ABSTRACT. In shared space environments, different types of road users share the urban space and frequently interact with each other, e.g., to negotiate priority. Instead of traffic rules, their interactions are governed by social protocols such as courtesy behavior and by informal rules like the rule of the road. Modeling the movement behavior of road users in shared places is crucial to evaluate the safety and efficiency of such environments. Based on the observation that this behavior differs based on environment topology, in this paper, we investigate two types of shared space topology: intersection zones and road zones. We use real-world data to investigate how road users behave differently in different shared space zones and model zone-specific movement behavior of road users. We validate our model by simulating various zone-specific scenarios involving pedestrian-to-pedestrian, multiple cars to pedestrians and also car-to-car interactions. The results indicate that our simulation model produces realistic behaviors.

ABSTRACT. It is a fact that a new or an improved transport system has great impact on land uses and real estate prices. While there are many studies, which examine the effect of a new urban public transport system on the values of neighboring properties, little effort has been done for the identification of the benefits of businesses. The present paper attempts to identify parameters which have significant impact on business revenue, as well as to quantify that impact. For the purposes of the study, a questionnaire-based survey took place, addressed to business enterprise owners and professionals in the areas of ten under-construction metro stations. The collection of the data followed by their statistical analysis, which concluded with the development of an ordinal regression model. The results indicate that that more benefits are expected by businesses which are located close to the metro stations, in areas with limited parking availability and especially by restaurants/café/bars. Models of that type can be very useful in cases of implementing alternative funding methods, such as Value Capture, in which it is necessary to estimate the benefits that all the parties will gain, in order to recover the appropriate proportion for financing the project.

ABSTRACT. Retrieving exhaustive information about individual mobility patterns is an essential step in order to implement effective mobility solutions. Despite their popularity, digital travel surveys still require significant amount of inputs from the respondent. As a consequence, samples are still relatively small in sizes and data are collected during a relatively short period of time – between a few months and a year. Driven by these motivations, the approach proposed in this paper uses mobile phone location history to automatically detect activity location without any interaction with the respondent. The proposed methodology uses raw location data together with a special indexing technique to calculate the probability of performing a certain activity in a certain location. Then, a heuristic rule improves this estimation by considering the value of the information over time (i.e. some time of the day carries more valuable information). Finally, GIS data about the number of facilities located in a certain area is downloaded in real-time to further improve the overall estimation. Results of this exploratory study support the idea that the proposed approach can reconstruct complex mobility patterns while minimizing the number of active inputs from the respondent.

ABSTRACT. A general procedure for the validation of a driving simulation environment for the analysis of gap-acceptance behavior was developed in this study. It allows to test whether a synthetic indicator of gap-acceptance behavior (the mean critical gap) shows significant differences when computed on the basis of field observations versus observations collected in the simulated environment. If such differences are not significant, driver behavior can be considered similar in the two contexts, thus supporting validation of the driving simulation environment . In order to demonstrate its effectiveness, the proposed procedure is applied to the case of a three-leg roundabout located in the Veneto region (Italy). The results show that the mean critical gap estimated in the field and the mean critical gap estimated in the virtual environment are not significantly different. The proposed procedure can be applied in various contexts in which gap-acceptance behavior is a central element in terms of safety and operational performance of the traffic system under analysis.

ABSTRACT. Application of extreme value theory (EVT) to road safety analysis is gaining interest, thanks to its ability to produce quick and reliable safety evaluations without the use of crash data. Traditionally applied to single collision types and single extreme variables (i.e. surrogate measures of safety), EVT can be further exploited to simultaneously model multiple collision types with the use of multiple extreme variables. In this paper two bivariate EVT approaches are applied for the safety evaluation of a three-leg unsignalized intersection. The first one considers two separate conflict points and a single surrogate measure of safety, the second one considers two surrogate measures of safety collected in a single conflict point. Each bivariate analysis was applied with both classic EVT methods: Block Maxima (BM) and Peak-Over Threshold (POT). The models produced good results, in particular with the POT method, and were able to significantly improve the univariate results when the two estimation datasets were correlated.

