ABSTRACT. Mathematical models of travel behaviour form a key component of the toolkit for transport planning. Traditionally, these models have been developed on the basis of travel survey data or stated preference surveys, but in the last few years, there has been growing interest in novel data sources. This presentation looks in detail at two specific topics in this context, namely a) the use automatically collected location data from mobile phone networks and GPS devices as an alternative to traditional revaled preference surveys, and b) the use of virtual reality data as an alternative to stated preference data. For both types of data, I present results from case studies and discuss the implications for model specification, estimation and use of results. In addition, the presentation discusses the potential use for new non-numeric data, such as video and social media feeds, looking at the possible benefits as well as issues with such data.

ABSTRACT. This paper addresses a variant of the well known selective pickup and delivery problem where travel times are considered uncertain. In this problem, a fleet composed of several vehicles with a given capacity should satisfy a set of customers requests consisting in transporting goods from a supplier (pickup location) to a customer (delivery location). The selective aspect consists in choosing the customers to be served on the basis of the profit collected for the service. This variant is characterized by the presence uncertain travel times in each arc. The goal is to find the solution that maximizes the net profit, expressed as the difference between the collected revenue, the route cost and the cost associated to the violation the time windows. This study introduces the problem and develops a solution approach to solve it. Very preliminary tests have been performed in order to show the efficiency of developed method.

Download: (a) Extended Abstract

ABSTRACT. In this research project, we implement a vehicle type dependent range constraint into a Vehicle Routing Problem (VRP) to consider the limited range of electric vehicles in urban freight transport planning due to the its restricted battery capacity and energy consumption. We apply this VRP in the route optimization jsprit which is linked to the microscopic agent-based simulation MATSim. In the framework of a case study focusing on food retail distribution in Berlin, Germany, we operationalize the range constraint and demonstrate the functionality and the effectiveness of this constraint using the distance from routing in a transport simulation network. Based on the simulation results, we analyze and discuss the impacts of the limitations of Battery Electric Vehicles (BEVs) on freight transport demand, road mileage performed and the resulting transport costs and greenhouse gas (GHG) emissions.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Transport of goods plays a fundamental role in the logistics chain, often representing the major cost item for the logistics operator. By improving route assignments to the vehicles of the fleet, it is possible to obtain significant time and cost savings. This paper presents a novel Ant Colony Optimization model aimed at solving Vehicle Routing Problems with Simultaneous Pickup and Delivery (VRPSPD), developed and implemented in a multi-agent simulation environment. The methodology is applied to the real case study of a logistics company in Italy in order to find an optimal set of routes able to transport palletized goods both from the main depot to different clients and from the clients to the depot or to the assigned port terminals (before being delivered to the final customer), with the objective of minimizing the cost function related both to the distance travelled and to the fleet size. Starting from the data on the daily orders provided by the company, different scenarios have been analyzed to compare the costs of transport actually incurred by the operator with those deriving from the route optimization process. First results highlight the validity of the method to reduce cost and scheduling and provide useful suggestions for large-size operations of a freight transport service.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Several logistic companies are starting to utilize electric vehicles (EVs) in their daily operations to reduce greenhouse gas pollution. The limited driving range of EVs requires visits to charging stations during their operation. The possible visits to charging stations have to be considered to avoid unnecessary long detours. Also the possible long charging times required to recharge the battery of the EV have to be considered to avoid delay in serving the customers. Therefore, the electric vehicle routing with time windows (E-VRPTW) is a common optimization problem in real-world logistic applications. An ant colony optimization based approach useful to solve the E-VRPTW problem is presented to output solutions efficiently.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. The Vehicle Routing Problem with Time Windows is concerned with finding optimal tours for vehicles with given capacity constraints to deliver goods to customers within assigned time windows. In one of our problem variant these time windows have a special structure, namely they are non-overlapping and each time window holds several customers. This is a reasonable assumption for Attended Home Delivery services. Our second variant considers general time windows which are overlapping and of arbitrary duration. In both cases we apply sweep algorithms which are known as simple, yet effective heuristics for the classical Vehicle Routing Problem. A carefully constructed benchmark set that resembles real-world data is used to prove the efficacy of both variants of our algorithms in a computational study.