ABSTRACT. The world we live in is constantly changing. Playing to Aveiro’s strong legacy of technological innovation, the Aveiro STEAM City project is changing the way city actors perform and interact, differentiating it from its competitors at home and abroad.

Aveiro is today a living trial city for intelligent mobility and transportation, with an advanced large-scale communications infrastructure, spread throughout the city. It is composed of fibre link technology (spread across 16km), reconfigurable radio units, 5G-NR radio and 5G network services. The access infrastructure covers 44 strategic points in the urban area of Aveiro, in the form of smart lamp posts or wall boxes on building facades, with mobility sensors such as radars, lidars and video cameras. Buses and garbage collection vehicles have also been equipped with communication units and sensors, which currently record mobility and environmental data, making a complete live map of these parameters in the city, and providing the required data for traffic monitoring and safe driving systems.

This talk will address the living lab, the services it can support, with a special emphasis on mobility and transportation, and how the interaction with citizens has been achieved.

ABSTRACT. The remarkable advancements in their sensing, computing and communication capabilities make vehicles and road infrastructures able to generate and collect an enormous amount of information, which fosters the development of data-driven models for context-aware decision making. In this work, we leverage Deep Reinforcement Learning (DRL) and develop an adaptive cruise control (ACC) application that aims at ensuring a safe, comfortable, and efficient driving experience. Our DRL framework accounts for, and properly weights, the different environmental conditions playing a role in adaptive cruise control, including vehicle stability. We evaluate and compare the performance of the proposed framework against standard ACC, by integrating it in the CoMoVe framework, which realistically represents vehicle communication, mobility, and vehicle dynamics. The results show that our solution provides very good performance in terms of safety, comfort and vehicle stability, under different traffic scenarios. Furthermore, they highlight the important role of vehicle connectivity in gathering additional data on the surrounding environment, hence improving the performance of the DRL scheme.

ABSTRACT. The paper introduces a novel approach to the classical adaptive traffic signal control (TSC) problem. Instead of the traditional optimization or simple rule-based approach, Artificial Intelligence is applied. Reinforcement Learning is a spectacularly evolving realm of Machine Learning which owns the key features such as generalization, scalability, real-time applicability for solving the traffic signal control problem. Nevertheless, the researchers' responsibilities become more serious regarding the formulation of state representation and the rewarding system. These Reinforcement Learning features are also the most fascinating and controversial virtues since the utilized abstractions decide whether the algorithm solves the problem or not. This paper proposes a new interpretation for the feature-based state representation and rewarding concept that makes the TSC problem highly generalizable and has scaling potential. The proposed method's feasibility is demonstrated via a simulation study using a high-fidelity microscopic traffic simulator. The results justify that the Deep Reinforcement Learning based approach is a real candidate for real-time traffic light control.

ABSTRACT. The rollout of driverless cars (DC) may reshape human mobility drastically. Urban roads may face a surge of car traffic if DC technologies continue to enhance the use of private cars. Shared DC use can remedy this by reducing the number of urban trips that non-shared-use DC could otherwise make. While recent research has proven the importance of socio-economic factors on shared DC use, the influence of present sharing behaviour, personality traits, and social norms on shared DC use has not been extensively explored to date. To address this gap in this study, we employed a panel of 500 Edinburgh-based respondents, through an online survey, to examine the likelihood of accepting the non-shared-use and shared-use DC options for regular urban trips. We have collected data on respondents' present carsharing and ridesharing behaviour, personality traits, social norms, and sociodemographic characteristics. To elicit the impact of these factors on the likelihood of accepting non-shared-use and shared -used DC options, ordered probit models were estimated. The model findings imply that frequent household car users and those influenced by social expectations to preserve the environment are willing to use non-shared-use DC. In contrast, city centre dwellers, cooperative and younger adults with sharing attitudes show a higher tendency towards shared DC. High-earning, working-aged and young respondents are more inclined to use a driverless taxi, whereas city-centre dwellers and those influenced by social expectations to share personal resources are more favourable towards ridesharing with a stranger in DC. These results can assist the policymakers, and transport planners shape policies for promoting shared DC use and transport service providers to deliver and operate shared DC fleets efficiently.

ABSTRACT. Autonomous Vehicles (AVs) can potentially reduce the accident risk while a human is driving. They can also improve the public transportation by connecting city centers with main mass transit systems. Creating a system that can provide a sense of security to the passenger, when the driver is missing, remains a challenging task. In this work, an image and audio-based approach, supported by novel AI algorithms, is proposed to increase the security and trust inside an autonomous shuttle. The two modalities, running in real-time, can detect petty crimes scenarios such as screaming, bag snatching, people fighting and vandalism and enable notifications to authorized personnel for proper actions.

ABSTRACT. Bus bunching is one of the main issues impacting bus service operations. It affects the reliability of bus services impacting waiting times at stops and risking user dissatisfaction with the travel experience. Although public transport is to be encouraged as the most environmentally friendly mass transit solution, Bus Bunching is a factor which can lead to use of private transport. Particularly on high-frequency, high intensity routes, headways can vary. A small delay causes an increase in the number of passengers at the next stop. This in turn leads to an increase in the dwell time (bus service time at a stop) and consequently it further increases the delay. At the same time, the next bus will have fewer passengers, shorter dwell times and will gradually catch up with the preceding bus, this being the most significant but not the only cause of bus bunching which is also the result of traffic flow, driving behavior and other, often random, factors. The advancement on Information Communication Technologies can offer new opportunities for transit operators to limit such an effect, ensuring a more attractive service for users. In this context, the paper, leveraging the opportunities offered by the innovations in ICT, proposes a machine learning-based procedure for addressing the issues causing bus bunching. The procedure is then applied to a real test case with positive results. As a result the proposed procedure may be incorporated in a decision support system to improve bus service operations.

ABSTRACT. People usually travel with different preferences to reach the location of their leisure activities, where transport mode is a decisive factor. Previous studies focus on developing discrete choice models for people where a trip purpose is considered as an explanatory variable. In this paper, discrete choice modeling is applied to model the behavior of leisure travelers, to understand the impacts of several factors on the transport mode choice, such as sociodemographic variables. A sample of 1100 travelers from a household survey in Budapest is used in the analysis. The sample includes the daily activity plans of leisure travelers and their sociodemographic as well as economic characteristics. A Multinomial Logit (MNL) model is applied, where the data are examined across travel time variables and travel characteristics, such as travel time, travel cost, gender, income, and car ownership. The developed models are used to predict the probability of using the different transport modes based on the characteristics of the travelers and their trips. The output of this study is important for decision-makers to understand the patterns of leisure tips in a city.

ABSTRACT. Travel using public transportation offers financial benefits to the people. However, reaching a destination by using a single mode of public transport seems to be difficult in metropolitan cities and results in taking of transfer at an interchange. As a result, the transfer facilities and their characteristics affect city dwellers' perception to ride using Public modes of transport. Also, an individual’s decision to ride in a multimodal public transport system is also governed by socio-demographic, travel as well as public transport characteristics. With this background, the present study aims to understand the factors influencing people's behavior to use public transport, including transfer for the Surat city. Based on 752 non-transit users' responses, the Structural Equation Model was developed using AMOS software. The results revealed that factors like ‘monthly income’, ‘in-vehicle time’, 'seat availability', 'out-vehicle time', and ‘ punctuality of time transfer’ play a crucial role in willingness to use multimodal public transport system.

ABSTRACT. A smart and sustainable city views public transport as a policy priority. Tools to support the decision-making of municipalities and bus operators are needed. The present study applied a generalized linear model using speed bus as dependent variable to assess the effect of factors related to time and urban space. Data from four bus lines in the city of Porto was considered, showing that all the analysed variables have impact on speed. The study findings confirm the relevance of bus lanes and traffic signals as spatial priority schemes. Overall, the present study leads to relevant conclusions that can support the decision-making of transit agencies and public transport operators. Additionally, the speeds obtained by the model can be used for calibrating simulation models.