ABSTRACT. At many intersections, turning drivers have to yield to traffic from the opposite direction and crossing pedestrians and cyclists. Still quite limited understanding of the situation makes this task very demanding for autonomous vehicles. These can be segregated from the flow by means of presignals and be provided with exclusive protected movements. Signal control manages this ``presorting and presignaling'' system to adapt to the demand. Those changes potentially influence traffic safety and traffic efficiency at intersections and should be evaluated before implementing the system. In a microscopic traffic simulation study, ``presorting and presignaling'' is applied to a range of existing German intersections. The model is calibrated using headway and vehicle trajectory data. Different scenarios are created by varying traffic demand, signaling aspects, the share of autonomous vehicles and their driving strategies, both for the unchanged and the presignaled intersection. Firstly, the capacity and the level of service are estimated. Secondly, advances in traffic safety are examined using conflict measures like time-to-collision and post-encroachment time. Preliminary results show that typical traffic demand can be served but should go side by side with intelligent traffic management.

ABSTRACT. Traffic infrastructures in developing countries are entirely different than developed countries in terms of pedestrian facilities. Pedestrians are not given due weightage and hence pedestrians crossing at urban midblock section is a common phenomenon in developing countries. Such crossing pedestrians not only interrupt the regular traffic but also put themselves at high risk of collision. Traffic stream distracted due to such crossing pedestrians experience delay and ultimately the quality of traffic flow is degraded. The present work is attempted to estimate delay caused to vehicles by crossing pedestrians. The speed models have been developed for different classes of vehicles observed on urban roads in India. The empirical data was collected from fifteen different midblock section with and without effect of crossing pedestrians designated as base and friction section respectively. The developed model further used to plot the speed-flow curves at different degree of pedestrian flow and thereby estimated delay due to crossing pedestrians. It is observed that with the increase in the pedestrian cross flow the vehicular delay increases. The study may be useful for delay estimation at such urban midblock sections influenced by pedestrian crossing for better travel time estimation and give some insight for designing pedestrian facility.

ABSTRACT. The use of Electric Vehicles for logistics necessitates routing based on battery levels and includes trips to the charging station to maintain sufficient charge. Battery consumption depends on several external parameters, and therefore fixed route solutions may not always remain feasible while execution. This paper mathematically verifies decomposing the Electric Vehicle Routing Problem into initial routing over customers and a recourse strategy for dynamic charging decisions. To accommodate the stochastic battery consumption, a charging strategy comparable to the dynamic inventory control policy is proposed. The optimal solutions are obtained from an exact formulation assuming deterministic battery performance. Two policies based on separate decision parameter for each node or a common parameter over the complete network are applied when battery consumption is stochastic. The vehicle movement is simulated over fixed routes and the battery levels are observed at the decision epochs where the charging decision is taken. The policies are compared with respect to the minimum battery level observed, number of times vehicle is charged and the number of infeasible solutions under Expected and Random Realizations of the battery performance. It is observed that the network-wide fixed parameters give comparable results to the node dependent parameters, using lesser information.

ABSTRACT. The relevance of travel time information will remain in future driving environments, as long as it meets accuracy requirements. Current travel time information systems usually rely on estimations obtained from spot-speed methods. Also, direct measurements of travel time are more and more used, according to the increasing penetration rate of the related technologies. However, in none of these cases is the information disseminated that expected from a real-time information system, i.e., accurate predictions of travel time. Researchers and private companies have developed methodologies able to provide such predictions. However, they are not generally implemented in practice because of their complexity or the unavailability of the necessary technology. This paper proposes a method that predicts travel times in real time, while being immediately applicable in most roads. Predictions are estimated from the spatial information given by the vehicle accumulation in each section, which is obtained from input-output diagrams. Previously, the original cumulative curves are corrected from drift by means of an algorithm that fuses the available direct and indirect measurements of travel time. The goodness of the methodology has already been proven by the authors when direct measurements are ATT. Now, its suitability when working with ITT will also be demonstrated.

ABSTRACT. This paper focuses on the optimization of a traffic light in the Kadıköy area –one of the central business districts in the polycentric form of Istanbul city- that is used by excessive amount of pedestrians. In the Kadıköy area of Istanbul, there has been an immense crowd at the coastline in either peak or off-peak hours because of the marine transit terminal. When the Kadıköy-Kartal subway’s terminal station is constructed at the heart of Kadıköy in 2013 without the crowd feasibility and analysis, it has become an inextricable situation for both vehicular and pedestrian traffic. In the studied area, there is also a tram line conflicting with the crosswalk that is specifically analyzed. When the pedestrians have a crossing gap, most of them make the decision of crossing without considering the signal phase. Likewise, when it is a pedestrian clearance phase, there can be situations where all the pedestrians cannot cross the street because of high density and insufficient green time.