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. The article deals with the development of a computer system, which allows us to recognize vehicles, track them, and measure the time needed to cross an intersection by each car in the lane. The main area of research is the analysis of the dependence of the intersection crossing time on the position of vehicles in the queue formed in the traffic lane. To count the vehicles in the queue and determine their category, we used the Yolo v3 neural network and the SORT tracker modified to return the object class. All vehicles are divided into three categories depending on their acceleration. We analyzed the collected data on the queue structure and the time of its unloading and demonstrated their direct interconnection.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. National resilience is urgent necessity under the disaster environment. This paper focuses on a link reliability estimation method for improving highway network reliability under the natural disaster environment, and also proposes simplified method for link reliability estimation. Road network reliability analysis consists of two phases. The former is the node-to-node reliability analysis on the bases of given a set of link reliability. The latter is estimation of link reliability. The road network reliability methodology developed by authors is that the input data is a set of link reliability and the output is node-to-node reliability. The problem of road network reliability analysis is that obtaining or estimating a set of link reliability as input data is difficult to obtain, although the estimating method for node-to-node reliability on the basis of link reliability has developed by Wakabayashi and Iida (1992). Normally, link reliability is estimated by sufficient statistical data. However, unfortunately, to establish estimation method is difficult, because there are few methods for obtaining data. Therefore, there have been few studies on link reliability estimation method. Disasters also have different characteristics for various hazards. First, difficulty in link reliability estimation is organized. Second, factors of disaster are categorized and their characteristics are compared. Third, a typical method of slope failure is explained. Fourth, a new method for flooding disaster is introduced. Lastly, the difficulty of link reliability estimation of earthquake disaster is summarized and discussed. Then, a pilot study using Fuzzy set theory is presented.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. There are 543 toll plazas across the National Highways and State Highways in India and most of them are currently working on the Manual Toll Collection (MTC) system. The government of India has recently adopted Electronic Toll Collection (ETC) system over the MTC system, but, due to the technical constraints of toll plazas, they are either operating with both ETC and MTC systems or with MTC system only. The motive of switching from MTC to ETC is to reduce congestion and delay time. However, due to the fix toll charges, the peak hour traffic causes congestion that ultimately returns to enormous delays to users. The study proposes the policy framework of Dynamic Toll Pricing (DTP) that alleviates the congestion at toll plazas by consisting of shifting traffic volume from peak/congested hours to off-peak/non-congested hours. This type of pricing proved to be one of the methods to control the traffic by elevating the toll in peak hours and thus giving discounts in off-peak hours. Such type of study about dynamic toll is not yet carried out for Indian scenario where traffic is highly heterogeneous. Hence, the main goal of the present study is to evaluate the DTP for the Indian condition and thus to examine the effect of discount and time saving on the shift to DTP. The observations revealed that most of the respondents, about 23.53 % are willing to shift in the morning peak hour than evening peak hours (20.59 %). Further, the utility function will be developed for knowing the probability of shift.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Agent-based transport simulation models are a particularly useful tool to analyze demand-oriented transport policies and new mobility services, which have both gained significant attention lately. Since travel diaries, a traditional source to create the transport demand in agent-based transport models, are often hard to procure and not policy-sensitive, alternative approaches to creating travel demand representations for simulation scenarios are sought. In this study, a particularly ecient approach based on Big Data and a new, aspatial activity-based demand model with comparatively low input data requirements is established. Home, work, and education locations are informed based on mobile-phone-based origin-destination matrices. Other activity locations are modeled within the scope of the coevolutionary algorithm of the agent-based transport model, which is also responsible for finding suitable travel options of the modeled individuals. As a result, a comparatively lightweight process chain to create an agent-based transport simulation scenario is established, which is transferable to other regions. A basic quality evaluation of the created tool chain is carried out against a well-validated transport simulation model of the same region.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. For the case of public transport, we consider the problem of demand estimation. Given an origin-destination matrix representing the public transport demand, the distribution of flow among different lines can be obtained assuming that it corresponds to a certain equilibrium characterized by an optimization problem. The knowledge of that matrix is expensive and sometimes unaffordable in practice. In this work, we explore its estimation through the numerical solution of a bilevel optimization problem.