Lines of research
- We are interested in developing new multidimensional data structures, which have applications in many fields as geographical information systems, proximity search, computer graphics, and many other. We also are interested the design and implementation of algorithms to process very large amounts of incoming information, for instance from the e-transactions or genetic data bases.
- C. Martínez, A. Duch, M. Serna, X. Messeguer
- A large amount of data can be represented as large graphs or digraphs, where the vertices represent the data and the edges some type or relation (proximity, dependency,
). One of the big issues in data science is the problem of reporting, i.e., inferring information from the massive data and formulate or formulating predictions. A particular line we
like to explore is the analysis of data coming from the fingerprint left by the execution of processes. One of the big challenges for manipulating that kind of data is their size, for instance any
on-line shop could store 106 transactions per hour. We want to develop techniques and
to analyse efficiently the data produced by on-line transactions. We also are interested in
an experimental framework for benchmarking the performance of graph algorithms
using a concurrent programming pattern called Dynamic Pipelines, on large big-data
- J. Carmona, E. Pasarella, J. Cortadella, J. Díaz
- We are interested in the design of efficient algorithms for comparing and analysing genomic data, as for instance, the alignment of genomes and the search of syntenic regions, as well as their statistical significance.
are also involved in methods for the taxonomic classification of metagenomic sequences and the monitoring of microbiomes, and we have
the TANGO software tool for the taxonomic classification of next generation sequences, which has been added to the BioMaS pipeline for the metagenomic analysis of amplicon
and to the MetaShot pipeline for the metagenomic analysis of shotgun sequences.
- G. Valiente, X. Messeguer, J. A. Subirana
satisfaction is the nutshell of many hard computational problems, such as scheduling problems, layout of large graphs, resources assignment problems, database query answering, etc. In all those problems, the
is to find a feasible solution that is constrained by a set of restrictions, or to find a close solution that satisfies as many restrictions as possible or is close to the optimal solution with
to a similar optimization goal. Our aim is to continue to work in two main lines: Closing the open gaps in the algorithmic complexity of the mildly tractable cases of the generic CSP, and nailing
the structural characteristics that make the intractable cases so difficult to solve. An important component for both lines is the formulation of the problem as a geometric
problem with semi-algebraic constraints.
- A. Atserias, J. Cortadella, J. Díaz, M. Garlik
- In this area we face various challenges in the design automation of nanoelectronic integrated circuits that require efficient algorithmic solutions. The
is focused on three topics: synthesis of layout for memory-dominated systems, architectural exploration of elastic systems and synthesis of asynchronous circuits.
first two topics are strongly related to the physical design (placement and routing) of large electronic systems.
third topic implies a consolidation of the research carried out for the logic synthesis of clockless control systems. A common aspect of all the aforementioned problems is the treatment of spaces of solutions that have a combinatorial explosion.
will often resort to techniques related to constraint programming using satisfiability and linear programming solvers.
for efficiently navigating along the solution space will also be used when required by the size of the solution space
- J. Cortadella, J. Petit, M.J. Blesa , J. Díaz, L. Machado
- We want to study logical models to help reasoning about the behaviour of big networks of systems, which could yield more efficient algorithm designs. We are referring to the kind of systems such as the internet of things. A system can be modelled as
attribute graph and the computational evolution of the network of systems as a sequence of graph transformations. We aim to a definition of a logic that allows us to reason
the properties of the graphs. This will allow us to define a set of tools to verify the correctness of the algorithms defined on our model.
- F. Orejas, E. Pino, E. Pasarella
- We consider large graphs like complex networks or large ad-hoc communications networks. We are interested in stochastic models for spread of disseminations on
kinds of graphs. For instance to model the behaviour of spread of infections and healing, on some specific networks as well as design efficient
and heuristics for particular problems, as the transmission of information. We also analise problems on static and dynamic networks from the perspective of game theory.
focus on three directions. Non-cooperative models for network formation, in particular the analysis of properties of the equilibria networks and of the dynamics creating such networks.
processes, as source for models of cooperation and decision making. We aim to extend the study of influence spread to other models of cooperation based on language acquisition
other social porcesses. Those models will be analysed as cooperative combinatorial optimization games as well as evolutionary non-cooperative games.
analysis in the angel-daemon framework, we will continue the study of uncertainty in the execution of orchestrations of web services. We also plan to extend the
analysis of uncertainty in weighted voting games to more complex decision systems.
- M. Serna, C. Àlvarez, M. J. Blesa, A. Duch, J. Gabarró, A. Messegué, J. Sierra, J. Díaz
ALBCOM Research Group
© Universitat Politècnica de Catalunya, 2017