ABSTRACT. The objective of the study is to analyse level of service (LOS) of urban roads from the perspective of road users. The Indian mixed traffic constitutes different type of vehicle classes, the behaviour of road users will also change with respect to the vehicle class they use for travelling. Hence the study models the variation of LOS perceptions with respect to different vehicle class travellers on urban roads. Video laboratory method was carried out for the collection of user’s perception data, in which the study participants is shown a series of video clips of various operating conditions and ask them their perceptions of LOS. Socio-economic information and travel habits were also collected from the study participants along with the video clip ratings, for the modelling purpose. Ordered probit model was selected for modelling the traffic operational quality in terms of quantitative traffic and road user characteristics. So this paper focusses on analysing the road users’ perception with respect to different categories of vehicle classes and identify characteristics that determine realistic urban road LOS in mixed traffic conditions. The present study will help the traffic engineers to understand the complex relationship between LOS and road-users, for planning the facilities more efficiently.

ABSTRACT. To provide the optimal allocation of requests to vehicles Routing algorithms typically assume the availability of complete and correct information about point-to-point travel times. In real applications, reliability of travel times is crucial, to avoid compromising the optimality and, more important, the feasibility of vehicle routing solutions, which could affect customer satisfaction or even violate contract terms, jeopardizing the overall service quality. Actually, non-recurrent events and traffic conditions make the estimation and the prediction of such travel times a difficult task, further complicated in real-time applications by the dynamicity of the information and the amount of needed estimates. In this paper we present a complete methodology to achieve an extremely fast computation of point-to-point travel times based on a dynamically updated and forecasted information of traffic conditions on a large network.

ABSTRACT. Central areas of large cities offer in general many advantages to their inhabitants. Typically, a large number of products, services, and opportunities are available in those urban zones, thus increasing life quality. Unfortunately, these benefits are associated with increasing transportation activities that can cause serious problems, such as traffic congestion, excessive energy consumption, and pollution. This paper aims at presenting a new transport system that consists of transporting freight in long-haul passenger vehicles. Two mixed integer mathematical programming models are presented: one for total cost minimization and the other for the travel time minimization. The problem under study was considered as a multi-commodity network flow problem with time windows, multi transport-lines, and multiple vehicles. Three heuristics based on mixed integer programming (MIP) were designed to solve it: size reduction, LP-and-fix, and a combination of these two procedures. The proposed approaches were validated in a case study designed around the intercity passenger transport system, in Ceará, Northeast of Brazil. Several operational scenarios were evaluated, taking into account the available freight capacities. The developed MIP heuristics produced high-quality solutions, in reasonable computational times, with the LP-and-Fix algorithm outperforming the other approaches.

ABSTRACT. Two-tier distribution systems have been proposed in the context of city logistics to mitigate the negative externalities of urban freight transport, particularly for larger and highly congested urban areas. In this paper, we focus on two-tier urban freight distribution systems based on mobile depots where little or no physical infrastructure is considered and where storage is not permitted. In these types of systems, coordination and synchronization between vehicles are essential, and the main objective is to have vehicles at the transfer locations in a need-to-be basis as a way to minimize their negative externalities. We propose a classification for these systems according to the level of mobility and accessibility of transportation modes operating at the first-tier. Furthermore, we characterize the main components and operational characteristics of the different systems, including the existence of multi-trips, the types of transport modes used, transported loads and the characteristics of the transfer sites (satellites). This work will hopefully contribute for a clearer characterization of two-tier urban distribution systems, to be later used in developing mathematical models for their design and planning.

