Research activities
CNR's research is organized in units (moduli) which are all linked to one of the eleven Departments of the CNR. Below we list the units of the Institute, indicating the year(s) of activity, the unit's coordinator, and the Department to which the unit is linked.
DYNAMICS OF FLUID AND BIOLOGICAL COMPLEX SYSTEMS (2005, 2006, 2007, coordinator Sauro Succi), Department of Materials and Devices.
Fluid-dynamics represents one of the leading-edge disciplines in modern
applied science. Traditionally devoted to the area of mechanical engineering, such as automotive and aeronautic sectors, under the relentless progress of micro and nano-technology, the science of fluids is nowadays strongly driven towards its emerging interfaces with material science and biology. This implies a substantial extension of its conceptual foundations and practical goals, both on conceptual and methodological
grounds, as well as on the application side. The activity of the present 'Commessa' is intended to cover all of these three complementary aspects, with projects encompassing a broad spectrum of activities, ranging from macroscopic fluid turbulence, all the way down to micro/nanofluidics, as well as quantum and relativistic fluids. It also includes a substantial activity on the development of advanced and innovative numerical methods for fluid systems at large.
MODELS, MATHEMATICAL METHODS AND NUMERICAL SIMULATION FOR THE DEVELOPMENT OF NEW MATHERIALS: RESEARCH AND TRAINING (2007,2008, coordinator Vanda Valente), Department of Materials and Devices.
The research is devoted to the development of mathematical methodologies and numerical tools for a deeper comprehension of a class of new materials as ferromagnets , liquid crystals, superconductors, ceramics, thin films, whose physical and chemical properties are widely used in the design and production of control devices and biomedical prostheses.
The research interests lie in the areas of control theory, partial differential equations, calculus of variations and scientific computing. The implementation of advanced computational methods, which requires also computer science know-how, is essential for the numerical simulation of the connected phenomena: controllability, phase changes, development and propagation of singularities, fractures and delamination in composite structures. The proposal is also supported by experiences developed in the framework of Research and Training European Projects.
QUANTITATIVE METHODS IN MANUFACTURING SYSTEMS (2007, 2008, coordinator Benedetto Piccoli), Department of Production Systems.
The goal of this research activity is to develop quantitative methods in manufacturing related to:
- control, robotics, scheduling and simulation which play a central role in operations management for manufacturing;
- software tool for biomedical devices with special focus on NMR;
- methods for the analysis and forecasting in complex financial and economic markets, which are sought e.g. to plan product demand and launch of new products;
- transportation networks, which are important in the context of distribution logistics and traffic;
- quality in manufacturing systems, another key factor in production to implement low cost policies;
- high performance computing, in distributed scenarios.
MATHEMATICAL MODELLING, SIMULATION AND OPTIMIZATION IN SUPPLY CHAINS (2005, 2006, coordinator Massimiliano Caramia), Department of Production Systems.
This research activity has the objective of putting together different methodological knowledge in practical contexts related to Project 1 of the Production System Department. In particular, the goal is twofold: on the one hand, the emphasis is on interdisciplinary to problem solving; and, on the other hand, the goal is to promote young researchers with such a vision. Mathematical modelling, simulation and optimization are the keywords of the research produced in this area; supply chains represent the applied context.
METHODS OF SCIENTIFIC COMPUTING AND DEVELOPMENT OF ALGORITHMS AND SOFTWARE (2005, 2006, coordinator Nicola Mastronardi), Department ICT.
The aim is the study of mathematical models and fast and reliable numerical algorithms for large scale problems arising in different fields, such as biomedicine, signal and image processing, remote sensing, fluid dynamics. The involved researchers have skills in different domains of Applied Mathematics, such as Statistics, Numerical Modeling, Numerical Analysis.
DEVELOPMENT OF MATHEMATICAL AND STATISTICAL METHODS AND OF RELATED SOFTWARE ORIENTED TO GRID COMPUTING IN A WORKFLOW MANAGEMENT ENVIRONMENT FOR REMOTE SENSING AND GENETICS (2006, resp. Umberto Amato), Department ICT.