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Autonomous vehicles offer today an important opportunity for a paradigm shift of transportation systems. These systems will become more flexible and efficient by adapting to travelers' demand. On the other hand, multi-agent systems offer an intuitive and powerful way to design complex distributed systems and agent-based simulation has proved to be well suited for studying the performance of mobility services and their impact on the users and the traffic in general. In this study, we followed an agent based modeling approach to design a mobility system based on a fleet of autonomous vehicles (robo-taxis and shuttles) and connected road side units, each of these elements represented by intelligent collaborative agents. In our model, the control is distributed among the agents and the global behaviour of the system emerges from local decisions. We then used a microsimulation tool to test the model on the real road network of the Paris-Saclay area. This area is becoming the largest hub for science, technolgy, innovation and education in France and is currently under constant development. This offers an opportunity for the design and the implementation of new models of transportation services. We designed our mobility system in an incremental manner by introducing new components and smarter behavior step by step. Our results suggest that for our area of study, a mobility service that combines buses and robo-taxis aided by connected road side units that allow to retrieve information about the traffic would perform better than a regular service.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Mode choice is influenced by a large variety of factors, as for example users’ socio-economic attributes or level of service for the different alternatives. In order to understand better what leads to temporal and spatial variations of modal split, we propose in this paper an analysis of a multi-day travel survey, with a series of descriptive statistics as well as inferential analysis on the correlation between mode choice and tour-specific attributes at both spatial and temporal levels. This paper discusses the importance of considering tour-based mode choice not only because it brings consistency between successive mode choices but also allows the inclusion of relevant tours’ characteristics such as activity types, distances, time of the day, and previous mode choices. A total of 5848 home-based tours done in 2008 are studied in the area of Ghent, Belgium. Identified patterns show the importance of modelling dynamic mode choice with trip chaining and time of the day. The modal share of car drivers differs of more than 40% between hours of the day and about 30% between different activities. Furthermore, the definition of activity spaces by principal mode choice and home-work locations introduces the calibration of probabilistic aggregate Gaussian fit to visited points.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. The goal of this research is to introduce the fundamentals of a game theoretic approach to model pedestrian roadway crossings, as well as road users interaction strategies in such scenarios. For this reason, the most influencing factors in users' decision and choice of strategy are employed to develop the model. This plays a central role in building a robust micro-simulation model to simulate the trajectory, and movements, of traffic participants in roadway crossing scenarios, in which there is no traffic control and management systems to conduct the traffic and users’ movements. The road users' interactions in a mixed traffic environment will become more challenging by the deployment of fully/partly automated driving systems, where the intentions and decisions of interacting agents must be predicted/detected to avoid conflict and improve the traffic efficiency.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Route choice for traffic assignment is an essential step in classic traffic modeling techniques, which have long been based on theoretical, empirically calibrated models. As probe vehicle data are becoming widely available, remarkable ongoing efforts attempt to improve the existing route choice models to produce more accurate estimations. This paper documents a study, conducted with real trajectory data collected on a motorway network in Germany. The aim was to analyze the usage of floating car data (FCD) to estimate the route choice on different scales. The logit model was used as a starting point. Then different levels of use of trajectory data in the route choice estimation process were tested and compared. The validation is made against reference data collected by means of Automatic Number Plate Recognition (ANPR) cameras on selected routes during consecutive weekdays. The drawback of using the logit model is discussed as well as the advantages of the FCD involvement in the estimation process. From the results, it can be concluded that an aggregated set of FCD can deliver a highly accurate realistic traffic assignment.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. The Dial-a-Ride Problem (DARP) aims to find a set of minimal cost tours for passenger vehicles in order to satisfy a set of transport requests. Each request requires to pick-up one or more passengers at a defined pick-up point and then drop-off the passengers at the desired destination. In this work, we consider a DARP that has been implemented by an Austrian mobility provider. The company focuses on rural regions that suffer from insufficient public transportation and offers a sustainable form of mobility. Moreover, the provider is especially interested in improving the sharing rates of the mobility service. Therefore, we propose a Large Neighborhood Search for solving the respective DARP. In a computational study, we compare different configurations of the service and identify the most promising configurations regarding sharing rates, passenger convenience and, hence, overall efficiency of the service.