ABSTRACT. Waiting time plays an important role in the cyclists' route choice, most likely because cyclists, after a stop, need to pedal harder to regain their previous speed. Literature review highlights that cyclists generally overestimate waiting time approximately three to five times higher than their actual waiting time. The estimation of the cyclists' waiting time is a complex task, as it depends on many factors such as infrastructure attributes, the cyclist’s individual behavior and even the phase during which the cyclist arrives at the traffic light. The aim of this paper is to validate a recent algorithm that estimates cyclists’ waiting time from GPS traces, using data obtained by a manual survey, and successively apply it to a big data sample of 270,000 GPS traces recorded in the city of Bologna, Italy, thus quantifying how the cyclists' characteristics and network attributes influence the cyclists' waiting time, which has not yet been addressed in literature due to the difficulty on estimating cyclists waiting times from GPS traces. Another manual survey allowed to both test the representativeness of the used big data and monitor how many cyclists pass with the red signal on different type of signalized maneuvers. Results show that waiting time represents a not-negligible share of the bike trip (11\% of total trip duration). On average, higher waiting times have been recorded on complex intersections by cyclists younger than 25 years old, by less frequent cyclists and by women. During rush hour, cyclists have recorded waiting times only 6\% above the daily average, demonstrating that traffic congestion has a limited effect on waiting times. Furthermore, approximately 14\% of all cyclists have crossed signalized intersection with the red signal, especially when the opposite traffic volume is not high and there is good visibility.

ABSTRACT. Ridesharing, being part of shared mobility, is rapidly improving its diffusion, due to the use of apps and smartphones, increasing the sustainability of transportation systems. However, informal ridesharing continues to be an alternative solution for users, to reduce the total travel costs, despite it is not managed by platforms. To investigate people’s behavior about shared mobility, we consider the 2017 Scottish Household Survey. Selecting users’ aspects usually relevant concerning such travel modes, and setting binary logistic regressions, the strongest bond emerged involves the non-possession of private vehicles and the usage of informal ridesharing (odds ratio equal to 16.504). This result indirectly confirms the important effect of ridesharing in the reduction of car ownership, underlining the interest to better understand the impact on mobility of such habitude.

ABSTRACT. Recent sustainable mobility scenarios, aimed at the decarbonisation in urban and non-urban contexts, seek to promote an increasing use of electric vehicles (EVs) or hybrid vehicles (PHEVs) by 2030, considering their attractiveness not only for private use but also as shared mobility alternatives by different categories of users. It is significant to investigate the binomial formed by users' characteristics and travel habits as influencing factors of the spread associated to these innovative services. This paper presents a descriptive and two-dimensional statistical analysis to assess the acceptance of shared e-mobility linked to the user’s profile. The variables have been investigated via an on-line questionnaire, in applying the Likert scale and a single and multiple bivariate analysis have been performed. The case study of Enna city is presented, characterized by a size typical of Italian small urban centers, with critical issues in terms of the transport services supply. A large sample of under 40 years has been selected assuming their greater propensity to use this innovative transport modes. The results highlight the correlations between the socio-demographic data, the choice of the vehicle power supply and the influence of the characteristics of the service. These factors, along with the mobility patterns, must be taken into consideration for the improvement and promotion of the service.

ABSTRACT. Floating car data (FCD) have recently become a popular tool of urban traffic engineering. Conventional FCD contain a series of probe cars’ timestamped locations and are used to estimate traffic speeds and travel times and identify congestions. In this study, we propose the enhancement of conventional FCD with car vision information: traffic measurements, collected by video cameras or lidars, installed on the probe cars. Given the limited penetration rates of probe cars, such vision information can significantly improve the accuracy of traffic estimation. Our experimental study is based on two simulated vision-enhanced FCD data sets: sensor-based real-world traffic data set with simulated observability and a simulation model with vision-equipped probe cars. We estimate the potential of vision-enhanced FCD for urban traffic flow estimation for different probe car penetration rates. A recently proposed temporal geometric matrix completion algorithm is utilized for traffic speed estimation given incomplete spatiotemporal traffic flow information. Empirical results show that the availability of vision-enhanced FCD leads to significant improvement: the coverage of 6% of spatiotemporal slots by probe cars gives reasonable, and the coverage of 9% – nearly optimal accuracy of traffic speed estimation. Thus, obtained experimental results support our hypothesis about the utility of vision-enhanced FCD to improve traffic estimation accuracy and discover related problems and limitations.

ABSTRACT. Geolocation data identifies the geographic location of people or objects, which may unveil the performance of some activity or operation. A good example is, if a vehicle is in a gas station then one may assume that the vehicle is being refuelled. This work aims to obtain vehicle-based operations from geolocation data by analysing the stationary states of vehicles, which may identify some motionless event (e.g. bus line stops and traffic incidents). Ultimately, these operations may be analysed with Process Mining techniques in order to discover the most significant ones and extract process related information. In this work, we studied the application of diverse approaches for detecting vehicle-based operations and identified different operations related to the bus services. The operations were also characterized according the distribution of their events, allowing to identify specific operations characteristics. The public transport network of Rio de Janeiro is used as a case study, which is supported by a real-time data stream of buses geolocations.

ABSTRACT. The viability of high-mileage vehicles, like taxis, in acting as potential probes for Vehicular Crowd-Sensing (VCS) has been largely confirmed in many experimental studies. However, these studies have been mostly carried out considering data from cities with regular, grid-based, road networks, or by abstracting the road network to a grid of cells. In this paper we investigate the potential suitability of taxis as probe vehicles, by evaluating the achievable spatio-temporal sensing coverage, computed over real trajectories from a swarm of 100 taxis in the city of Porto (PT). Our results confirm that as few as 100 taxis have the potential to effectively sense complex urban road networks, for many VCS-based use cases. On the other hand, we found that the probing frequency might be inadequate to support use cases requiring higher sampling rates. As a consequence, recruiting more vehicles and/or devising specialized routing/incentitivazion mechanisms might be necessary.

ABSTRACT. Life is a complex phenomenon, so is the task of driving. Mostly human performance is controlled by sympathetic and parasympathetic systems functioning in human body. It is quite dynamic in response to numerous stimuli the driver acquires from the human-vehicle-environmental ensemble. Geometry is one of the fundamental stimulus for a driver, driving on a non-urban highway, especially when there is minimal interference from other vehicles. Consequently, it is quite natural that drivers adopt higher speeds. His/her mental workload will be based on the input he/she gets from the above system and the rate of reception of input will be more with speed. Any inability to cope with the workload may lead to occurrence of crashes. Hence, it is very important to take into account the influence of geometry on driver workload while designing highways. One way to assess the workload is to capture the physiological responses of a driver during driving.

Objectives This study explores the factors that significantly influences driver workload. Secondly, it studies whether the factors are the same for both car as well as bus drivers. There are several physiological measures which can be measured on a driver. Heart rate, galvanic skin resistance and rate of eye blinking are a few measures. The study incorporates these measures along with their deviation from base condition as candidate measures for workload. Thirdly, interval estimate regression models are developed for predicting car and bus driver workload.

Scope Scope of the study is limited to horizontal curves of two lane non-urban highways of Kerala. Curves with sufficient preceding tangent length are considered.

Methodology

The study included driving experiments done on 114 horizontal curves of gradient less than 2 percentage, each curve being driven over by 30 car and bus drivers. Subjects were equipped with sensors for collecting workload measures. They were given a static cooling period as well as base period. Cooling time helps to get familiarize with the body attachments and base period data is used for normalizing the physiological measurements. GPS coordinates were continuously logged along with the workload data. Radius of curve, length of curve, length of tangent section, superelevation at curves, degree of curvature, deflection angle and minimum available sight distance at curves were the geometric variables considered. The physical characteristics of drivers were measured through questionnaire and laboratory experiments. It includes age, driving experience, previous crash history, health issues, reaction time, depth perception, peripheral vision and vision acuity.

