Abstract

This project aims at advancing in the development of sophisticated models, optimization and simulation techniques for solving and evaluating the performance of different transport problems, with the common feature of a notorious dynamic behavior, observed in all of them. The major goals of the project are to systematize the methodologies to model various dynamic transport problems, to formulate sophisticated methods that incorporates the proper complexity of each situation, to work with real data (offline and online) in the context of the proposed models, to develop efficient solution algorithms that should be consistent with the theory, and to constructs detailed simulation platforms of real systems to evaluate and validate the different dynamic optimization methods proposed. In this project, the research is focused on dynamic vehicle routing problems, including pick-up and delivery for passengers and a dispatch of technicians for a maintenance-repair service, both problems formulated under a dynamic framework in the sense of making routing decisions in real-time according to unknown demand coming up during the working day. In addition, we study a joint location-routing problem for a distribution of tests in schools. The second family of problems we deal with are oriented to the operation of traditional (fixed route) public transport systems. In this case we orient the research on planning on optimization of operational schemes. The third line of research deals with network design problems mostly oriented to the decision of implementing transfer points for dynamic systems. Thus, we show an integration of fixed route public transport services with a pickup and delivery service, the integration between lines for a fixed corridor, some optimization schemes for solving the pickup and delivery problem with transfers, and also some heuristic rules to improve the operation at transfer points. The fourth topic is oriented to a dynamic multi-user and multi-scale traffic equilibrium model for urban networks, where the dynamic component appears in the fashion to model equilibrium and users’ interactions. Finally, we describe the potential simulation platforms to be constructed in order to test and validate the operational schemes and optimization models described.