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Shared mobility is still a pioneer concept for Czech citizens. The volume of car-sharing services in Czechia is very low but it has been growing substantially during the last years. To support this emerging market, it is especially important to know who the car-sharing users are, and what is their motivation to use these services. Furthermore, from policy-making perspectives, it is important to know if car-sharing represents an environmentally friendly alternative to car-ownership in Czech conditions. We answer these questions using data about clients of the largest and oldest car-sharing company in Czechia called Autonapůl. Our findings indicate, that a typical car-sharing user is a man, more educated, with median or higher income, younger or mid-aged, often not owning a car (car-sharing substitutes car ownership). An important feature is that they have a high environmental consciousness. Economic aspects and being more flexible are important factors for the existing car-sharing users as well. However environmental motivation is usually not so important as flexibility and economic aspects. We compare our research results with the situation in the countries with a more developed car-sharing market.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. The paper presents the results of a stated preference (SP) experiment designed to investigate preferences towards Shared Autonomous Vehicles (SAVs) before and after experiencing a ride in a shared fully-automated vehicle in a real (but constrained) environment. The ride was provided in a four-seat electrically-powered fully-automated vehicle, at a large out-of-town shopping mall located to the north of the city of Bristol (UK) during January 2020.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Carsharing companies can customize their service by adopting different pricing schemes and offers with the aim of increasing fleet use and profits. Different business models have been developed such as round-trip and one-way. Is it clear that, even though many aspects of the business model and operations are the same, the different way in which these services are supplied leads to a diverse response from the users. In this work, we analyze how a fixed pricing scheme affects the behavior of the members of two different carsharing systems: two-way and one-way, explicitly considering their different income distributions to analyze social equity aspects. The policy is simulated in MATSim, an agent-based simulator able to generate realistic mode choices based on individual activity-travel behavior. Scenarios with a synthetic population of carsharing members for the city of Berlin are analyzed. We aim to provide an experimental analysis that addresses the different behavior of different demand sectors, categorized by income, in function of the supply distributed on the territory. Simulation results show that the two services are not in competition between each other: the two-way service is used as a substitute for private cars while the one-way system is preferred from agents who choose to use multiple types of modes during the day. The response from the different income classes tends to be similar for both services since all the users within the same purchase power have the same degree of acceptance for both systems.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Nowadays, multiple passenger ridesharing and its variants look one of the more promising emerging mobility concepts. However, the real implementation of these systems accounting specifically requirements for users should raise certain challenges, even further for disabled people, which inherits requirements that need special consideration: short ride times, specific vehicle characteristics depending on the mobility handicaps, narrow time windows constraints, etc. This paper presents the real case study of the public transport service that Barcelona city offers to people with reduced mobility, which could program weekly taxi trips within the city at a specific time. The proposed routing algorithm integrated into this service management system is based on a tabu search heuristic approach used to minimize the dimension of the heterogeneous fleet and the total traveled time. Furthermore, this work exhaustively analyses the operational factors of this mobility service to analyze how they affect the service performance and car-sharing factors. The obtained results show that certain operational decisions could make better use of the resources allocated to the sharing services.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Nowadays, connected cars are uncommon on our streets, but their percentage is expected to grow uninterruptedly. These devices would provide a lot of data in real-time that can be used for road operators to improve the traffic. This research is focused on one of these applications, which is shockwave damping on freeways. This work evaluates two shockwave detection methods that use probe vehicle data and fixed sensors data and one mitigation algorithm that uses variable speed limits to resolve shockwaves. This paper analyses through microscopic traffic simulation the performance of the selected algorithms and how their parameters affect them in several scenarios. Also, the effect of the penetration rate of probe vehicle data is evaluated. Finally, the best algorithm with the best parameters configuration is applied to a realistic model of the AP7 freeway in Girona (Spain). The obtained results show that the algorithms applied greatly reduce the total travel time in this network.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. In the study we summarize in this paper, we aim to develop a solution that helps to reduce the gasses exhausted from motorizedvehicles in the area of a university campus focusing on the emissions from shuttle services serving the internal area of the campus.We are motivated by the idea of evolving the possible extensions of campus based research to city based, considering that campusesare small scale settlements, with their dormitories, social centers, faculties, and etc. In order to evaluate daily emissions from shuttlevehicles serving the campus-wide area in a ring-based fashion, we employ different emission calculation methods in terms of thescale, i.e., macroscopic and microscopic, by using GPS trajectory data that is obtained on site by measurements. We discuss thetemporal and spatial effects of acceleration of shuttle services on emissions for approximately 12 hours of operation.