Analysis and Results Various analysis were done on the data to explore the solutions to various research questions. Analysis of Variance study was done in search of answers to following questions: 1. Whether heart rate of young and old drivers are significantly different or not. 2. Whether driving experience has any influence on heart rate while driving. 3. Whether there exist significant difference in the heart rates of a driver by profession and other drivers. 4. Whether workload of car and bus drivers are significantly different

Correlation study and scatterplot analysis were conducted to identify the variables which are significantly correlated with workload measures. Further, interval estimate of regression analysis was done. Sight distance is the most significant variable for car drivers were as sight distance and deflection angle are significant variables for bus driver workload. Heart rate, deviation of heart rate, galvanic skin resistance and deviation of galvanic skin resistance are the best candidate measures for driver workload.

ABSTRACT. Traditionally, Random Utility Maximization (RUM) models have been widely applied to travel mode choice modelling. Currently, Machine Learning (ML) models are being applied as an alternative to RUM models, since they provide better results in terms of prediction capability and they can manage large volumes of data. In this paper, a comprehensive comparison between classic RUM models and ML models, including single and ensemble classifiers as well as Deep Neural Networks (DNNs), is provided in order to assess systematically the performance of different models over two different datasets which have different sizes and nature of data. Numerical experiments show Random Forest (RF) is the best classifier in terms of accuracy index and the computational cost to train the model.

ABSTRACT. Many traffic-related applications, e.g. traffic demand modeling, rely on conventional data collection methods such as travel surveys. These methods can be demanding in terms of cost and time, which results in low network coverage and limited representativeness. On-motion sensors, e.g. smartphones, offer the opportunity to replace such methods and compensate for the aforementioned drawbacks by collecting positioning data automatically. GPS data consist of positioning records each of which has geographical coordinates and a timestamp associated. These data, however, require cleansing and processing before being put into use. Information about the used transportation mode is missing from this kind of data unless travelers were specifically asked to report it. In the literature, supervised machine learning (ML) algorithms were successful in inferring transportation modes from GPS data. However, these algorithms, unlike unsupervised ML algorithms, require training data that are not always available. This paper aims to investigate the capability of unsupervised ML algorithms to infer transportation modes from real GPS data extracted from smartphones. Therefore, we used two datasets to benchmark different unsupervised ML algorithms with different input attributes. The paper also investigates the feasibility of using a pre-trained model for unlabeled real data. Finally, we compared the best performing unsupervised setup to the supervised ML algorithms recommended in the literature. The results suggest that the recommended unsupervised setups can reach an overall inferring accuracy of 93%.

ABSTRACT. In order to enhance the efficiency and reliability of bus transportation systems, it is important to predict travel time precisely in the planning phase. With the precise prediction of bus travel times, the cost can be reduced for bus companies, such as from planning fixed buffer times between trips, while a better service can be provided for passengers. In this study, historical bus travel data from a bus line of the HOCHBAHN bus company in Germany in 2019 are used for the analysis. We develop two neural network models, namely multilayer perceptrons and a long short-term memory neural network, for predicting the bus travel time between timepoints for a trip occurring at a given time of day, on a given day of the week, and in given weather conditions. Both neural networks are combined with a genetic algorithm and a Kalman filter to improve accuracy. The genetic algorithm is implemented to tune the neural network parameters, such as the number of hidden layers and neurons in a hidden layer, while the Kalman filter algorithm is used to adjust the travel time prediction for the next trip using the latest bus operation information. In the experiment, we test our models month by month and split data for each month into three parts: the data of the first two weeks for training, one week for validation and the last week for prediction. The experiment results show that the hybrid model is powerful for predicting the bus travel time. In particular, the combination of the multilayer perceptrons with the genetic algorithm and Kalman filter provides the best travel time prediction (with an improvement of 56.2% compared with the planned bus travel time from the bus company). In order to make a recommendation for bus companies to plan vehicles with more accurate bus travel times, we test our hybrid models with larger training sets to predict the travel times e.g. in August. However, due to the structure of the planning problem, the plan for buses should be estimated before a new month begins. We cannot consider the real information during the planning, therefore we apply our hybrid models without the Kalman filter and they provide on average about 26% improvement compared with the planned travel time.

ABSTRACT. Electric micro mobility systems such as e-bikes and e-scooters represent sustainable travel alternative especially for specific classes of trip distances. Moreover, the coverage and accessibility of transit services can be completed through the implementation and promotion of these systems. Aiming to investigate the potential demand that can be moved from private cars to environment-friendly modes (e.g., e-micro mobility), a methodology based on collection of data by probe vehicles is presented. To test its goodness, it is applied to the city of Rome (Italy) with challenging results.

ABSTRACT. The objective of this research is to compare two methods to compute airport capacity: Static Analysis and Dynamic Analysis. For this purpose, the two methods have been evaluated in two different scenarios, the first one is related to normal operations and the other to an exceptional event. In the context of the coronavirus COVID-19 pandemic, they can be considered to coincide with a "Pre-Covid" and a "Covid" scenario. Nowadays, the calculation of airport capacity is very much linked to static analysis, a tested historical method, based on empirical formulas dictated by the International Air Transport Association (IATA). The dynamic analysis is a state-of-the-art method that uses software packages to simulate a great variety of non-ordinary situations. It can incorporate a wide range of information related to the specific case study such as airport layout, entry/exit routes, etc. Despite this, this method is still little used as it does not currently have specific guidelines. The following study evaluates the two methods, highlighting their strengths and weaknesses.

ABSTRACT. The continuous growth of air transport and the shortage of capacity have made multi-airport systems (MAS) important configurations to meet demand in large metropolitan areas. Nowadays, MAS are acknowledged to cater more than 2 billion of passengers internationally, well over half of global traffic. In a multi-airport system, different airports serve the same (or part of the same) catchment area instead of enlarging capacity at only one site. As a result, air transport supply distributes across different airports, and so are passenger flows. From a supply perspective, the presence of multiple airport sites may lead to duplication of infrastructures, losing the benefits of economies of scale albeit reducing airport congestion. In turn, this may be beneficial to passengers by potentially fostering accessibility to air transport services (de Neufville, 1995) and boosting the number of flights. More importantly, it may also enable differentiation strategies at both airport and airline level. Over the last decades, the literature has investigated multi-airport systems from different standpoints, from the analysis of passenger allocation using airport-airline-access choice modelling (e.g., Pels et al., 2000, 2001, 2003; Bas¸ar and Bhat, 2004; Hess and Polak, 2005, 2006; Ishii et al., 2009; Marcucci and Gatta, 2011), to the optimization of MAS operations (Sidiropoulos et al., 2018) including the utilization of runways and airspace and the efficient planning of arrival-departure capacity trade-offs (e.g., Ramanujam and Balakrishnan, 2009; Clarke et al., 2012; Murc¸a and Hansman, 2018; Murc¸a et al., 2018; Schefers et al., 2020; Yin et al., 2020). By contrast, studies focusing on airlines’ service planning within a MAS context have been fairly scant and so are decision support tools aimed at practically supporting airlines in this task. A key issue of airline planning in a multi-airport system is the decisions concerning whether (and how) to diversify services across airports, or rather concentrating them into a single airport. In fact, serving a given market—e.g., Milan-London, Barcelona-Paris, London-New-York—by more than one airport pair has been actually increasingly common for both low-cost and full-service carriers examples. Allocating flights over the different airports that operate in the same MAS can be beneficial for a number of reasons; from a supply perspective, it may be driven by capacity considerations, such as resource constraints at a given airport or the availability of slots, and may also serve as a risk mitigation strategy; from the demand perspective, pursuing a multi-airport supply strategy allows to exploit airport specificities in providing diversified services and, above all, pursuing a better matching of the geographical distribution of demand to reduce the generalized access cost as a key driver of passenger choice. In addressing this topic, we propose an integrated modelling framework to provide decision support to airlines in carrying out frequency planning activities in a MAS context, that is, to determine how many and how to distribute flights across time and airports. First, we develop a grid-based demand model. Relying on passenger location and coupled with air transport service attributes, we investigate the number of passengers travelling from each spatial grid in a given time-window to jointly consider the effects of time-of day preferences, air-service characteristics and access costs. Second, we develop and apply an evolutionary algorithm to solve the airline frequency planning problem in a MAS context explicitly integrating the grid-based demand model. We applied the modelling framework to a three-airport system in Europe. We consider a single business-oriented route operated by a full-service carrier under a semi-monopolistic market structure. In particular, we focus on the corporate frequent-flyer passenger component, which represent the key and most strategic passenger segment on the route. This allows us, first, to focus on a relatively unexplored, but increasingly important passenger category, second, to emphasize the role of planning choices under strong time-of-day preferences for departure time, and third, to calibrate a grid-based spatial demand model leveraging on passengers’ origin location data. Results provide strong evidence that the degree of diversification of a multi-airport system plays a significant role in determining passenger demand. The two-airport structure is found to be the preferred configuration, able to balance the advantages of having a denser service on the territory and the disadvantage of replicating the service in different nearby airports. A three-airport structure indeed implies an over-dispersion of the service, that turns to be detrimental in the eyes of passengers. In a second step, the evolutionary algorithm developed on the basis of our grid-demand model effectively highlights the practical applicability of our framework. In detail, going through numerous MAS supply configurations following the Darwinian principle of the fittest, the algorithm identifies a near-optimal configuration that enhances passenger demand by +9.3% compared to actual values. Ultimately, the implementation of the algorithm further highlights how MAS airport diversification contributes to mitigate the loss of passengers following the closure of one of the airports belonging to the multi-airport system.

