-friendliness and convenience, shared e-mopeds are suitable for helping to reduce the environmental impact of urban transport. However, its traffic-related, economic and environmental effects are merely represented in academic research. Therefore, this study investigates the ability of an e-moped sharing system to substitute passenger car trips, and the resulting economic and environmental effects. First, we model fleets of 2500, 10,000 and 50,000 shared e-mopeds in Berlin, based on a passenger car scenario generated by the multi-agent transport simulation framework MATSim. Afterwards, the total cost of ownership and a life cycle assessment are conducted. The results indicate that a substantial part of all passenger car trips in Berlin can be substituted. The larger the fleet, the more and longer trips are replaced. Simultaneously, the efficiency in terms of fleet utilization decreases. The scenario with 10,000 e-mopeds offers the lowest total distance-based costs for sharing operators, whereas a fleet consisting of 2500 vehicles exhibits the lowest environmental emissions per kilometer. Already with today’s grid mix, the use of shared e-mopeds results in a significant reduction in environmental impact compared to conventional and battery-electric passenger cars.

Analysis of Electric Moped Scooter Sharing in Berlin: A Technical, Economic and Environmental Perspective

Electric moped scooter sharing services have recently experienced strong growth rates, particularly in Europe. Due to their compactness, environmental-friendliness and convenience, shared e-mopeds are suitable for helping to reduce the environmental impact of urban transport. However, its traffic-related, economic and environmental effects are merely represented in academic research. Therefore, this study investigates the ability of an e-moped sharing system to substitute passenger car trips, and the resulting economic and environmental effects. First, we model fleets of 2500, 10,000 and 50,000 shared e-mopeds in Berlin, based on a passenger car scenario generated by the multi-agent transport simulation framework MATSim. Afterwards, the total cost of ownership and a life cycle assessment are conducted. The results indicate that a substantial part of all passenger car trips in Berlin can be substituted. The larger the fleet, the more and longer trips are replaced. Simultaneously, the efficiency in terms of fleet utilization decreases. The scenario with 10,000 e-mopeds offers the lowest total distance-based costs for sharing operators, whereas a fleet consisting of 2500 vehicles exhibits the lowest environmental emissions per kilometer. Already with today’s grid mix, the use of shared e-mopeds results in a significant reduction in environmental impact compared to conventional and battery-electric passenger cars.

Improving emergency response in the era of ADAS vehicles in the Smart City

Management of emergency vehicles can be fostered within a Smart City, i.e. an urban environment in which many IoT devices are orchestrated by a distributed intelligence able to suggest to road users the best course of action in different traffic situations. By extending MATSim (Multi-Agent Transport Simulation Software), we design and test appropriate mitigation strategies when traffic accidents occur within an existing urban area augmented with V2V (Vehicle-to-Vehicle), V2I (Vehicle-to-Infrastructure) capabilities and Advanced Driving Assisted cars (ADAS). Further, we propose traffic congestion models and related mechanisms for improving the necessary time for emergency vehicles to respond to accidents.

Potential Urban Air Mobility Travel Time Savings: An Exploratory Analysis of Munich, Paris, and San Francisco

The advent of electrified, distributed propulsion in vertical take-off and landing (eVTOL) aircraft promises aerial passenger transport within, into, or out of urban areas. Urban air mobility (UAM), i.e., the on-demand concept that utilizes eVTOL aircraft, might substantially reduce travel times when compared to ground-based transportation. Trips of three, pre-existent, and calibrated agent-based transport scenarios (Munich Metropolitan Region, Île-de-France, and San Francisco Bay Area) have been routed using the UAM-extension for the multi-agent transport simulation (MATSim) to calculate congested trip travel times for each trip’s original mode—i.e., car or public transport (PT)—and UAM. The resulting travel times are compared and allow the deduction of potential UAM trip shares under varying UAM properties, such as the number of stations, total process time, and cruise flight speed. Under base-case conditions, the share of motorized trips for which UAM would reduce the travel times ranges between 3% and 13% across the three scenarios. Process times and number of stations heavily influence these potential shares, where the vast majority of UAM trips would be below 50 km in range. Compared to car usage, UAM’s (base case) travel times are estimated to be competitive beyond the range of a 50-minute car ride and are less than half as much influenced by congestion.