Aim of the research is to cover the sequel of the research that goes from the development of mathematical and statistical methods for solving complex problems to the implementation of the related software in a cluster and grid environment, until the development of interfaces that make easy to use software in distributed computing environments. This purposes well fit the paradigma of e-Science that aims at improving the global cooperation in science by sharing computer equipments, data and software geographically distribuited. In particular remote sensing and genetics applications will be consider. In addition the scientific community will be trained on the use of hardware/software tools oriented to geografically distributed computing.
AUTONOMIC APPLICATIVE SERVICES FOR CRITICAL INFRASTRUCRURES(2005, coordinator Gianfranco Mascari), Department ICT.
Large distributed organizations (as CNR)have a lot of applications, information and hardware resources spread over national and international territorries. Such organizations are software intensive distributed systems. However in many cases each organization unit (institute, departement, etc) have access to its own local resources. Grid computing and Autonomic computing attempt to overcome the limitations of traditional distributed computing in terms of accessibility and self management, respectively, of the resources. Our research activity aims to develop an integrated methodology (models and architectures)for
- safe, ubiquous and homogeneous acces to distributed resources;
- inteligent resources management;
- use of critical applications.
QUANTITATIVE MODELING OF COMPLEX BIOLOGICAL SYSTEMS (2007, coordinator Filippo Castiglione), Life Sciences Department.
Mathematical biology aims to the understainding of those phenomena that appear too complex to be analysed without the help of the exact sciences. With the mathematical or computational models it is possible to attempt
a quantitative description of the laboratory experimental data. Mathematical models need to first produce results that are perfectly in line with experimental observations, then it should produce valuable predictions. The first goal of the present research area is to work out models with predictive values on the basis of clinical or experimental data coming from medical doctors and biologist that collaborate with us. A second goal is to provide a set of mathematical tools to describe the dynamics of complex system at many levels of description. A final goal is to create a reference point within the CNR and for all institutes of the Life Sciences Department to foster a genuine interdisciplinary research between mathematics and biology.
ANALYSIS AND SYNTHESIS OF HETEREOGENOUS DATA FOR THE MONITORING ASSISTED OF THE DECAY OF CULTURAL HERITAGE (2007, 2008, coordinator Laura Moltedo), Cultural Heritage Department
The Cultural heritage object and the environment where it is embedded constitute a system about which functions have to be developed and realized in order to process data related to the heritage object (for instance image) and the environment (physical chemical data). In this framework, the goal of the research is to provide analysis and synthesis tools, that allow the state of conservation monitoring for the study case object.
Digital restoration functions are applied to historical importance images; segmentation techniques are applied to color and gray level images in order to extract decay regions or remove the noise, together to compression techniques. Data originated from laser scanner constitute the tool for 3D modeling and physical-chemical data are processed by means of 2D and 3D visualization and animation techniques. Then appropriate functions are developed in order to define virtual sensors.
Different types of chemical-mechanical decay are studied by numerical mathematical models.
PARAMETER CHOICE RULES (2005, 2006, coordinator Gabriele Inglese), Department Earth & Environment
In order to recover the vertical profile of athmospheric parameters (temperature, concentration of athmospheric chemical species, ...) we must solve an ill-posed nonlinear integral equation. Approximate solutions of the problem comes from an optimization procedure and cannot be obtained without using regularization methods (Tikhonov, stopping iteration or maximum entropy). These methods require the choice of a parameter related to the weight of the penalization in the cost functional that must be minimized.
DEVELOPMENT OF USER-FRIENDLY SOFTWARE FOR THE ANALYSIS OF REMOTE SENSED DATA (2005, 2006, coordinator Umberto Amato), Department Earth & Environment
Processing of remotely sensed data often requires use of resources (i.e., data, computing, software) geographically spread at different sites. Aim of the research is to develop suitable interfaces that make software produced for remote sensing applications friendly run by users remotely connected. In particular, these interfaces will be developed both in advanced programming languages and in special environments (e.g., Matlab). They will be particularly suited for cluster computing architectures.
DEVELOPMENT OF CLASSIFICATION AND SEGMENTATION METHODS (2005, 2006, coordinator Italia De Feis), Department Earth & Environment
The research aims at developing statistical methods for remote sensing, particularly regression of multivariate data, classification and segmentation of images. Several sensors will be considered (e.g., interferometers like IASI, imaging radiometers like MIVIS, SEVIRI) onboard satellites and aircraft. Applications will be concerned both with land features and atmospheric corrections for detecting land features.
