Roger Xavier Lera-Leri

Roger Xavier Lera-Leri

Auto-generated profile, based on publicly available information

Roger Xavier Lera-Leri

Artificial Intelligence Research Institute (IIIA-CSIC)

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 February 2024 in Knowledge-Based Systems
Verified
Journal Article
Open access
RLRoger X. Lera-Leri
Roger X. Lera-Leri
ELEnrico Liscio
Enrico Liscio
FBFilippo Bistaffa
Filippo Bistaffa
CJCatholijn M. Jonker
Catholijn M. Jonker
MLMaite Lopez-Sanchez
Maite Lopez-Sanchez
Francisco Salas-MolinaFrancisco Salas-Molina
Francisco Salas-Molina

2 June 2025 in Annals of Operations Research
Verified
Journal Article
Open access
PDF file
RLRoger X. Lera-Leri
Roger X. Lera-Leri
FBFilippo Bistaffa
Filippo Bistaffa
TTTomas Trescak
Tomas Trescak
JRJuan A. Rodríguez-Aguilar
Juan A. Rodríguez-Aguilar
Sensibly planning the subjects to study during a university degree is one of the most crucial tasks that impact the future professional life of a student. Nonetheless, to the best of our knowledge, no automated solution is available for students who want to plan their desired degree path and maximize the skills required by desired or target job(s). In this paper, we consider the Degree Planning Problem (DPP), which aims at computing degree plans composed of university subjects for students during the completion of an undergraduate degree. Specifically, we aim to obtain the best set of skills matching the requirements of students’ preferred job(s). To achieve this objective, we propose a flexible and scalable approach that solves the DPP in real-time by means of a non-trivial formalization as an optimization problem that can be solved with standard solvers. Finally, we employ real data from our University’s Bachelor in Information and Communications Technology to show, through several use cases, that our approach can be a valuable decision-support tool for students and curriculum designers.

16 October 2024
Verified
Book Chapter
Open access
PDF file
RLRoger Xavier Lera-Leri
Roger Xavier Lera-Leri
FBFilippo Bistaffa
Filippo Bistaffa
Research Project Planning (RPP) is a central task routinely tackled by research institutions, which aims at planning the dedication of a set of researchers involved in a set of research projects, with the goal of meeting the preferences of each researcher as much as possible while efficiently utilizing the available budget. Despite its importance, in real-world institutions, RPP is solved manually due to the lack of automated solutions. To overcome this limitation, we put forward a flexible and scalable approach to provide robust project plans to users (e.g., the administrative staff of a research institution) by modeling the RPP as a constrained norm approximation problem, hence enabling the use of modern off-the-shelf optimization solvers. Results on real-world data provided by our research institution show that our approach can compute optimal plans that reflect the preferences of users (i.e., meeting the preferred dedication of researchers and efficiently spending the available budget) in a matter of seconds. Furthermore, we show that our approach can compute plans for thousands of projects and researchers within minutes, hence being able to solve RPP problem instances much larger than the ones typically encountered in an average-sized institution.


17 July 2025 in arXiv
Verified
Preprint
Open access
PDF file
RLRoger Xavier Lera-Leri
Roger Xavier Lera-Leri
FBFilippo Bistaffa
Filippo Bistaffa
AGAthina Georgara
Athina Georgara
JRJuan Antonio Rodriguez-Aguilar
Juan Antonio Rodriguez-Aguilar
Following the recent push for trustworthy AI, there has been an increasing interest in developing contrastive explanation techniques for optimisation, especially concerning the solution of specific decision-making processes formalised as MILPs. Along these lines, we propose X-MILP, a domain-agnostic approach for building contrastive explanations for MILPs based on constraint reasoning techniques. First, we show how to encode the queries a user makes about the solution of an MILP problem as additional constraints. Then, we determine the reasons that constitute the answer to the user's query by computing the Irreducible Infeasible Subsystem (IIS) of the newly obtained set of constraints. Finally, we represent our explanation as a "graph of reasons" constructed from the IIS, which helps the user understand the structure among the reasons that answer their query. We test our method on instances of well-known optimisation problems to evaluate the empirical hardness of computing explanations.