A Concept Agent-Based Simulation Model to Evaluate the Impacts of a Shared Space Network

Abstract In shared space, all road users are encouraged to legally occupy the same road space with little physical or visual separation. Complex social interactions that appear in these spaces create complicated ecosystems, which make the estimation of future impacts difficult using classic modeling techniques. This study introduces a concept Agent-Based Simulation Model to represent this complex ecosystem at a network level. The proposed model is based on the methodological framework of Multi-Agent Transport Simulation (MATSim). Two critical factors are integrated in this iterative simulation process, i.e. traffic complexity level that influences network supply and perceived safety that affects mode and route choices.

Quantifying Health & Economic Benefits of Bicycle Superhighway: Evidence from Patna

Abstract The active mode of transport is good for health as well as it is one of the most important modes of sustainable transport. In many previous studies/reports, a cost-benefit analysis for new infrastructure is provided; however, it rarely includes the benefits due to physical activity. The present study attempts to quantify the benefits of physical activities. A comparison is made between business as usual and bicycle superhighway (BSH) scenario. The protective benefits are calculated for different values of average cycling duration. For reference, cycling for 3h/week results in the relative risk of 0.72. The average cycling duration is estimated using the output of a multi-agent transport simulation model. This subsequently is used to determine the number of deaths prevented each year for both scenarios. With the help of a simulation model of Patna, India, it is shown that a substantial increase in the longer bicycle trips is observed. This results in saving of more than 750 deaths each year.

Population downscaling in multi-agent transportation simulations: A review and case study

Simulating the dynamics and evolution of metropolitan transportation systems serving millions of travelers remains a difficult task, beyond existing standard software's abilities. MATSim (Multi-Agent Transportation Simulation) is the only high-resolution spatially-explicit framework that allows intrinsic population downscaling - simulating the entire system's dynamics based only on a fraction k of the traveling population. Till now, the choice of k was dictated by hardware performance, and a common rule was not to downscale below 10%. We investigate downscaling in MATSim by comparing the aggregate and disaggregate statistics that describe the dynamics of car traffic in full-scaled and downscaled simulations of the Sioux Falls test case road network. Simulations with 25% or higher shares of the traveler population preserve all major urban traffic statistics. Within the 10–25% interval, downscaling becomes unstable for some of the statistics. For scenarios that are downscaled 10% and below, statistics can substantially deviate from the full-scale model. We further discuss the problems related to a multimodal transportation simulation's downscaling that also includes public transportation.

Multimodal transportation system protection against sea level rise

Transportation infrastructure resilience is an important component of a region’s ability to recover from natural disasters. While Sea Level Rise (SLR) is becoming inevitable with climate change, little is known of the impact of protection strategies on multiple modes of transport. The present paper proposes a framework where a range of coastal protection strategies are undertaken in the case of one meter of SLR (expected by the year 2100). The methodology incorporates high-resolution hydrodynamic simulations using the Coastal Storm Modeling System (CoSMoS) and traffic simulations using the Multi-Agent Transport Simulation (MATSim) to quantify the potential impact of SLR and protection strategies on both the highway and public transit systems in the San Francisco Bay Area.The results of the traffic simulations are analyzed at the regional and Transportation Analysis Zone (TAZ) levels. Modeling results show that coastal protection of one area will affect the transportation system (sometimes negatively) in areas beyond its vicinity. The improved spatial resolution, and the integration of highway and transit networks in a unified model, reveal transportation phenomena that were not identified in previous studies. By quantifying the impacts on commuters’ mobility in different TAZs, the methodology can be used to develop effective and inclusive strategies against SLR for a given region of interest.

Potential of Private Autonomous Vehicles for Parcel Delivery

Using the same vehicles for both passenger and freight transport, to increase vehicle occupancy and decrease their number, is an idea that drives transport planners and is also being addressed by manufacturers. This paper proposes a methodology to simulate the behavior of such vehicles within an urban traffic system and evaluate their performance. The aim is to investigate the impacts of resignation from fleet ownership by a transport service company (TSC) operating on a city-wide scale. In the simulation, the service provider hires private autonomous cars for tour performance. Based on assumptions concerning the operation of such vehicles and TSCs, the software Multi-Agent Transport Simulation (MATSim) is extended to model vehicle and operator behavior. The proposed framework is applied to a case study of a parcel delivery service in Berlin serving a synthetic parcel demand. Results suggest that the vehicle miles traveled for freight purposes increase because of additional access and egress trips. Moreover, the number of vehicles en route is higher throughout the day. The lowering of driver costs can reduce the costs of the operator by approximately 74.5%. If the service provider additionally considers the resignation from fleet ownership, it might lower the operation cost by another 10%, not taking into account the costs of system transfer or risks like vehicle non-availability. From an economic perspective, the reduction of the overall number of vehicles in the system seems to be beneficial.