ABSTRACT. This paper addresses this cockpit crew training scheduling problem. More specifically, the research objective was to develop a modelling approach that could contribute to the computation of robust crew training schedule capable of dealing with schedule disruptions. Throughout this research, robustness is defined as the capability to deal with or absorb the negative effects of disruptions. This capability is expressed as a combination of scheduling cost and expected recovery cost. To attain our research objective, a modelling framework integrating several models and solution methods were developed and applied. More precisely, the modelling framework consists of a Training Scheduling & Assignment Model (TS&AM), a Disruption Generator (DG) algorithm, a Rule-Based Recovery (RBR) model, and a Neural Network (NN) algorithm. The TS&AM integrates scheduling of courses and assignment of trainees, instructors and simulators. The output roster serves as input for a data-driven disruptions impact simulator, composed by a DG model based on Monte-Carlo Simulation and the RBR model. This disruptions impact simulator generates information about possible disruptions and the respective recovery costs. The NN feedback algorithm learns the recovery costs experienced in the disruption impact simulator and updates these costs in the TS&AM to generate more robust rosters.

ABSTRACT. This paper investigates the contribution of emerging technologies (i.e., internet of things and big data) for providing route advice to goods vehicles driving within cities. More precisely, the implementation and application issues related to the introduction of opportunities offered by emerging technologies in gathering real-time data, in obtaining knowledge and support for providing route suggestions will be pointed out. The main results concern the formalization of the trade-off that it is necessary to activate between the different emerging technologies (starting from internet of things and big data) to improve the learning process of path utility. This process is formalized within the general theory of dynamic transport systems. Thus, the possibility of using multiple technologies in an integrated way is pointed out, identifying the road ahead for researchers and technicians both for specializing the general technologies related to the transport systems, and for improving the current systems that provide the user with real-time and personalized information on the route.

ABSTRACT. Up to now, bulk transports have been carried out via a hub-and-spoke network in the general cargo sector. However, it is expected that the use of autonomous vehicles will enable a more flexible delivery. Such developments may, economically, make sense for shippers. From an ecological point of view, also negative effects can be expected due to enhanced transport performance. For this reason, in the framework of the research project "`Digitalisation in transport sector"' funded by Federal Environment Agency, we investigate the impacts of automation on general cargo transport at the logistics network level as well as the resulting economic and ecological effects. For assessing the impacts of autonomous vehicles on logistics network structures and on freight transport routes ex-ante, an instrument for strategic transport and logistics network planning is needed. We develop an efficient meta-heuristic to find new facilities and adjust the network, while thereby considering the routing characteristics by tackling the large-scale location routing problem i.e. LRP. By the linked approach, we can estimate the exact logistics network and coordinates and measure exact transport distances, driving transport lead times and number of necessary vehicles on the infrastructure network. In the framework of the case study food retail distribution in Germany, we operationalize this approach and evaluate the new optimized network as well as investigate the logistics effects. In fact, this research reveals that the utilization of autonomous vehicles significantly enhances transportation ranges, and the number of tours, while reducing the number of operating facilities. Other impacts such as shifting towards a more direct transportation mode choice have also been observed due to the reorganized network.

ABSTRACT. Tourists are often more vulnerable than residents in sudden disaster situations due to lack of knowledge regarding evacuation routes and safe areas. To establish protocols and schemes for the first step of evacuation to safe zones, it is necessary to gather tourists’ likely behavior during an evacuation. However, there are few actual data available. In this research, we conducted a VR (Virtual Reality) experiment assuming a sudden disaster situation and estimated tourists’ route choice based on the experiment. The experiment was conducted with a newly developed application. In the experiment pictures of intersection in the touristic Higashiyama area of Kyoto, Japan, where shown to participants and they could choose the direction they want to proceed until reaching an open space or designated shelter. As a result, we could quantify the preference in terms of road width and direction (left, right, straight). The results reveal the tendency to select wide roads and to proceed straight. If the participants were put under time pressure these tendencies are intensified. We believe this information can be used to help city planners where route guidance is needed for smooth evacuation.

ABSTRACT. The increasing levels in automated functions in vehicles are expected to reshape the future of road transport. Soon, connected, automated and conventional vehicles will share the roads, and it is essential to understand the impacts of such new road environment under several components such as traffic performance, climate change, air quality. This paper is devoted to present a study focusing on a multi-criteria traffic assignment model based on minimizing travel time, distance travelled and pollutant emissions considering various automated vehicles (AVs) with different driving behaviours scenarios in a road network of conventional vehicles (CVs). For that purpose, a case-study of an intercity corridor was used considering current traffic demand. Results show AVs introduction can in fact significantly contribute to reducing emissions provided their behaviour is a combination in its majority of cautious AV. Results suggest replacing 20% of AV aggressive fleet by cautious yields worst results regarding emissions, even when compared to 100% aggressive AVs. The proposed approach is relevant for decision making, particularly for strategic policy making and planning, and can help authorities achieving their sustainable mobility goals, specially to anticipate the impacts of AVs introduction in the near future.

ABSTRACT. In a railway network, traffic is often perturbed and trains must be rerouted and rescheduled. To do so, we propose an asynchronous algorithm. It is based on the idea of decomposing the problem considering at each step the smallest possible portion of the network and subset of trains. We prove that this algorithm guarantees to find an overall feasible solution if it exists, for networks with specific characteristics.

ABSTRACT. Unexpected events may perturb operations and generate conflicts that must be addressed promptly to limit delay propagation and other negative impacts on the network. The real-time railway traffic management problem deals with disruptions in railway networks, including tracks, deadlocks, and stations. When they happen in station areas, new decisions involving train platforming, rerouting, ordering and timing must be made in real time. This paper explores a mesoscopic approach to deal with disruptions at rail stations. A mathematical programming-based model is proposed to determine re-routing and re-scheduling decisions for railway traffic in a station area. The key steps of the approach, which simulate what happens in real-time traffic management, are: i) an off-line preprocessing stage of the set of feasible routes originally planned, and a second preprocessing stage to analyse the disruption which sets the necessary parameters for ii) an integer programming model that seeks solutions which minimise deviations from planned train schedules and assign new and appropriate platforms (if necessary). Computational experiments show that realistic instances can be solved near to optimality using CPLEX in very short times. This allows to consider this methodology for solving real time traffic management problems.