NUMERICAL SIMULATION OF PHYSICAL AND BIOLOGICAL ASPECTS OF THE ANTARTIC SUBGLACIAL LAKES (2006, coordinator Daniela Mansutti), Department of Earth and Environment.
The physical and biological processes in the antartic subglacial lakes are relevant to the understanding of the climate changes both during recent times and geological eras and also for the informations that can be extrapolated on the primordial forms of life. In particular this research is structured according to the following issues: 1) physical characterization of the Concordia lake (evolution of the icy crust, of the thermo-fluid dynamics of the basin and of the ice/water interface); 2)study of the origin of the subglacial idrographic net in Antartics (area Dome C) and in the icy satellites such as Europa, the jovian moon; 3) study of the evolution of a subglacial lake's biota due to ecosystem changes. In each case, first, the mathematical and numerical models will be determined and characterized, then, the computational codes will be built and validated, and finally, tests based on real data will be developed.
DEVELOPMENT OF INNOVATIVE DIAGNOSTIC METHODOLOGIES FOR QUANTITATIVE IMAGING, QUALITY AND SAFETY CHECK, OPTIMIZATION OF TREATMENT PLANS (2005, resp. Umberto Amato), research of the main unit with the same name at the Department of Medicine
The referent Institute of the present research is IBB. Image diagnostics is a fundamental tool of modern medicine, both for understanding physiological and pathological processes, and because the continuous improvement of the accuracy of the methods permits faster and more accurate diagnoses. General objective of the research is to advance in the technology of Image Diagnostics and Radiotherapy giving new tools to the research in medicine and clinics. Particular objective of the research is to develop statistical methods for the quantitative analysis and synthesis of images coming from several equipments (MR, CT, PET, etc..) to use both to improve knowledge in Phisyopathology and to better characterize and monitor deseases.
A MULTIDISCIPLINAR APPROACH FOR THE DEFINITION OF MOLECULAR NETWORKS REGULATING MENDELIAN AND MULTIFACTORIAL INHERITANCE TRAITS (2005, 2006, 2007,coordinator Francesca Carfora), Life Sciences Department
Aim of the research is the identification of genes possibly related to pathologies of interest on the base of quantitative traits. Statistical linkage analysis uses genealogy, health status and DNA information to investigate the genetic component of a disease. In particular, studies on isolated populations such as the ones in the Cilento area (more than 1500 living inhabitants in a pedigree of about 10000 people) revealed to be very useful. The identification of a few significant traits among the thousand analyzed, basing on a limited number of observations, is a statistical problem requiring the development of methods for model selection, hypothesis testing, dimensionality reduction which take into account the complexity of the involved genealogies.
CELLULAR AND MOLECULAR ANALYSIS OF SYNTHETIC VACCINE-INDUCED IMMUNE RESPONSE (2005, 2006,2007 coordinator Antonia Vecchio), Life Sciences Department
The immune system is repeatedly stimulated by a vast array of antigens in a normal environment. Thus, in peripheral lymphoid system, mature lymphocytes compete with each other for the possibility either to expand during antigen-specific immune response, or to proliferate in response to growth factors. Still, the total number of CD8 T lymphocytes in peripheral lymphoid system (i.e. in blood, lymph, spleen, lymph nodes) doesn't change, being kept in constant equilibrium by cell proliferation and death. The kinetics of cell proliferation and death of mature CD8 T lymphocytes are still largely unknown. A better knowledge in this area would be really important for the rational design of vaccine protocols. In our project, we will integrate immunology and mathematical knowledge to develop mathematical models based on ordinary differential and/or Volterra integral equations to simulate proliferation of mature CD8 T lymphocytes. This will allow us to explain by a mathematical model the experimental findings on CD8 T cell proliferation and death, as well as to make predictions on long term maintenance of the immune response.
STATISTICAL METHODS ORIENTED TO GRID COMPUTING FOR GENETICS (2006, coordinator Claudia Angelini), unit of research in Bioinformatics and molecular modelling at Molecular Design Department.