Assessment of the DRT System Based on an Optimal Routing Strategy

Demand responsive transport (DRT) is operated according to flexible routes, dispatch intervals, and dynamic demand, is attracting a lot of attention. The biggest characteristic of the DRT service is that the vehicle routes and schedules are operated optimally based on real-time travel requests of using passengers without fixed operating schedules. This study analyzed the feasibility of implementing the DRT service by analyzing the benefits for the users and cost of the operator from the effects of increasing public transportation use and providing personalized mobility service based on DRT implementation by the introduction of DRT using multi-agent transport simulation (MATSim). Through the simulation, the DRT is expected to provide convenient, fast, and cost-effective mobility services to customers; provide an optimal vehicle scale to providers; and, ultimately, achieve a safe and efficient transportation system.

A multi-dimensional rescheduling model in disrupted transport network using rule-based decision making

Abstract Apart from daily recurrent traffic congestion, unforeseen events such as flood induced road damages or bridge collapses can degrade the capacity of traffic supply and cause a significant influence on travel demand. An individual realising the unexpected events would take action to reschedule its day plan in order to fit into the new circumstance. This paper analyses the potential reschedule possibilities by augmenting the Within-Day Replanning simulation model implemented in the Multi-Agent Transport Simulation (MATSim) framework. Agents can adjust day plan through multi-dimensional travel decisions including route choice, departure time choice, mode switch, and trip cancellation. The enhanced model not only improves the flexibility of MATSim in rescheduling a plan during an execution day, but also lays the foundation of integrating more detailed heterogeneity decision rules into the travel behaviour simulation to cope with unexpected incidents. Furthermore, the proposed rescheduling model is capable of predicting the network performance in the real-world picture and gives a hint on how best react to transport disruptions for transport management agency.

Spatiotemporal Implications of Population Downscaling: A MATSim Study of Sioux Falls Morning Peak Traffic

Abstract Computer hardware is steadily advancing; however, simulating real urban transportation systems that serve millions of individual travelers is still a difficult task. The Multi-Agent Transportation Simulation (MATSim) is the only agent-based traffic model that includes intrinsic downscaling - procedures of changing network parameters in order to simulate the dynamics of the system as a whole while activating only a fraction k of travelers. In this paper, we present the MATSim’s downscaling procedure and compare the dynamics of car traffic in the downscaled and full-scaled scenarios of the Sioux Falls test case. We compare aggregate and disaggregate statistics that represent Sioux Falls daily traffic, focusing on the morning peak. We conclude that downscaling up to k = 0.25 preserves all major statistics of urban traffic, within the interval of k between [0.1, 0.25]. Some of the statistics replicate well the statistics of the full-scaled runs, while downscaling below k = 0.1 can easily result in substantial deviations from the dynamics of the full-scale model.

Scalable Model Predictive Control for Autonomous Mobility-on-Demand Systems

Technological advances in self-driving vehicles will soon enable the implementation of large-scale mobility-on-demand (MoD) systems. The efficient management of fleets of vehicles remains a key challenge, in particular to achieve a demand-aligned distribution of available vehicles, commonly referred to as rebalancing. In this article, we present a discrete-time model of an autonomous MoD system, in which unit capacity self-driving vehicles serve transportation requests consisting of a (time, origin, destination) tuple on a directed graph. Time delays in the discrete-time model are approximated as first-order lag elements yielding a sparse model suitable for model predictive control (MPC). The well-posedness of the model is demonstrated, and a characterization of its equilibrium points is given. Furthermore, we show the stabilizability of the model and propose an MPC scheme that, due to the sparsity of the model, can be applied even to large-scale cities. We verify the performance of the scheme in a multiagent transport simulation and demonstrate that service levels outperform those of the existing rebalancing schemes for identical fleet sizes.