ABSTRACT. A railway company has to deal with many interrelated planning tasks. In this work we consider the Railway Crew Scheduling Problem (RCSP) for the Rail Cargo Austria (RCA), which is the largest railway company for freight transportation in Austria and one of the largest in Europe. The RCSP aims for determining the most efficient combination of shifts. For RCA's purposes these are shifts for engine drivers. A shift consists of a sequence of consecutive scheduled trips of locomotives and tasks over a given period of time. Each trip must be covered by exactly one shift while operational, legal and labor constraints are satisfied. The RCSP is known to be difficult to solve and there exists a wide range of solution approaches resulting from many research activities. We present and investigate a matheuristic to tackle the RCSP. We use a breadth-first search construction heuristic and formulate a Set Partitioning Problem with the aim of minimizing the overall paid working time, while all RCA specific constraints are incorporated. Although schedule efficiency and employee satisfaction are in general conflicting, a cost-efficient schedule will not be implemented if it does not reach acceptance of the crew. In a computational study we evaluate the proposed approach on it's practical applicability on real-world data provided by the RCA, based on the Austrian railway network. The focus is on analyzing the effects of different conditions on crew schedules. In the course of this evaluation we make use of algorithms from previous work on the Locomotive Scheduling Problem (LSP). The LSP is concerned with assigning locomotives to a train schedule while costs are minimized, and the obtained optimal locomotive schedule forms the input for the RCSP. In combination with our work on the LSP, this work provides a building block for an application for RCA that generates both, a locomotive schedule and a crew schedule.

ABSTRACT. In recent years, the railway system has been facing the various challenges of the “digital age”. To increase its attractiveness, capacity, sustainability, and security, it needs to improve its' everyday operational and planning process. This can be enabled using new generation digitised and automated Traffic Management Systems (TMS). Nowadays, railway dispatchers need a TMS that offers precise and real-time traffic information as fundamental condition for effective traffic management, and whose performance can be further improved by increasing the availability and diversity of sources and data, for which an effective data management platform is required. The EU-funded OPTIMA project is designing and developing a communication platform to manage the connection between several services supporting TMS applications and the TMS applications, and enable their testing. It represents one of the steps required for the development and implementation of a new generation of TMS. This paper described the concept for the OPTIMA platform which will link TMS applications used by railway dispatchers with infrastructure systems such as signalling and interlocking systems, maintenance, and energy management, as well as with data services (such a weather information) with the aim of providing standardised interfaces and common data structures as basis of a common and standardised communication.

ABSTRACT. A thorough understanding of driver behavior is an important step to improve the environmental performance of road traffic. Accurate data analysis tools can be valuable to identify these concerns. The present study explores relationships between driving patterns, tailpipe emissions, and road differentiation by using Principal Component Analysis Biplot. This statistical methodology is suitable to identify patterns hidden in data and can be used as a visualization tool. In this study, the key variables included were: speed, engine speed (RPM), acceleration, and vehicular jerk (first derivative of acceleration), as kinematic variables, and carbon dioxide (CO2), nitrogen oxides (NOx), and VSP (Vehicle Specific Power) mode, as pollutant emission variables. For this purpose, second-by-second vehicle dynamics and tailpipe emissions data were collected in three passenger cars with different powertrains (gasoline, diesel, and hybrid) along with different types of routes (one partly urban/rural and two motorways with variations in traffic volumes) in Aveiro Region (Portugal). Results revealed that Biplots allowed to distinguish different driving behaviors, separate route types (urban/rural from motorways), establish some remarks about emissions, and also present the correlated variables in a single plot. Therefore, this technique can be considered as a useful visualization tool to explore real traffic-related data.

ABSTRACT. The aggregated traffic models based on the Macroscopic Fundamental Diagram represent a promising tool to design strategies for ecological routing and mitigate network-wide emissions. To this end, we must first characterize the relationship between path emissions and distance traveled on aggregated networks, i.e. a regional network. In this paper, we investigate this relationship on a 1-region and 4-regions networks. We utilize an accumulation-based MFD model to mimic the traffic dynamics in the network and utilize the COPERT IV model to estimate the travel emissions, focusing on the carbon dioxide CO2. We show that in some cases there is a linear relationship between the total emissions of CO2 and the average travel distance or travel time of paths on regional networks. However, this is not always true as the traffic dynamics play an important role in the drivers' path choices, i.e. on the network loading, and therefore on the total path emissions.

ABSTRACT. Mobility is vital to the economic development of cities by promoting accessibility for goods and commuters. At the same time, transport systems generate negative external effects. Within the transport sector, road transport accounts for 35% of nitrogen oxides (NOx) emissions, making it the most contributing mean of transport. One way to combat some of the problems associated with the transport sector is platooning. Platooning can be described as a group of vehicles that drive close together, acting as one. This convoy or platoon also has the advantage of reducing the fuel consumption of all the vehicles within. Optimizing platooning formations that encompass light and heavy-duty vehicles that travel the highways will be a necessity for the future of transportation. This optimization can be performed by mixing vehicle type order, intervehicle distance, number of vehicles and speed of the vehicles. Numerical modelling will be used to support the development of an autonomous control system for platoons. The system will automatically advise the formation of platoons with specific data (order, speed and number of vehicles) to minimise road traffic externalities. As drag coefficient plays a major roll on platoons’ fuel consumption at highway speeds, this work takes a deep look at its influence as a steppingstone. In this context, the work aims to analyse the influence of drag coefficient (ζ) as its portraited on the modelling approach of vehicle-specific power (VSP) and consequent vehicle emissions of NOx and carbon dioxide (CO2). Different platooning scenarios are compared in terms of drag and pollutant emission reductions.

ABSTRACT. Vehicles with Internal Combustion Engines (ICE) still represent the most prevalent form of road transport in Europe, being an important source of both greenhouse gases and air pollutants. In response to these concerns, Portable Emission Measurement Systems (PEMS) have been widely used by researchers to measure tailpipe emissions and to detect cheating of emissions regulations by manufacturers. This paper introduces four different approaches to estimate carbon dioxide (CO2) and nitrogen oxides (NOx) emissions for these vehicles. These approaches were based on: i) speed intervals (≤50 km.h-1, 50-90 km.h-1, ≥ 90 km.h-1); ii) internally observable variables (IOVs); iii) vehicle specific power (VSP); and iv) driving volatility indicators. The development of IOVs models was made by testing the most significant parameters on CO2 and NOx emission rates, which included the engine revolution per minute (RPM), manifold absolute pressure (MAP), and intake air temperate (IAT). VSP-modal approach centred on binning emission rates in 14 models that reflects deceleration, idling, cruise, and acceleration states. Driver volatility was characterized by means of vehicular jerk (i.e., first derivate of acceleration) using nine combinations of vehicular jerk types. To obtain real world emissions, data were collected from one petrol and one diesel passenger cars using an integrated PEMS. IOVs and jerk models based on the product of MAP and RPM presented similar CO2 compared to measured values for both vehicles, but they resulted in higher overestimation of NOx than a VSP-modal approach. The proposed methodology can be extended to other individual ICE or alternative fuel vehicles for which it may be expensive to get emissions, engine, and dynamic data.