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Capacity estimation at bottleneck locations along urban motorways are of importance for control purposes. If the capacity is known, the control strategy can be designed to not exceed capacity, or at least prolong the time until exceeding the capacity and thereby entering congested traffic conditions. In this study we show how the probability distribution of capacity can be estimated using automatic identification of traffic breakdowns and clustering methods.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. The conversion from single-entity level characteristics of traffic flow to comparable system-level characteristics shaped a new era for traffic monitoring and control. Since then landmark studies explored network-level traffic flow relationships across entire urban networks or regions and cities, mainly based on simulation data but also with empirical data. Although, the ability to observe and monitor the traffic state of the system on a network-wide level depends on the availability of existing traffic surveillance systems, adequately deployed such as to cover a complete network. To overcome this deficit, we propose a method to estimate a network's Macroscopic Fundamental Diagrams (MFD) using traffic flow mechanics at the microscopic level and exploiting class-type traffic information that can be obtained from online traffic maps. This valuable information depicted on maps is extracted based on image processing techniques, able to simultaneously perform discretization of the urban space-and the road network therein- in seamless pixels and further capture the color-coded traffic information in a suitable data structure valuable for meta-analysis. Then, the fundamental traffic flow mechanics are used for connecting the captured pixels properties with macroscopic traffic phenomena, especially with the well-defined (MFDs). The validity of the method is tested by comparing the estimated MFDs to ground-truth MFD obtained using empirical data from loop detectors. The results are providing valuable evidence on the operational characteristics of large urban areas, while at a meta-analysis stage it was able to capture spatio-temporal phenomena of urban mobility, like concentration, hysteresis and homogeneity. Since online traffic maps provide almost global coverage the proposed method is practically feasible and offers a novel approach for monitoring large-scale traffic systems.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. AV subnetworks is a way to deal with automated traffic and its technological need that will likely increase during the AVs deployment period. This strategy carries many benefits, yet some inconveniences are worth to mention. One of them relies on the fact that the design of AV subnetworks is often in practice focused on mitigating congestion in the peak-hours. However, designing for the most congested hour can be quite delicate when such a strategy is fixed throughout the day. The remaining part of the day involves different mobility patterns and shifting trips patterns throughout the day, i.e., different Origin-Destination pairs. When such O-D pairs are inside these AV subnetworks, CV owners cannot drive, and therefore a new mode of transport is necessary. This paper focuses on the lower-level decision problem, i.e., the traffic distribution during the transition period while AVs are being deployed in urban areas and AV subnetworks are expanding. A nonlinear mathematical programming model is presented to perform the trip distribution, where walking appears as an alternative. The main objective of this paper is to study the impacts of AV subnetworks from a CV owners’ perspective. A novel formulation guarantees that CV trips starting inside AV subnetworks throughout the day aren’t ignored – this means an alternative mode of transport, in this case, walking. This paper evaluates throughout the day when such situations would likely occur in a case study of the city of Delft, in the Netherlands, in two scenarios with AV subnetworks. The experiments revealed that walking is somehow inevitable when AVs reach 75% of the vehicle fleet – increasing travel costs up to 26.0% and 43.8%.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. This paper describes a data exploration study using Floating Car Data to analyze mobility patterns of geographical spaces. The objective is to build a mobility-related typology of territorial zones by investigating the related vehicle movements. Mobility features at the level of stays and trips are recovered from daily vehicle trajectories. Based on the features of stays that take place in each zone and the related trips, we characterize the functional mix of human activities of the zone and its interactions with the rest of the territory. A multi-feature clustering analysis is conducted to feature out such zonal mobility patterns in terms of trip generation and attraction, activity time and duration and zonal users’ anchor places (the individual homeplace and workplace), as well as the temporal variation among hours and days. Further spatial statistical analysis is conducted to identify spatial subsets of zones with respect to their mobility patterns, thus aiding to understand the territorial organization. Overall, this study provides a data-driven approach to study mobility interactions with territorial spaces, by spatial segmentation, characterization and differentiation.