we are happy to invite you to the seminar of Dr. Moises Rodriguez Madrena (Universidad de Sevilla - Facultad de Matemáticas) that will take place on Thursday, January 27, 15:00-16:00.

The seminar will be held in-person room 18 (Aula 18) of the Department of Business Studies of Roma Tre University (Via Silvio D'Amico, 77 00145 - Roma, Italy), and online via Teams.

Date and time: January 27 (Thursday), 15:00 pm

Speaker: Dr. Moises Rodriguez Madrena (Universidad de Sevilla - Facultad de Matemáticas)

Title: Location Problems: New Advances and Applications

Abstract:  Location problems are among the most important problems in the area of Operation Research. In general, a location problem consists of determining the position of one or more facilities to satisfy the demand of a set of demand points, such that some criteria (represented by objective functions) are optimized (minimize or maximize). In this seminar, we present two new location problems that arise from two different applied scopes.

(1) In the literature of transportation research field, it is frequent to address routing or distribution problems where the movement between points is given by the combination of different transportation modes. For instance, movements within or outside an urban area could be different: within the city boundary the movement is slow due to the layout of the streets and traffic, while outside this boundary in the rural area the movement is fast. This situation can be modeled as regions endowed with different norms, where distances within the regions are the ones induced by the corresponding norm. In this framework, we address the single-facility Weber location problem considering different Lp-norms at different polyhedral regions.

(2) Given a set of tradable assets, the classical portfolio selection problem consists in determining the amount of capital to be invested in each asset to build the most performing portfolio. A relatively recent promising strand of research is to exploit clustering information of an asset network to develop new portfolio optimization paradigms. Following this idea, we endow the asset network with a metric based on correlation coefficients between assets’ returns and show how classical location problems on networks can be used for clustering assets. More precisely, in the bi-objective (gain-risk) portfolio selection model, we add a third criterion which is represented by the objective function of the standard p-median location problem. In this way, we are able to evaluate the effect of different degrees of clustering on the selected portfolios.

We tackle the above problems using Mathematical Programming tools (MISOCP and MILP formulations for the first and the second problems, respectively). The usefulness of our approach is validated by reporting some computational experiments.

Venue: Room 18 (Aula 18) - Department of Business Studies - Roma Tre University -  Via Silvio D'Amico, 77 00145 - Roma, Italy.

Teams link: virtual room