ABSTRACT. For some time, many studies on maritime transport have been analyzing the issue of the competitiveness of container ports, as they represent the fastest and most immediate access doors to internal markets: in particular, the competitiveness of supply chains in integrated and multimodal container transport is really based on efficiency port nodes and their ability to be competitive. This paper intends to provide a methodology for evaluating the ranking of container ports in the Mediterranean, applying an analysis based on the joint use of Principal Component Analysis (APC) and Data Envelopment Analysis (DEA): the first allows to reduce the number of variables with a few significant factors (without losing significance or representativeness of the phenomenon), while the second allows for the comparison of ports, identifying the most performing ones. Respect to the joint use of PCA and DEA, several authors have used this, also in different fields and specialization. Chen and al. (2010) have used an integrated model to evaluate the operational efficiency of iron ore logistics in China: they used two stages: starting from 15 original indicators, in first stage they used a PCA to reduce to 6 synthetic indicators; then in second stage they used a DEA with 4 inputs and 2 outputs. Poldaru and Roots (2014) used a combined PCA-DEA approach to measure the quality-of-life scores for 15 Estonian counties: also, in this case, the authors have used two stage: first, they analyzed original data using DEA; then, they applied PCA to reduce the numbers of variables and then they applied DEA using new variables (factors). Chandrasekaran and Madhana Gopal (2013) have used the combined PCA-DEA approach to evaluate the efficiency of commercial banks in India, analyzing inputs and outputs separately. Mehdi Tavakoli and Shirouyehzad (2013) have used PCA-DEA to analyze the human capital management in some private banking service organizations: first, using a questionnaire, they used a DEA with several inputs and outputs; then, they used a PCA to reduce the correlations between variables and then again DEA to evaluate the performance of organizational units. The 33 main container ports of the Mediterranean were analyzed, both transshipment and gateway and mixed: for each of these some characteristic variables of the infrastructure were evaluated (quay length, yard size, number of QCs, depths, etc.) and the historical series of TUEs moved from 2002 to today. The Mediterranean has been identified as a test area precisely due to its nature as a "closed sea" of limited width which allows ships in transit along the Suez Gibraltar route to be able to easily use the ports of both the north and south shores, making these ports strongly competitive with each other. The analysis showed that without this procedure, the use of the DEA alone is strongly influenced by the type and number of the input and output variables that are generally used to carry out the ranking between ports, and therefore how the latter vary , the final results are also modified: for this reason the authors have adopted a joint APC-DEA technique that allows to strongly reduce the variability determined by the often arbitrary choice of inputs and outputs, bringing everything back to a few representative synthetic factors.

Keywords: Data Envelopment Analysis, Principal Component Analysis, benchmarking, container port, Mediterranean sea Bibliografia Beckman M., Mc Guire C. B., Winsten C. B. (1956), “Studies in the Economics of Transportation”, Yale University Press, New Haven CT. Bovy P.H. L., Stern E. (1990), “Route Choice: Way Finding In Transport Networks”, Kluwer Academic Pub, Dordreckt, The Netherlands. Branch, A. E. (1986), Elements of Port Operation and Management. London: Chapman and Hall Ltd. Brooks, M. R. (1984), “An alternative theoretical approach to the evaluation of liner shipping (Part 1. Situational factors)”, Maritime Policy and Management, 11, 35-43. Brooks, M. R. (1985), “An alternative theoretical approach to the evaluation of liner shipping (Part 2, Choice Criteria)”, Maritime Policy and Management, 12, 145-155. Chandrasekaran, R. Madhanagopal R.and Karthick K. (2013) A Stochastic Frontier Model on Investigating Efficiency of Life Insurance Companies in India International Journal of Mathematics Trends and Technology 4 (10) ; Cazzaniga Francesetti D. (2006) “I criteri di scelta dei porti internazionali e i porti italiani”, Università di Pisa Chen, Y., Du, J., David Sherman, H. & Zhu, J. (2010). DEA model with shared resources and efficiency decomposition. European Journal of Operational Research, 207(1): 339–349 De Langen P.W., 2007 “Port competition and selection in contestable hinterlands: the case of Austria”, European Journal of Transport and Infrastructure Research 7 (1), 1–14. D’Este, G. M. and S. Meyrick (1992), “Carrier selection in a RO/RO ferry trade Part 1. Decision factors and attitudes”, Maritime Policy and Management, 19 (2), 115-126. Ferraro A.M. (2006) “Indici di valutazione dell’efficienza di terminal marittimi per container”, Technical Report n. 5/2006, Laboratorio di logistica, Rende (CS), Italy. Foster T., (1978) “What’s important in a port”, Distribution Worldwide 78 (1978), 33–36. Gi-Tae Y., Roe M., Dinwoodie J. (2008), “Evaluating the competitiveness of container ports in Korea and China”, Transportation Research Part A: Policy and Practice, 2008, vol. 42, issue 6, 910-921. Holguin-Veras J., Jara-Diaz S. (1999), “Optimal pricing for priority service and space allocation in container ports”. Transportation Research B 33, 81-106. Kek Choo Chung (1993), “Port Performance Indicators”. Transportation, water and urban development department the world bank - Transport No. PS-6, December 1993. Lirn T.-C., Thanopoulou H.A. and Beresford A. K.C. (2003), “Transhipment Port Selection and Decision-Making Behaviour: Analysing the Taiwanese Case”. International Journal of Logistics: Research and Applications Vol. 6, No. 4, 2003. Malchow, M. and A. Kanafani (2001), “A Mathematical Analysis of the Significance of Port Charges.” Proceedings of the The International Association of Maritime Economists Annual Conference 2001, 923-936, Hong Kong. Martinez-Zarzoso I., Garcia-Menendez L. and Suarez-Burguet C. (2003), “Impact of transport costs on international trade: the case of Spanish ceramic exports”, Maritime Economics and Logistics 5 (2003), 179–198. Murphy, P., J. Daley and D. Dalenberg (1991), “Selecting links and nodes in international transportation: an Intermediary’s perspective”, Transportation Journal, 31 (2), 33-40. Poldaru and Roots (2014) A PCA–DEA approach to measure the quality of life in Estonian counties Socio-Economic Planning Sciences 48(1) Ogden, K.W. and A.L. Rattray (1982), “Analysis of freight mode choice”, Proceedings Of the 7th Australian Transport Research Forum, Hobart, Australia, 249-276. Saleh, F. and B. J. Lalonde (1972), “Industrial buying behaviour and the motor carrier Selection decision”, Journal of Purchasing, 8, 18-33. Sanchez R.J., J. Hoffman et al. (2003), “Port efficiency and international trade: port efficiency as a determinant of maritime transport costs”, Maritime Economics and Logistics 5 (2003), 199–218. Sayareh and Lewarn, 2006 Sayareh, J., Lewarn, B., 2006. Efficient supply chains through effective seaport organizations. In: Proceedings of the IAME Annual Conference 2006, 12–14 July, Melbourne, Australia. Slack, B. (1985), “Containerisation and inter-port competition”, Maritime Policy and Management, 12 (4), 293-304. Tavakoli M.M., Shirouyehzad H. and Dabestani R. (2013)Efficiency evaluation of a private bank's branches with service quality approach by data envelopment analysis International Journal of Services and Operations Management, 2013, vol. 16, issue 4, 427-442 Tongzon J. (2009) “Port choice and freight forwarders”, Transportation Research Part E 45 185-195. Tongzon J. (2002), “Port Choice Determinants in a Competitive Environment”, Proceedings of the IAME Panama 2002 Conference, 13 – 15 November 2002, Panama. Tongzon, J. and S. Ganesalingam (1994), “Evaluation of ASEAN port performance and efficiency”, Asian Economic Journal, 8 (3), 317-330. Ugboma C., Ugboma O. and I. Ogwude (2006) “An Analytic Hierarchy Process (AHP) approach to port selection decisions – empirical evidence from Nigerian port”s, Maritime Economics and Logistics 8 (2006), 251–266. UNCTAD (1992), “Strategic Planning for Port Authorities”. United Nations. UNCTAD (1996), “Port Organization and Management”. United Nations. Willingale, M. C. (1984), “Ship-operator port-routeing behaviour and the development process”, in B.S. Hoyle and D. Hilling (eds), Seaport Systems and Spatial Change, New York: John Wiley & Sons, 43-59. Wilmsmeier G., Hoffman J. and. Sanchez R. (2006), “The Impact of port characteristics on international maritime transport costs”. In: K. Cullinane and W. Talley, Editors, Port economics, Research in Transportation Economics vol. 16, Elsevier (2006). Wilson, F.R., Bisson, B.J., & K.B. Kobia (1986), “Factors that determine mode choice in the transportation of general freight”, Transportation Research Record, 1061, 26-31.

