Transportation I
Transportation I
Contributed Session
Time Slot: Friday Morning
Room: AUD_A
Chair: Martina Fischetti
Optimizing a real-time carpooling service
Time: 9:00
Maurizio Bruglieri (Politecnico di Milano), Peruzzini Roberto, Pisacane Ornella
According to a recent survey, in Europe, the average occupancy rate of most frequent car trips is 1.7 persons/car (including the driver). Moreover, this value has decreased over the last decades. This decreasing trend implies an increase in the number of private cars on the road and then, in the traffic congestion. In addition, more cars on the road also means an increase in the harmful emissions. A possible solution to this problem, more flexible than public transport, is given by carpooling. Carpooling (also known as ridesharing) refers to a mode of transportation in which travelers share a vehicle for a trip with others that have similar itineraries and time schedules, and split travel costs (e.g., those due the fuel consumption). Carpooling is a service that can combine the flexibility and speed of private cars or taxi services with the reduced cost of fixed-line systems. Besides saving travel costs, carpooling also allows reducing the number of circulating cars and thus in mitigating traffic congestion and reducing air pollution.By effectively using new communication technologies, such as smartphone and global positioning system (GPS), there are some attempts to enable real-time ridesharing systems, i.e., a system which supports an automatic ride-matching process among participants on very short notice or even en-route.In this work, we want to investigate the benefit of using autonomous vehicles in a real-time ridesharing service. Indeed, if on one hand, such a service has the clear advantage of not requiring any participant that acts as a driver, on the other hand, it requires that the routes of each vehicle are planned by a central system to optimize simultaneously different objectives such as the total number of passengers served and the total traveled distance. Then, this central system acts as a dispatcher of the requests and provides the users with their assignment to a vehicle that may already have passengers on board. In the assignment process, such a dispatcher has to take into account the level of service and then, to avoid situations in which passengers already on board make a long detour in order to allow the vehicle also serving other users. The resulting decision problem is a challenging Vehicle Routing Problem. To solve it, we propose a mathematical programming based approach in which, in each iteration, a lexicografic Mixed Integer Linear Programming model is solved in order to find partial solutions. Finally, in order to address real-life instances in reasonable computational time, we also develop a metaheuristic in which we combine moves taken from the literature with those ad hoc designed for the problem.Preliminary numerical results on instances derived from a carpooling service in Melbourne show the effectiveness of the proposed metaheuristic compared to the mathematical programming based approach.
Bundles generation and pricing in crowdshipping
Time: 9:20
Giusy Macrina (University of Calabria), Archetti Claudia, Guerriero Francesca
In this work we investigate the crowdshipping paradigm, an innovative framework of delivering which is gaining more and more success in the last mile delivery context. The main idea of crowdshipping is to exploit the capacity of ordinary people, namely occasional drivers (ODs), who offer their own vehicles and free time for performing shipments, allowing the delivery companies to provide a more flexible, faster and cheaper delivery service. Deciding which customer requests should be served by the internal own fleet of vehicles and which, instead, should be assigned to the ODs, is a challenging issue, as well as determining the compensation that should be offered to the ODs to minimize the expected total distribution cost. Hence, we develop a bundle-based strategy to decide how many, and which parcels should be assigned to the ODs. We then focus on matching and pricing problems, with the purpose of maximizing the numbers of assigned parcels and minimizing the costs. We propose a pricing scheme to determine the compensation for the ODs. We also assume that each OD has a willingness to serve function, that is a compensation for which the OD will accept to deliver a bundle. Hence, we simulate two auction strategies: a static and a dynamic one. We carry out an exhaustive computational study to compare the pricing strategies and the two auctions, outlining several managerial insights.
A pick-up rule for sharing transportation costs in itinerant events
Time: 9:40
Amparo Mármol (Universidad de Sevilla), Diego Borrero, Miguel Angel Hinojosa
Fixed route problems model situations in which a company must visit several customers located in different places, traveling along a fixed route that starts and ends at the same point. When the objective of the fixed route is the transportation of material and the installation of equipment for the celebration of itinerant events, there are certain differentiated characteristics that must be taken into account when allocating the corresponding transportation costs. In this paper, we firstly propose some cost-sharing rules for the fixed route problems that represent these situations. Subsequently, we introduce a TU game representing the costs that the different coalitions should bear in these situations, the pick-up game. We prove that pick-up games are convex, thus the existence of stable allocations is assured. Moreover, one the cost-sharing rules previously proposed is highlighted in this context the pick-up rule, since we prove that it coincides with the Shapley value, the nucleolus and the tau-value of the pick-up game.