Agent- and activity-based large-scale simulation of enroute travel, enroute refuelling and parking behaviours in Beijing, China

This paper develops an agent- and activity-based large-scale simulation model for Beijing, China (MATSim-Beijing) to explicitly simulate enroute travel, enroute refuelling and parking behaviours, as well as the associated vehicular energy consumption and emissions, based on MATSim (Multi-Agent Transport Simulation), which is a typical integrated activity-based model. In order to take into account heterogeneous parking and refuelling behaviours, the MATSim-Beijing model incorporates several Multinomial Logit (MNL) models to predict individual choices about the maximum acceptable times of walking from trip destination to parking lot, of diverting to a refuelling station and of queuing at a station, using the data collected in a paper-based questionnaire survey in Beijing. A Sensitivity Analysis (SA) -based calibration method was used to estimate the model parameters by searching for an optimal parameter combination with the objective of minimize the gap between simulated and observed traffic flow data, exhibiting a relatively good performance of decreasing the Mean Absolute Percentage Error (MAPE) by around 23%. Further, the calibrated model was used to investigate whether and how the population scaling and network simplification, which were two commonly used approaches to speeding up large-scale traffic simulations, might influence model accuracy and computing time. The results indicated that both approaches could to some extent influence model outputs, though they could significantly reduce computing time.

Demand responsive transport: Generation of activity patterns from mobile phone network data to support the operation of new mobility services

Abstract Demand Responsive Transport (DRT), covering the first/last mile of a journey, plays a pivotal role in the delivery of a seamless integrated door-to-door service, which is a fundamental requirement for the implementation of Mobility as a Service (MaaS). Business models currently in use do not deliver sustainable and durable DRT in urban areas. This can be minimised using transport modelling tools ahead of the operation phase. However, transport models are not fit for purpose when it comes to model on-demand shared mobility services and the integration of these services in a complex public transport ecosystem. This paper focuses on how to model demand for ride-shared mobility services and how to plan for these services when running in integration with mass transit. An Agent Based Model (ABM), built in the open-source Multi-Agent Transport Simulation (MatSim) platform for Bristol (UK), has used an activity-based approach to model demand for two New Mobility Services (NMS). This was then generated using anonymised and aggregated Mobile phone Network Dataset (MND), both as a trip-based and trip chains dataset to assess the capabilities of MND. Results show that the simulations built using the trip chains MND datasets (722,752 agents generated) lead to better insights in users’ travel patterns. An advanced method using additional data sources covering land-use (location of business, services and transport facilities) was used to infer purpose and mode of transport during the multimodal journeys. The output of the ABM predicts demand for two flexible on-demand services, identifying best routes to maximise the number of users served and quantifying the benefits in the integration with public transport services and in modal shift from private cars. This is expected to be useful either for Local Authorities for transport planning purposes, and for operators looking at financially sustainable DRT.

Development of a map-matching algorithm for dynamic-sampling-rate GPS signals to determine vehicle routes on a MATSim network

The rapid development and proliferation of global positioning system (GPS)-enabled systems and devices have led to a significant increase in the availability of transport data, more specifically GPS trajectories, that can be used in researching vehicle activities. In order to save data storage- and handling costs many vehicle tracking systems only store low-frequency trajectories for vehicles. A number of existing methods used to map GPS trajectories to a digital road network were analysed and such an algorithm was implemented in Multi-Agent Transport Simulation (MATSim), an open source collaborative simulation package for Java. The map-matching algorithm was tested on a simple grid network and a real and extensive network of the City of Cape Town, South Africa. Experimentation showed the network size has the biggest influence on algorithm execution time and that a network must be reduced to include only the links that the vehicle most likely traversed. The algorithm is not suited for trajectories with sampling rates less than 5 seconds as it can result in unrealistic paths chosen, but it manages to obtain accuracies of around 80% up until sampling sizes of around 50 seconds whereafter the accuracy decreases. Further experimentation also revealed optimal algorithm parameters for matching trajectories on the Cape Town network. The use case for the implementation was to infer basic vehicle travel information, such as route travelled and speed of travel, for municipal waste collection vehicles in the City of Cape Town, South Africa.