ABSTRACT. This paper deals with inventory sharing and routing in the context of decentralized supply chains. The supply chain considered in this paper consists of a single manufacturer distributing its products through a network of independents Points of Sale (POS). The problem is modeled as a 1-leader - n-followers Stackelberg game. A new mixed-integer bi-level program is developed, in which the manufacturer decides first on inventory levels and the distribution routes, considering each follower’s (POS) response function that minimizes the follower’s own cost. A trade-off solution to manage conflict of interests between the parties involved in the supply chain is also proposed. To solve the mixed-integer bi-level program an original hybrid Genetic Algorithm coupled with deep reinforcement learning is developed and used to solve a set of large-size instances. The gap analysis shows that the proposed hybrid algorithm performs quite well and that inventory sharing allows the network to improve its service level.

ABSTRACT. 1. Introduction

Lean/ Agile Management is a modern concept within the Supply Chain Management. The term refers to the ability to manage market changes, in a manner that responds to customer requirements but at an acceptable cost. Supply chain agility requires coordination and integration across individual firm functions. Main characteristics include Customer sensitivity, Virtual Integration, Process Integration, Network Integration, and Measurement (Christopher 2000, Van 2001). Lean/ Agile Supply Chain allows its links to easily adjust to the external conditions and respond to changes in the business environment. Leanness and agility of the supply chain correspond to its flexibility and resilience. In this paper we are considering a strategic problem of Fleet Composition for a Copy Paper Supply Chain and we are solving it along the concept of Lean/ Agile Management. Fleet Composition consists of defining the type and number of vehicles in a fleet that on one hand satisfies a complete fulfillment of incoming transportation orders (market demand meeting) and on the other hand allows avoiding high fixed costs associated with oversized and as a result underutilized fleet (Redmer et al 2012). According to this definition our major objective is to find most desirable fleet composition (size of a fleet and type/make of vehicles) for the delivery of copy paper blocks of different sizes within the two echelon supply chain (between paper manufacturing plant, regional warehouses and retailer’s stores/ shops).

2. Problem Statement

The considered decision problem has a strategic character and belongs to a classical category of “matching supply and demand” problems. We want to define the most suitable fleet operating between two echelons of the paper physical distribution/ supply system: Paper Manufacturing Plant and two different Warehouses W1 and W2; and further from warehouses W1, W2 to 25 clients (retailers) denominated by R1, R2… R25 (see figure 1). The assignment of retailers to specific warehouse is predefined. The analysis is carried out as a case study for a supply chain operating in Poland. The manufacturing plant, warehouses and retailers are located in the vicinity/ neighborhood of a large/ mid-sized Polish city of Poznan. The manufacturing plant owns a fleet of vehicles, and due to the increased demand, shall require investing in purchasing more vehicles. It is assumed in our analysis that some vehicles (trucks, vans) can be also sold and a portion of delivery services can be outsourced to a third party logistics (3PL) company.

This research group play the role of analysts whose task is to rank the variants from the most desirable to the least in order for the decision maker represented by the CEO of the company to choose from according to set of preferences concerning the stakeholders. The stakeholders are; the Company top Management who are mainly concerned about decreasing costs and magnifying profits, Employees represented by the plant workers, vehicle drivers and Warehouse employees who are concerned about their comfort, safety and additional wages emerged from deliveries, and finally the retailers.

The study assumes the deliveries of product in two different stages. First stage involves the deliveries of products from the manufacturing plant to two different warehouses. Second stage involves deliveries from warehouses directly to retailers located in Poznan.

3. Problem Formulation and Structuring

The decision problem is formulated as a Multiple Criteria Ranking Problem. The critical components of each MCDP are: definition of the consistent family of criteria, modeling of the DM’s and stakeholders preferences and computational experiments. In our analysis the proposed criteria is as follows: C1 - Monthly Cost [zl] -a minimized criterion denominating the overall logistics/ transportation costs associated with monthly fuel cost, salary and third party logistics costs. C2 - Investment Cost [zł] – a minimized criterion denominating the cost of purchasing new trucks and/or selling the ones that won’t be used in the future. C3 - Capacity Utilization [%] – a maximized criterion denominating how much of total trucks capacity is used during deliveries. C4 - Delivery Time [min] - a minimized criterion denominating the quality of services provided to the customers. Each of the vehicle models proposed by the project group achieves different results in the field of mobility and speed of delivery. C5 - Safety (1-10 Scale) - a maximized criterion associated with the average percentage of drivers' accidents per 1000 cars of a given model and the average number act of vandalism per year in the Wielkopolska region. C6 - Environmental Friendliness (Tones of CO2) - a minimized criterion associated with the amount of carbon dioxide in Tones emitted from the fleet during delivery. C7 - Employees Comfort (1-10 Scale) - a maximized criterion denominating comfort of Warehouse workers, Drivers and Plant workers associated with the number of routes as the more routes they have routes, the more setting ups for loading and unloading, and more paperwork. Comfort of Drivers has two sub criteria also, one related to the trucks flexibility and the other to the features of the trucks.

The variants are constructed intuitively with the application of expert knowledge. Each variant constitutes a certain fleet composition which either includes using the current owned fleet, purchasing/ renting vehicles as required or outsourcing a fleet. Each variant includes different types and number of trucks A, B, C…etc. with different characteristics. Examples of variants are as follows: • 1xA, 1xB, 1xC, 1xD, Using owned fleet only. • Selling truck A and buying additional trucks C and D. Fleet composition: 1xB, 2xC, 2xD • Selling the A and buying additional truck B and C. Fleet composition: 2xB, 2xC, 1xD The behavior of each fleet variant is simulated and certain evaluation parameters are generated. Parameters are aggregated into criteria. Examples of the parameters are the fuel cost, salaries, inventory, CO2 and percentages of car accidents. Based on the data the evaluation matrix is created.

4. Computational Experiments and Results The computational phase is performed with the application of multiple criteria ranking method called AHP (Analytic Hierarchy Process). The decision maker responsible for problem solving is a company top management whose objective is to create most desirable fleet composition. Authors of study are analysts who provide comprehensive analysis of the decision situation and provide a ranking of variants that helps the decision maker choose the one that fulfill the objectives for all parties as much as possible. Final result of alternatives ranking is created using AHP software. Each variant has different score that was evaluated using the objective function. The variant with highest value is considered the most desirable among other options. It is the alternative that fulfills the requirements of decision maker as well as the stakeholders the most. Based on the ranking, the top management of the company can decide which variant will be chosen.

ABSTRACT. The negative impacts of urban freight transport have fostered the design of new distribution systems for inner city deliveries. One of the proposed city logistics solutions is the establishment of two-echelon distribution systems, where freight originating from the periphery of the city is transferred at intermediate locations (satellites) from larger vehicles (urban freighters) to smaller more environmentally friendly vehicles (city freighters), deemed more suited to operate in the city centre.

In this paper, we address a variant of the two-echelon (capacitated) vehicle routing problem (2E-CVRP) arising in the context of city logistics encompassing characteristics such as a heterogeneous fleet of city freighters, time windows, vehicles synchronisation at satellites, direct deliveries, the possibility of freight transfers at customers, and multiple trips by city freighters.

We propose a heuristic solution method for this problem based on a Variable Neighbourhood Search (VNS) heuristic, as well as a set of auxiliary evaluation metrics designed to support the decision making process.

Preliminary results show the flexibility of the proposed heuristic to account for different problem characteristics, as well as the impacts of considering different best insertion criteria for the initial solution construction. Additionally, the proposed evaluation metrics have allowed for an improved assessment of the trade-offs between different city logistic solutions, and may be considered when developing decision support systems.

ABSTRACT. The aim of this paper is to estimate static OD matrices combining traditional data sources and Floating Car Data (FCD), testing several scenarios with different penetration rates and representativeness of probe vehicles. For each scenario, the accuracy of results is related to the cost of the data acquisition. To the best of authors’ knowledge, this paper represents the first attempt to define a relationship among FCD penetration rate and representativeness, traditional data sources, OD matrices accuracy and cost of the estimation, thus contributing to the literature on the subject and useful considerations of practical applications. The procedure was applied to two real cases for which ground-truth matrices were available, in order to calculate goodness of fit indicators for the estimated matrices. In this way, it was possible to define the optimal solution, which can maximize the reliability of results and minimize its cost. The procedure can be adapted to every real case, thereby providing sound basis for resource allocation in many applications.

