Graduate Seminar 2018/2019

The Graduate Seminar ("Seminario Dottorato" in Italian) started in 2006. It runs about twice per month, usually on Wednesday afternoon, except in Summer. Seminars are usually given by PhD students and PostDocs of the Department of Mathematics, but occasionally also by Senior Researchers. It is assumed that each Student of the Doctoral School will give a talk in the Seminar during his/her doctoral studies.

The Graduate Seminar is a double-aimed activity. On the one hand, speakers have the opportunity to think how to communicate their researches to a public of mathematically well-educated but not specialist people, by preserving both understandability and the flavour of a research report. On the other hand, people in the audience enjoy a rare opportunity to get an accessible-but-precise idea of what's going on in areas of mathematics that they might not know very well.

All speakers are required to prepare a short report on the the topic of their talk, which are collected in a booklet at the end of the  year.

The Graduate Student is organized by Corrado Marastoni and Tiziano Vargiolu.

 

List of 2018-2019 Seminars
(Click on title for abstract)

 

Abstracts of 2018-2019 Seminars

Nicola Gastaldon, Exact and Meta-Heuristic Approach for Vehicle Routing Problems

Abstract. The Vehicle Routing Problem (VRP) includes a wide class of problems studied in Operations Research and relevant from both theoretical and practical perspectives. In its basic formulation, the problem is to find a set of routes for a given fleet of vehicles through a set of locations, so that each location is visited by exactly one vehicle and the total travel cost is minimized. Such problem is often enriched with many attributes rising from real-world applications, such as capacity constraints, pickup and delivery operations, time windows, etc. VRP belongs to the class of combinatorial optimization problems, and it is very hard to solve efficiently and researchers have developed many exact and (meta-)heuristic algorithms. The former takes advantage of the structure of the mathematical model to obtain a speedup through decomposition methods. The latter exploits heuristic techniques to obtain solutions that trade off quality and computational burden, such as evolutionary algorithms and neighborhood search routines. In our research, we consider the VRP arising at Trans-Cel, a freight transportation company based in Padova. We devised a Tabu Search heuristic implementing different neighborhood search policies, and now embedded in the tool supporting the operation manager at Trans-Cel. The algorithm runs in an acceptable amount of time both in static and dynamic settings, and the quality of the solutions is assessed through comparison with results obtained by a Column Generation algorithm that solves a mathematical programming formulation of the problem. Current research aims at developing data-driven techniques that exploit the information available from the company's repositories to support stochastic transportation demand arising in real time.

Top