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Dynamic Traffic Assignment (DTA) models are widely used in transportation system management. Calibration is a crucial step to improve the reliability and the accuracy of DTA models. We present a systematic framework to offline calibrate the supply and demand component of a DTA model. The essence of model calibration is an optimization problem, aiming to minimize the discrepancy between field conditions and simulated traffic measurements. Instead of relying on a single traffic measurement, a multi-objective function is formulated with different traffic measurements for the supply and demand component respectively. As the calibration process is a nonlinear and stochastic problem, heuristic algorithms: the Genetic Algorithm (GA) and the Simultaneous Perturbation Stochastic Approximation (SPSA) Algorithm, are implemented as solution techniques. To overcome the limitations of standard GA and SPSA, such as high running time, we introduce the Islands Genetic Algorithm (IGA) and SPSA with Principal Component Analysis (PC-SPSA) to solve the calibration problem. A case study on a network of Munich, Germany, is used to validate the proposed methodology. The promising results indicate that calibration of the supply and demand component of a DTA model using multi traffic measurements improves modelling accuracy. In addition, IGA outperforms standard GA in terms of convergence speed and solution quality.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. The Park and Ride (P&R) facilities were important emerging mobility facilities for reducing auto-dependency in cities. However, the estimation of the usages of the P&R facilities was still difficult due to the lack of related data in many cities. To this end, the objective of this paper is for mining passenger intermodal (P&R usage) information from Automated Fare Collection (AFC) data. This is accomplished by an approach of multiple data sources fusion based on AFC data, P&R facility information, Transport Survey data (EGT), and GTFS data in Greater Paris. Three main steps are considered: (1) P&R related rail transit EGT O-D trip generation, (2) P&R related rail AFC O-D trip generation, and (3) mining P&R related intermodality by a supervised learning model. We recover the intermodal O-D trips from the full population of AFC data of the 3 rail transit modes (metro, train-RER, and tram), and classify the P&R users. The usage of P&R facilities are obtained by an unsupervised learning model, and the P&R clients are classified in three groups: two groups during morning peak hours (one with short distance, and another with long distance) and one group during evening off peaks.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. This study addresses the estimation of Origin-Destination (OD) matrices from Automated Fare Collection (AFC) data, following an aggregated approach. The methodology includes the implementation of the Trip Chaining Method (TCM) to estimate the alighting-stop of each AFC record, followed by the identification of transfers amid consecutive AFC records. The identification of transfers allows for the subsequent aggregation of trip-legs into full-trips. The set of full-trips is then used to build OD matrices. This methodology was applied to the case study of Porto, considering 20,000 smart cards over the year of 2013. The results were analyzed through a quantitative approach, complemented by the visualization of OD matrices built at two geographic levels: a detailed perspective using Public Transportation (PT) stops as matrix entries, and a top-level perspective using the PT geographic zones. Data were segmented by type of day, seasonality, and user frequency.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Origin-destination (O-D) matrix estimation is an essential input to dynamic traffic assignment models and traffic simulation models that are important tools to transportation planning. Despite the extensive study of efficient O-D matrix estimation in the literature, there is no research work that assumes faulty traffic sensors. This work proposes a model-based approach to achieve robust and good quality O-D matrix estimation using traffic flow dynamics in the presence of sensor faults. The approach consists three stages: (i) identifying the level of faulty behaviour of each sensor using a novel fault-tolerant optimisation algorithm, (ii) isolating faults, and (iii) improving O-D matrix estimation performance by adaptively reducing the sensor faults effects and resolving the state estimation problem. A path-based cell transmission model (CTM) is developed to capture the dynamics of traffic networks and associate link count observations with the path demand pattern. A state space model is utilised to associate link densities (observations) with per path densities (state vector). Simulation results show the effectiveness of the proposed methodology in terms of performance in estimating the parameters of interest, yielding estimation performance very close to the one obtained with healthy measurements, irrespective of the fault magnitude.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Bike-sharing systems are an important part of many cities’ transportation systems and they are constantly being introduced in more and more cities worldwide. Thus, the strategic decisions for these systems are essential both for their successful operation and the efficient operation of cities’ transportation systems. The present paper aims to develop a methodological approach for determining the optimal locations for installing bike-sharing stations, taking into account the operators’ perspective. Through the developed methodological approach, it is sought to select locations which maximize the demand and the area (built environment) coverage and at the same time minimize the needs for bike redistribution within the day. Thus, the optimal selection of locations for bike-sharing stations is being set as a multi-objective optimization problem. The proposed methodological approach is being applied in the city of Thessaloniki, Greece, where a dock-based and a dockless bike-sharing system operate. The results indicate that the selected stations slightly vary based on the assigned weights in each of the three objectives; higher weight in the demand coverage objective results in more selected stations close to the city’s waterfront where bicycling demand is higher, while higher weight in the area coverage results in more selected stations in the inner city.