ABSTRACT. The assessment procedures of city logistics scenarios require the simulation of freight transport demand and hence the estimation of freight vehicle origin-destination (O-D) flows. These O-D flows can be obtained from the simulation of delivery tours. Therefore, the paper recalls a system of models able to simulate delivery tours using an aggregate approach proposed by the authors and presents such an advancement in sub-model specification and calibration. It uses a two-step procedure for simulating the choice of tour pattern (e.g. single/multiple direct or single/multiple peddling) and the decision of number of stops (trips) made in a selected tour pattern. The data used in this study are from a set of commercial vehicles operating in the Veneto Region (Italy) for which floating car data (FCD) of 60 operating-days were available. The analysis allowed us to investigate the current patters of delivering tours in the Veneto Region and to point out relationship with socio-economic characteristics and land use of the zone served.

ABSTRACT. The arrival of automated vehicles could significantly reduce the operating cost of mobility services. This fact has encouraged researchers to propose door-to-door services instead of the current fixed routes. However, a comparison between these two alternatives is required in order to identify when (depending on the development degree of the automated vehicles) and where (depending on the characteristics of the area of service) the implementation of each service is the most competitive solution. This research compares the two types of transit services to supply first/last-mile solutions in suburban areas. By means of an analytical approach, the results show that fixed routes remain the most efficient alternative unless the new technology reaches a certain degree of development that allows a high reduction of operating costs. In this case, the applicability of door-to-door services will significantly increase under certain circumstances: small areas of service, short distance trips and high values of time.

ABSTRACT. The concentration of both employment and services in a specific area of a town generates positive effects, but also impacts. Some of them are congestion of a large city center, inefficiencies in public transport systems, and others. Many countries face these impacts in which their cities are large and disperse with a central business district that concentrates the majority of activities. Urban subcenters seek to approach economic activities to peripheral urban spaces. The objective of this research is to analyze and evaluate the feasibility of implementing urban subcenters. The analysis is based on a continuous approximation model that considers a total cost function, including users and agency costs: The model can compare urban structures of different transit modes (subway, tramway, bus rapid transit). The model addresses with mathematical optimization (Karush-Kuhn-Tucker conditions) on a circular city with circular and radial roads. The model is applied to several urban subcenters scenarios, and it is applied to a real case: Santiago, Chile. The results of the modeling show that implementation of subcenters generates interesting improvements in the functioning of public transportation systems. These outcomes will be better if the transit is adapted to optimal urban schemes for obtaining consolidated subcenters.

ABSTRACT. Public transport services are currently executing or planning a fundamental transition from traditional buses to electric buses. During this transition phase, the public transport offering is fulfilled with a mixed fleet across multiple bus terminals, which poses operational challenges for optimal vehicle scheduling, a problem not yet addressed in literature. As researchers in Transport Engineering and Operational Research at the University of Luxembourg, in collaboration with the Roma Tre University, we support the Ministry of Transport of Luxembourg and Volvo buses by modelling and simulating this transition phase, to help them manage and solve such challenges. In this work we develop a mathematical model of the problem and implement a time-based decomposition framework, through which we can optimize real-life daily instances. This method is tested using the main urban bus lines that connect Central Station, Findel Airport and three other major terminals within Luxembourg City. The objective is applying our approach to a wider set of bus lines of the urban system, providing (near) optimal solutions that explicitly consider the energy constraints arising from electric bus operations, while establishing an advantageous trade-off between delaying trips, to implement quick-charging of electric buses, and performing the same trip with costlier traditional buses.

ABSTRACT. This paper presents the service optimization problem of three different on-demand transit systems operated by taxis and vehicles with semiflexible routes. This problem is aimed at minimizing the total cost of the system that consists of the temporal cost experienced by users and the operating cost incurred by transit agencies. A compact set of estimations of the user performance and operating cost is provided, based on geometric probability. The optimization tool calculates the optimal time headway (or maximal waiting time in case of taxis) and the total cost per passenger served. Moreover, the most efficient on-demand transit service can be easily identified in each demand domain. Results show that taxi systems are only preferable for low demand densities (λ_d<30 pax/km2-h). In other demand domains, bus systems with low flexibility in the layout present the lowest operating and total cost per passenger. The estimation of unit operating costs allow decision makers calculating the subsidies needed to make the system profitable for transit agencies

ABSTRACT. The model presented in this work is based on the agent-based simulation framework MATSim. We describe a new approach for the development of a MATSim scenario. In cooperation with the City of Hanover, available data was centrally collected, analysed and clustered. Using data fusion, the dataset was combined and enriched with additional information from open data sources in order to improve the level of detail of the model. In combination with the German mobility survey “Mobilität in Deutschland 2017”, the developed approach is specifically adapted to the regional characteristics of the city of Hanover, but yet transferable to other scenarios.

