» Investigación » SISIUS - Investigación en la USE

Responsable: Juan Antonio Mesa López-Colmenar
Tipo de Proyecto/Ayuda: Plan Nacional del 2009
Referencia: MTM2009-14243
Fecha de Inicio: 01-01-2010
Fecha de Finalización: 30-06-2013

Empresa/Organismo financiador/es:

• Ministerio de Ciencia e Innovación

Equipo:

• Investigadores:
  • Rosario Arriola Hernández
  • Eva Barrena Algara (alta: 01/03/2011)
  • María Teresa Cáceres Sansaloni
  • Gilbert Laporte
  • María Cruz López de los Mozos Martín
  • Antonio José Lozano Palacio
  • Francisco Alonso Ortega Riejos
  • Miguel Ángel Pozo Montaño
  • Alicia de Los Santos Pineda
  • Anita Schöbel

Resumen del proyecto:

Location analysis and network design are two strongly related fields that have evolved in parallel and are both well studied in the literature. In these fields, one aims to model real situations that often involve high budgets, whose functioning is planned to a long term horizon. Nevertheless, sometimes the available input data are not completely reliable, and consequently long term results can be wrong.

According to IEEE Standard Computer Dictionary, the robustness of a system can be defined as the degree to which a system or a component can function correctly in the presence of invalid inputs or stressful environmental conditions. If the underlying space is a graph or a tree, algorithms and results about their complexity are known for the classic objectives (median, centre, cent-dian regret). But for other objectives such as equity, there are almost no results, when the context is discrete or continuous and when one wants to locate non-punctual structures. In this project we intend to enhance the knowledge related to those contexts. On the other hand, in the network design field, several efforts towards the evaluation of the robustness of a network, even towards its optimization, have been made. In contexts more complex than the topological one (when there is a metric involved and the network is used for flow transportation and paths and lines must be designed, with their frequencies, in order to transport this flow) the robustness concept is not well established.

A second goal of this project is to analyze possible indexes of robustness in these cases, which is related to the first one. In the network design field, some research has been devoted to determining robust networks (especially in the traffic network field with equilibrium models) by using Stochastic Programming and Robust Optimization. Other strategies such as creating redundancies, minimizing interrelations and recovery robustness will be dealt with in this project with the aim of designing robust complex networks.

As a final goal we want to integrate models of location and network design in a robust way. In particular, we will consider railway networks, due to its special complexity both from the modelling point of view and from the computational size of the optimization problems involved. Exact methods will be used to solve small instances, but for larger instances heuristic, approximate and metaheuristic algorithms will be needed. Those non-exact algorithms will be tested and compared by means of the small instances. This project is an extension of the MTM2006-15054, more precisely of its line 3: equity and robustness objectives. It is also related with the VI Framework Programme EC, under contract no. FP6−021235–2: IST-021325, ARRIVAL (Algorithms for Robust and on-line Railway Optimization: Improving the Validity and reliAbility of Large scale systems).