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Bicycle sharing systems (BSS) have been implemented in cities worldwide in an attempt to promote cycling. Cycling as a mode of transport has the potential to provide transport alternatives for those marginalized by car-based mobility, to reduce traffic related diseases and injuries, noise and air pollution, and to promote an active lifestyle and improve public health. The three Southern European island cities included in this research, Limassol (Cyprus), Las Palmas de Gran Canaria (Spain) and the Valletta conurbation (Malta), exhibit characteristics considered as barriers to cycling, such as hot summers and high humidity, hilliness and car-oriented culture and infrastructure. Thus far, cycling modal share is low: under 1%. However, bicycle sharing systems and policies promoting cycling have emerged in these cities too. In this research a year of trip data, shared by the BSS operators, is used to analyse the system use on an aggregate and station level by analysing the origin-destination matrices to identify spatial patterns, and by assessing different usage types to capture the behaviour of users. Particular attention is paid to the influence of tourism on the system use, through a classification of BSS use for leisure, commuting or other purposes, and analysis of the spatial influence of tourist accommodation and points of interests and the temporal influence of tourist arrivals. The comparative analysis between the three cities shows that despite sharing commonalities, the cities exhibit differences in their user base and their shared bicycle use.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Shared mobility is an innovative transportation strategy defined as the shared use of a vehicle, bicycle or other mode which enables users to gain short-term access to transportation modes on an as-needed basis. Bike-sharing systems have rapidly expanded around the world with important implications for urban areas, which generally experience problems such as recurrent congestion, air pollution and undesirable livability of the city. Considering the benefits regarding cycling and implications deriving from bike-sharing services implementation, this paper presents an in-depth analysis to investigate a variety of determinants, barriers and motivation that can influence the willingness to cycling and join bike-sharing. The study focuses on a specific target group represented by university students and their preferences have been collected through a structured questionnaire in applying the Likert Scale. A statistical analysis has been realized based on a chi-squared test, deriving the difference between expected and observed frequencies for several combinations of the analyzed attributes. First results highlight the differences between the impact of economic, environmental and social factors for students cycling and provide useful suggestion to define the way for a well-thought-out design of a bike-sharing transport service.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. In recent years, an increasing number of electric scooters (e-scooters)-sharing companies has appeared around the world. A typical company distributes a shared fleet of e-scooters in a city and the vehicles can be easily rented through a smartphone application, paying a per minute fee. However, a major issue that has soon become apparent is that a consistent part of the users is prone to park the e-scooters without caring about the rules of the road, abandoning them in locations and positions that greatly reduce urban decorum and may interfere with pedestrians and other vehicles. Many local governments have thus started to take actions, such as bans and fines, against e-scooter sharing companies. In order to cope with the issue of bad parking and do not compromise acceptance of e-scooters by city residents, some companies have begun to include correcting the position of wrongly parked scooters as an important part of their operations. In this work, we address the problem of optimally managing the actions of a set of agents who are hired by a sharing company expressly for repositioning e-scooters in order to guarantee urban decorum. We call these agents beautificators, since their fundamental task is to reposition scooters over short distances (even just a few meters), so to fix inappropriate and disordered parking made by users. We stress that such repositioning must not be confounded with traditional relocation made in vehicle-sharing systems to rebalance fleets in the service area: rebalancing is made over medium and long city-distances and is primarily aimed at guaranteeing a balanced distribution of vehicles in the service area, better satisfying the demand and increasing the overall profit. To the best of our knowledge, such optimization problem has not yet been considered in literature and we propose to model it by Integer Linear Programming and solve it by means of a matheuristic. Computational tests on realistic instances defined in collaboration with professionals of a major e-scooter sharing company are reported and discussed, showing a remarkable performance of our modelling and algorithmic approach.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Due to modern communication and information technology, shared taxi services are on the rise. While most research and practical projects focus on services operating in densely populated areas, the advantages of individual, flexible and shared transportation services in sparsely populated urban areas have not been well explored yet. Using a constraint programming formulation, this work investigates request management policies to evaluate the economic potential of shared taxi services. In particular, computational experiments demonstrate the impact of pre-bookings on efficiency and service quality of these services in rural areas.