ABSTRACT. The concept of logistics tender or request for quotation is gaining importance among the logistics and transportation companies, since many large manufacturers are using this reverse auction system to transport their products worldwide. This paper describes a real-life case in which historical data is used to develop machine learning models able to predict the probability of success in a given tender based on the actual level of some decision variables. Some preliminary results are given, and a discussion on how these models can be integrated into optimization models is also highlighted as potential future work.

ABSTRACT. The Kernel Logistic Regression is a popular technique in machine learning. In this work this technique is applied to the field of discrete choice modeling. This approach is equivalent to specifying non-parametric utilities in random utility models. A Monte Carlo simulation experiment has been carried out to compare this approach with Multinomial Logit models, comparing the goodness of fit and the capability of obtaining the specified utilities.

ABSTRACT. Short-term urban traffic forecasting is an important problem of transportation engineering. Many modern forecasting models utilise information about a spatiotemporal structure of traffic flows – relationships between flow characteristics at distant road links that appear with time delays. Accurate identification of this structure is critically important for models’ forecasting accuracy and interpretability. This paper proposes application of an ensemble learning technique for learning the spatiotemporal structure. The proposed ensemble combines three spatiotemporal feature filtering methods that widely used in traffic modelling – travel time-based, which utilises information about road connectivity and travel times between road segments, cross-correlation-based, which uses the correlation structure of dependencies, and a graphical model-based, which discovers conditional relationships between traffic flows. The resulting ensemble is used for specification of features in the spatially regularised vector autoregressive model and applied to a real-world data set (Minneapolis, USA). Extensive experiments demonstrate promising results of the proposed ensemble learning in terms of model forecasting accuracy, robustness of estimated structures and parsimony of resulting model specifications.

ABSTRACT. This work proposes stochastic optimization for the airline fleet management problem, considering uncertainty in the demand, operational costs, and fares. In particular, a multistage tree is proposed, compounded of strategic and tactical nodes. At the former ones, fleet composition decisions are made, while at the latter ones, aircraft assignment decisions are formulated. Computational experiments are based on a small air network with seven strategic nodes and fourteen tactical nodes (i.e., seasons) where two fleet types are available to be included: Airbus 320, and Boeing 737. These results provide the optimal fleet planning and assignment at both strategic and tactical scopes. Finally, it is shown the superior performance of the stochastic version of this problem against the deterministic one.

ABSTRACT. Incoming air traffic in a given airport can be provided by a great diversity of air routes. However, airports comprise a limited number of runways. The reduction in the number of paths that aircraft can transit takes place in terminal arrival routes, which act as an interface between incoming routes and approach trajectories. This occurrence entails that air traffic has to be carefully managed in terminal arrival routes in order to prevent possible bottlenecks. This work presents an optimisation model that manages not only approach and landing operations, but also terminal arrival routes, deciding on runway assignment, terminal arrival route selection and aircraft trajectory. The proposed integrated model leads to a mixed integer non-linear problem. For its resolution, a Benders decomposition is proposed. On the one hand, the master model deals with runway assignment and terminal arrival route selection, making use of a set of binary variables. On the other hand, the sub-model deals with the trajectory calculation problem, managing a set of continuous variables and minimizing a combination of fuel consumption and delay.

ABSTRACT. When an unplanned disruption causes the temporary closure of an airport, incoming flights are re-routed to one (or more) nearby ones. As a consequence, traffic in the alternate airport increases and the efficiency, punctuality and regularity of operations may be compromised. The purpose of this work is to determine the impacts on the alternate airport airside operations due to the presence of diverted flights. If the number of aircraft to be serviced increases, ground handling operators are subjected to an additional workload, probably resulting in delayed departures and knock-on delays. A discrete-event simulation model of both aircraft landing-and-takeoff (LTO) cycles and turnaround operations is built by using AnyLogic. The model is applied to the case study of Lisbon “Humberto Delgado” airport. When the number of incoming flights increases upon a certain threshold, departure delays spread over the day, which should call for emergency actions and contingency plans.