ABSTRACT. This study explores the spatial relations between job and housing locations by mining trajectory data. The objective is to establish a data-driven method to recognize employment cores areas and identify their catchment residential areas. More specifically, mobility traces are firstly mined to detect the home and work places of each individual according to temporal patterns of the day and repetition patterns over distinct days. A spatial density distribution analysis is then conducted to identify the employment cores and sub-cores over the territory. The core-periphery patterns between the cores and other areas are further investigated by building a connection graph based on the home and work locations over zones. The graph-based clustering algorithm is employed to partition the graph so as to identify bonded zones as communities and interpret the catchment areas pertaining to different employment cores. A case study is applied with the field-collected Floating Car Data in the Paris Region to showcase the method applicability. Overall, this study offers a referential instance of analyzing urban spatial dynamics by digital traces at a large scale. The results of which may contribute to the planning guidance corresponding to the up-to-date changes.

ABSTRACT. Individuals' decision to include an activity in their daily agendas is based on their needs. They take decisions of activity participation, considering the activities and schedules of other household members. This results in activities that involve joint participation of household members. Limited modelling research is available on the notion of the generation process of joint activities of household members. Besides, the time span for most of the operational models is a single-day. The current paper's focus is on modelling joint-activity generation where the time span is based on multiple-days (one week). We used a latent class model (LCM) and estimated it using the data from the National Travel Survey (NTS (UK)) for the years 2013-14. The model predicts the probability of alternatives for an individual. Each alternative represents a series of binary outcome (i.e. 1 or 0) in relation to generation of a specific type of joint activity for every single day of the week. Four latent classes are identified based on model-fit parameters: frequent joint-activity performer, weekdays joint-activity performer, non-frequent joint-activity performer, and weekend joint-activity performer. The results show that both household and individual attributes impact the individual's weekly joint-activities' generation. The available time window to the individual after mandatory activities also affect which latent class the individual belongs. Future research will focus on integrating the model in some functional activity-based model.

ABSTRACT. The present work describes a driving simulator experiment aimed at analysing how different feedback modalities can affect vehicle lateral control over time. Participants performed the same track 4 times: the first time (baseline) no feedback was delivered, whereas the next three times either an acoustic or multimodal feedback was presented in case of error. About 25% of the participants in the initial sample were tested again after a month without any feedback. The results showed that the positive effect of the feedback was maintained over time. Thanks to the precision teaching technique, drivers improved the investigated driving parameters while maintaining a correct position inside the lane.

ABSTRACT. The main field of investigation of the paper is the dynamic traffic assignment. In particular, in this context different models can be applied concerning the dynamic network loading itself and in particular the (within-day dynamic) traffic flow propagation on links. Due to the advances in technological application, more recently different traffic flow models and new enhanced strategies have been proposed and applied in the era of the Intelligent Transportation Systems and in particular in presence of cooperative and connected communications (V2V, Vehicle to Vehicle communication; and V2I, vehicle to infrastructure communication). In more detail, the paper aims at investigating the application of a new proposed traffic flow model that has been formalised to support the application of traffic control strategies (also including the procedure of speed optimisation) in presence of connected and automated vehicles (CAV). In terms of traffic flow modelling, the literature identifies three main modelling approaches: macroscopic, mesoscopic and microscopic modelling. A further approach has been recently investigated, called hybrid traffic flow modelling, which is also suitable for applications at different scales (multi-scale). This paper discusses the results achieved through the application of a new hybrid traffic flow model based on the combination of two sub-models: a space-time discrete model based on aggregate macroscopic variables representation (the Cell Transmission Model), and a space-time discrete model based on a disaggregate model (the Cellular Automata model). In particular, the main focus of the paper is on the application of the model at a large scale network considering the implicit path choice based on the Logit and C-logit behaviour models. The achieved results are also compared with respect to the application of an explicit path enumeration. The analyses are carried out in terms of travel time spent, network total delay, queue lengths and degree of saturation. Results highlight the suitability of the proposed approach and the implicit path choice modelling slightly outperforms the case of explicit path choice enumeration in terms of performance indicators. Peer-review under responsibility of the scientific committee of the 24th EURO Working Group on Transportation Meeting.The main field of investigation of the paper is the dynamic traffic assignment. In particular, in this context different models can be applied concerning the dynamic network loading itself and in particular the (within-day dynamic) traffic flow propagation on links. Due to the advances in technological application, more recently different traffic flow models and new enhanced strategies have been proposed and applied in the era of the Intelligent Transportation Systems and in particular in presence of cooperative and connected communications (V2V, Vehicle to Vehicle communication; and V2I, vehicle to infrastructure communication). In more detail, the paper aims at investigating the application of a new proposed traffic flow model that has been formalised to support the application of traffic control strategies (also including the procedure of speed optimisation) in presence of connected and automated vehicles (CAV). In terms of traffic flow modelling, the literature identifies three main modelling approaches: macroscopic, mesoscopic and microscopic modelling. A further approach has been recently investigated, called hybrid traffic flow modelling, which is also suitable for applications at different scales (multi-scale). This paper discusses the results achieved through the application of a new hybrid traffic flow model based on the combination of two sub-models: a space-time discrete model based on aggregate macroscopic variables representation (the Cell Transmission Model), and a space-time discrete model based on a disaggregate model (the Cellular Automata model). In particular, the main focus of the paper is on the application of the model at a large scale network considering the implicit path choice based on the Logit and C-logit behaviour models. The achieved results are also compared with respect to the application of an explicit path enumeration. The analyses are carried out in terms of travel time spent, network total delay, queue lengths and degree of saturation. Results highlight the suitability of the proposed approach and the implicit path choice modelling slightly outperforms the case of explicit path choice enumeration in terms of performance indicators.

ABSTRACT. The aim of this paper is to merge both macroscopic models: bi-dimensional model with the GSOM (Generic Second Order) model in order to overcome the missing data for large networks. The exchange between artery of the GSOM model and cells edges of the bi-dimensional are assumed through the supply and demand concept. Indeed variable disaggregated from bi-dimensional to GSOM and aggregated from GSOM to bi-dimensional. Simulation is based on Godunov scheme and finite volume method.

ABSTRACT. The article develops methods to improve the accuracy of modeling and predicting the traffic capacity of intersections based on the fuzzy factors of intensity of pedestrian flow and flow discontinuity. Initial data was collected and analyzed using convolutional neural networks (YOLOv3). The article justifies the need for clustering signal-controlled junctions of the transport network to form a limited number of generic management algorithms for them. We analyzed the differences between clusters by the average values of independent factors. Regression modeling was completed for each cluster. Statistical analysis showed a significant improvement in the quality of the models within each cluster compared to the complete model of transport junctions, exceeding the maximum permissible quality thresholds. We built a model using fuzzy logic methods which reflect the distribution field of the influence of the most important factors.

ABSTRACT. The GSOM model (Generic second order model) is a macroscopic traffic model designed to take into account the dynamics of vehicle and driver attributes. It has recently been extended to the multimodal case. One prominent feature of the resulting multimodal GSOM model is that it distinguishes two flows: vehicular (car or public transportation) and passenger. The flow of pasengers is subordinated to the flow of vehicles. On the other hand, traffic on regional sized networks cannot be modelled with any great precision over very dense surface subnetworks of the regional network. Thus various approximate methods are being developped to address this difficulty, notably the MFD (macroscopic fundamental diagram) approach and bidimensional methods. The paper concentrates on the latter and aims first to develop a multimodal bidimensional model. In a second step the paper investigates how to interface the multimodal GSOM and the bidimensional models. At a regional scale the resulting model describes the structuring network (main roads, motorways, train tram metro lines...) following the GSOM methodology, and describes the flow on dense networks of small to medium capacity streets and lines as a bidimensional flow over a bidimensional medium.