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. This paper summarizes the existing work on implementation models of charging stations. It aims to draw up a comparative overview of the approaches that have been used until 2020 to simulate and assess the impacts of location strategies of charging infrastructure. In particular, three categories of approaches are identified. Then, we analysed, for each approach, both technical and economic factors , in order to provide a complete analysis to stakeholders involved in EV charging infrastructure design and planning. Finally, a list of recommendation for future research works are provided in order to develop models that are closer to reality.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. This work studies the real-time train rescheduling problem for a busy metro system. We aim to analyse how train delays propagate in the network due to small disturbances, while considering various types of recovery strategies. We compare optimization and rule-based algorithms and integrate them into the SIMSTORS traffic simulator. We use the heuristic and exact algorithms implemented in the AGLIBRARY solver. The resulting simulation-optimization framework is used to investigate the following operational issues: how to design suitable periodic or event-based strategies, how to setup the traffic prediction horizon dynamically, how to decide the frequency and the length of the optimization process. The proposed closed loop framework is used to evaluate several disturbance scenarios and simulation-optimization periods for a practical case study: the Santiago Metro Line 1, in Chile. Preliminary results on various settings of the framework show that the best performance, in terms of train delay minimization and system reliability, is achieved when using the optimization algorithms compared to the rule-based traffic regulation that implements a hold-on strategy.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Public transport (PT) agencies are expected to provide passenger-oriented services. There is a need to understand the user’s perception of information technologies, especially within specific socioeconomic and demographic groups. Because of this, and although public transportation agencies are integrating various systems, this work aims at assessing which type of real-time information (RTI) and display platform (DP) for PT are perceived as the most useful among different socioeconomic and demographic groups citizens. Surveys were disseminated in two different readiness level areas, to understand citizens preferences based on age, academic qualifications, public transportation usage, and access to such type of technologies. There was a total of 655 respondents, 196 in Portugal and 459 in Sweden. The most valued RTI were arrival/departure time (ADT) and trip planning (TP). Results suggest there are clear differences between the preferences of young and older users. While in Portugal, high percentage of young people prefer TP as RTI, in Sweden it was found that the percentage of respondents that prefer ADT or TP increases with the increase of age. With respect to DP, panels and apps are those with more votes. Results show in general, younger people prefer Apps and older people prefer RTI provided through panels. Moreover, the findings suggest that areas served with less technologies on information for public transport give more value of having access to RTI. These results are important to support public authorities on designing an integrated system with such technologies to be possibly implemented at a regional level.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Transport networks constitute the backbone of urban and regional systems as they provide the required channels for spatial interaction. To that end, an important dimension of transport systems concerns their reliability. In brief, reliability measures have evolved the lines of accounting for the impact of recurring and non-recurring congestion on the performance of the system. Especially for the case of public transport systems, reliability can be seen as an important intrinsic characteristic of the system due to the multi-chain character of the trips, where a potential delay might translate into a subsequent missed connection, and hence a disproportionate travel time increase for the affected users. Nevertheless, the majority of developed reliability indicators focus on the supply side of the system, failing to incorporate passengers’ dimension into their formulation.

Download: (a) Extended Abstract (b) Presentation video

ABSTRACT. Accessibility is one of the key measures of urban transportation planning, which quantify how easy is the access to a facility. Public transport accessibility concerns of the access level of geographical locations to public transport. In this paper, accessibility is used as an indicator to estimate social exclusion based on the maximum distance that someone has to walk to reach the public transport. The concept of the 6-minute walking distance (6MWD) is applied to measure accurately the walking ability for different groups of the population. A real life case study is conducted to get insight into the transportation network of the Porto Metropolitan Area, Portugal. For this purpose, geographic, demographic and infrastructure data were collected and integrated. Also, webservices are used to measure walking distances between locations. The results of this study allowed to characterize regions by different levels of accessibility, providing insight into the social exclusion in public transport. This assessment is used not only to identify inequities but also to get an overview of the service quality of public transport.

Download: (a) Extended Abstract (b) Presentation video