Decision support system in urban traffic management
Time: 10:00
Siniša Vilke (University of Rijeka, Faculty of Maritime Studies), Jasmin Ćelić, Frane Tadić, Borna Debelić
Traffic congestion is an unavoidable problem in large cities. Therefore, traffic management centers must respond quickly to traffic disruptions and apply the most appropriate measures to solve problems in the traffic network. The causes leading to traffic congestion are usually related to regular or extraordinary traffic events and traffic condition.Due to a regular traffic event such as a traffic collision on the main traffic corridor, illegal parking and emergency interventions, actions such as detection, closure and speed limits are mostly used. During an extraordinary traffic event such as manifestations or roadwork, applied actions include the partial or complete closure of the road, change of route, change of travel methods, etc.Solving this complex task requires not only a quick response, but also a lot of expertise given the large number of possible actions that need to be considered during the decision-making process. Therefore, simulation models or decision support tools are increasingly used. In this project, the simulations are intended to demonstrate the possibilities of using simulation tools in the decision-making and traffic management process as innovative methods within intelligent solutions for traffic management in cities.The purpose of prototyping is to test innovative scenarios and standard operating procedures that will be used to support decision making. The prototype created in this way can be further developed through validation and calibration procedures to obtain a real traffic situation in an area, based on which different solutions affecting traffic flow can then be tested.As part of the development activities of the CEKOM Connected Traffic scientific research project, an advanced information and communication system is being developed to support decision-making and urban traffic management. The improvement of traffic management within this project will also have a positive impact on the development of mobility and the increase of traffic safety in the city of Rijeka. The innovation of this platform is reflected in its approach to connectivity and Big Data analytics, as well as its real-time traffic management. In this paper, the authors focus on testing innovative scenarios and standard operating procedures for the purpose of decision support to demonstrate the effectiveness of traffic management in urban areas by applying innovative solutions (modeling and simulation). The development of the simulation model in terms of functionality and sustainability covers the narrower and wider urban area of Rijeka and the application of a multi-layer modeling methodology that integrates macro and micro level traffic simulations. Reliable and accurate measurement of vehicle load on individual roads was used to create the simulation. Since the simulation will be validated with some new traffic routes, new realistically calculated traffic load data is expected as one of the outputs of the project.By applying intelligent traffic solutions and information and communication technology, it is possible to integrate large amounts of processed data through the platform, thus achieving the validation of the prototype and the generation of traffic conditions, which increases the quality of traffic management.* The paper is the result of research activities of the scientific project Connected Traffic implemented within CEKOM for smart cities (CEKOM – Center of Competence for smart cities, the city of Rijeka), funded by the EU ESIF fund, started in March 2020. and ends in March 2023.
How the European Union uses Operations Research for transport analyses
Time: 10:20
Martina Fischetti (Joint Research Center), Aycart Javier, Duma Davide, Gualandi Stefano, Ibáñez Juan Nicolás, Tomasi Claudio
One of the main European Union (EU) policy priorities under the European Green Deal is to achieve climate neutrality by 2050. Transport is a key player in that task, as it is a major consumer of energy, and extensive contributor to greenhouse gas emissions. In this context, the EU mandate includes the monitoring of the performances of both public transport and road transport across the EU member states. This is key to be able to inform the relevant policy decisions on the topic. One of elements of analysis that the EU employs is the use of comprehensive data to compute performance and accessibility-to-opportunities metrics associated with different types of transport. Underneath these measures lies a number of operations research (OR) challenges. The road transport analysis, for example, is a routing problem defined on (very) large networks. Moreover, if traffic information is considered, the problem turns into a time-dependent routing problem.In this presentation we will give an overview of some of the OR transport-related challenges faced by the EU and some of possible solution methods proposed by the Joint Research Center (JRC) of the European Commission. The work presented is developed in collaboration with the Department of Mathematics of the University of Pavia.