ABSTRACT. This paper presents results of simulation analysis of demand elasticity effects arising as a consequence of major road improvements in urban areas. These projects, often desired to bring network-wide benefits, also induce additional traffic volumes which eventually deteriorate traffic congestion. Our objective is to investigate how this might be prevented with additional changes in road network to preserve key gains provided by new road investment. We present results for selected case studies of major road projects currently underway in the second-largest Polish city of Krakow. We carry out investigation in macrosimulation transport model, and highlight the impact of demand elasticity phenomena - firstly, how it indeed induces traffic volumes which “fill up” the space apparently freed up by new ring road projects – and then secondly, how further reduction of road capacity leads to final traffic congestion relief. As shown in our analysis, certain measures (indicators) might be potentially useful in estimating the “suppresion” of induced traffic effect – i.e. related to network travel times and incremental costs of induced traffic. Findings from our study could be used to derive a generic method for evaluating urban road system development scenarios, and be of practical relevance for road investment appraisal process.

ABSTRACT. This paper presents an assignment model able to reproduce mixed traffic flows made of automated vehicles (AVs) and conventional ones (CVs). In order to apply to large urban networks, such a model is specified in the framework of the meso-simulation approach. First results of a test case application to a network of realistic dimensions are reported to show the capability of the proposed assignment model to support the assessment of future mobility scenarios characterized by the presence of both AVs and CVs.

ABSTRACT. In this paper, we analyze the relation of penetration rate of vehicles with Cooperative Adaptive Cruise Control (CACC) and three parameters, i.e Total Time Spent, Number of Stop-and-Go Movements and Total Emission, in three test networks, which have different levels of complexity and different traffic control strategies. Our analysis shows that the possible effects of CACC on environment and traffic in near future is distinguishable. We further discuss the compatibility of existing urban traffic control management strategies with CACC.

ABSTRACT. In this paper, we propose to extend the regional dynamic traffic assignment framework for Macroscopic Fundamental Diagram traffic models discussed by Batista and Leclercq (2019), to account for bounded rational drivers. We consider drivers with indifferent preferences and with clear preferences for more reliable travel times. Monte Carlo simulations are used to account for explicitly calculated distributions of trip lengths as well as the regional mean speeds. The Method of Successive Averages is used to calculate the regional network equilibrium. Preliminary results considering bounded rational drivers with indifferent preferences are discussed.

ABSTRACT. The mobility data derived from mobile phones may provide hints regarding land-use. Activity zones, be residential or productive, feed the global mobility once acting as origin and/or destination of trips. This research presents an approach to characterise the predominant activity of the sectors of a case of study, the metropolitan area of Malaga (Spain), using mobility patterns. The methodology is tested and compared with the socio-economical information provided by the Official General Statistics and Economic Information in order to quantify the reliability of the approach.

ABSTRACT. A widely used taxonomy for train timetabling models, classifies them according to the point of view used, in: i) operations centric models and ii) passenger centric models, which tries to maximize customers satisfaction. This work proposes a passenger centric timetabling model based on machine learning techniques. The proposed model uses the concept of strategy of a passenger as a set of lines to satisfy his/her particular origin-destination trip. A demand forecasting model has been developed based on kernel functions and it allows the estimation of the number of passengers in each strategy of the network, taking into account the existing correlation between strategies. The resulting model captures the elasticity of demand with respect to the characteristics of the designed timetable such as price, travel time, arrive time, etc.

ABSTRACT. OD flows provide important information for traffic management and planning. The prediction of dynamic OD matrices gives the possibility to apply anticipatory traffic management measures. In this paper, we propose an OD prediction approach based on the data obtained by Automated Number Plate Recognition (ANPR) cameras. The principal component analysis (PCA) is applied to reduce the dimension of the original OD matrices and to separate the main structure patterns from the noisier components. A state-space model is established for the main structure patterns and the structure deviations, and is incorporated in the Kalman filter framework to make predictions. We further propose three K-Nearest Neighbour (K-NN) based long-term pattern recognition approaches. The proposed approaches are validated with field ANPR data from Changsha city, P.R. China. The results show that the observed OD flows can be accurately predicted by our proposed approaches. Which prediction method performs best depends on the quality of the available data: for regular, periodic OD matrices the Kalman filter is better, for irregular OD matrices the pattern recognition, that looks at different time periods in the historical data, gives better results.