Microarray data analysis is recognized to be very useful both for detecting genes that play a crucial role in several pathologies (by identifying differentially expressed genes) and for understanding their functional regulation mechanism (by finding their binding sites). Recently, several statistical methods has been proposed in the literature for the analysis of genetic data, mainly tailored for analyzing most of the 'classical' experiments. However, for what concerns microarray experiments whose aim is to monitor gene expression level along the time (for example when one wants to investigate the effect of a given treatment in a cell as a function if the time after the treatment) then very few methods are available within the scientific community. Aim of the present research is to develop and to implement on cluster and/or grid efficient statistical methods particularly suited to deal with time course microarray data analysis. The software packages developed as final product of our research will be available for the scientific community. Additionally we will also study new algorithms for clustering temporal profiles and for the detection of bindings site in co-regulated genes. The latter turns to be very useful for understanding regulation mechanisms of genes
APPLICATIVE SERVICES FOR COMPLEX INFORMATION PROCESSING and COMPLEX STRUCTURES SERVIZI APPLICATIVI PER IL TRATTAMENTO DI INFORMAZIONI E STRUTTURE COMPLESSE (2006, coordinator Gianfranco Mascari), Department of Materials and Devices.
A double-triad is obtained by considering adaptation between: - Objectives of Services Consumers: specification, realization, evaluation, and - Constraints of Services Providers: specification, satisfaction, evaluation. A second double-triad is obtained by considering adaptation between: - Request: preparation, realization and evaluation, and - Supply: preparation, execution and evaluation. A fractal double-triad is obtained by considering the entanglement of these adaptation patterns (inspired by the quark-antiquark double triads of particle physics).
MODELING AND SIMULATION (2005, 2006, coordinator Riccardo March), a general research project which is not linked to a particular Department.
Modeling and simulations arise in science and technology, thanks to new mathematical and computational techniques that allow to manage complex problems. For instance the applications of fields such as the calculus of variations and partial differential equations have been extended to new subjects.Moreover the contribution of interdisciplinary know-how has permitted developments which are deeply innovating.
The subjects of the present research are both theoretical and applicative: variational problems with free discontinuities in image analysis, applications of General Relativity, Neumann problems for nonlinear elliptic operators, nonstandard Banach spaces, theory and applications of hyperbolic systems, computational complexity, distributed computational systems.
STATISTICAL METHODS ORIENTED TO GRID COMPUTING FOR GENETICS (2007, 2008 coordinator Claudia Angelini), unit of research in Mathematical and statistical methods for genetics and proteomics wothin the interdipartimental projects Bioinformatics.
Microarray data analysis is recognized to be very useful both for detecting genes that play a crucial role in several pathologies (by identifying differentially expressed genes) and for understanding their functional regulation mechanism (by finding their binding sites). Recently the activities of the research unit were focused on the development of novel statistical methods able to monitor the change in the gene expression's profiles as a function of time (for example when one wants to investigate the effect of a given treatment in a cell as a function if the time after the treatment) . In this framework we developed new models and related software also graphical user friendly interface. The trial versions of the software are available from the authors and accessible to researchers worldwide. Research currently deal on improving the proposed methodologies and apply it in broader contexts. In particular the activities will be also devoted on statistical algorithms for clustering and on new methods for identifing gene network. This research is designed to help to understand the mechanisms of genetic regulation.
BAYESIAN METHODS FOR VARIABLE SELECTIONS WITH APPLICATION IN GENOMICS,(2007, 2008 resp. Claudia Angelini), Unit of research RTL
Genetics is a research area in great and promising expansion to new important discoveries that can help in the understanding, treatment and prevention of many diseases. However it also poses new and important challenges for mathematics, statistics and information due to large amounts of data that are involved, which have to be analyzed and interpreted. The objective of this research is to develop and provide innovative statistical tools for handling genetic data and to study their mathematical properties. In particular, recent works show that Bayesian methods for the selection of variables can be very useful for analyzing microarray data, where p>> n (where p is the number of variables and n the number of samples). However, so far such methods lack of a rigorous mathematical validation. Aim of this unity of research in to focus on the study and development of such methods (and more in general to develop methods for the selection of models) with particular interest on their statistical and computational properties. In parallel, the units of research will propose the application of proposed methods to real problems, such as determining the sites of transcription of co-regulated genes in similar species