myUniCal
myUniCal
University of CalabriaHRS4RUniversity of CalabriaHRS4R
University of CalabriaHRS4R
Search
Search
Address bookAddress book
Italian English Chinese French Portuguese Espanol Arabic Russian
William SPATARO - Professori Associati

William SPATARO

Professori Associati

Informatica (INFO-01/A)

Contacts
No items here
Since January 2005, he is an Assistant Professor at the Science Faculty of the University of Calabria for the scientific sector INF/01 (Computer Science) and member of the Department of Mathematics. Teaching activities - Lecturer of “Parallel Algorithms and Distributed Systems”, Master Computer Science degree course, University of Calabria, Italy. - Lecturer of “Computer Science”, Natural Scinces and Culturage Heritage Bachelor degree course, University of Calabria, Italy. Main lines of research The following outline summarizes the research topics: 1 - Modeling and Simulation of Complex Natural Phenomena with Cellular Automata. 2 - Optimization with Evolutionary Algorithms and Multi-objective Evolutionary Algorithms. 3 - Development of Parallel Applications for Distributed Architectures and Load Balancing Algorithms. Research projects He has participated in several national (FIRB, PRIN, etc.) and European (HPCN-Esprit, Environment and Climate Programme) regarding modeling and parallel computing. Peer reviewing He has referred several papers for the following international journals: ACM Transactions on Modeling and Computer Simulation, Applied Mathematics and Computation, Journal of SuperComputing, Fundamenta Informaticae, Mathematical and Computer Modelling, Computer and Geosciences, Annals of Geophysics. He is member of the Scientific Committee for the International Conference on Scientific Computing (CSC), the European Modeling and Simulation Symposium (EMSS) and the International Conference on Parallel Processing and Applied Mathematics (PPAM).
SCIENTIFIC AREAS OF INTEREST / RESEARCH Parallel Computing, Modelling, Numerical Simulation, Cellular Automata, Genetic Algorithms. POSITIONS HELD Assistant Professor, Scientific Field INF/01 – Computer Sciences, Department of Mathematics at the University of Calabria. Member of the Department of Mathematics and member of the “PhD in Mathematics and Computer Science” scientific board in the same department. INTERNATIONAL SCHOOLS “COURSE ON THE STRUCTURE OF PARALLEL MACHINES” held in Trento in June 1991, organized by the Fibonacci Institute of Trento. AWARDS WorldComp 2010 – The 2010 World Congress in Computer Science, Computer Engineering, and Applied Computing. “Studies on Cellular Automata”. RESEARCH ACTIVITIES The research is located within the Computational Science field and, therefore, in a interdisciplinary context, touching a variety of fields: from modeling to Parallel Computing, from optimization to the study of complex systems. However, the current representative research regards computer science applications (mostly by means of Parallel Computing) to the modeling and simulation of geological processes, studies that are having an increasing interest in the scientific community, both for theoretical aspects and, more specifically, for possible applications related to the assessment and mitigation of geological events (such as debris flows, invasion lava, pyroclastic flows, floods, etc.) The studies were substantiated in the development of a methodology for the use of Cellular Automata (CA), parallel models that are computationally equivalent to Turing Machines, for the study of natural and artificial complex systems that evolve on the basis of local interactions at a macroscopic level and using, for their optimization, techniques based on the Genetic Algorithm computational paradigm. Model implementation is performed on sequential and parallel computers, using Message Passing (MPI, Message Passing Interface) techniques, shared-memory (eg OpenMP) models and, recently, techniques based on the promising GPGPU (General Purpose computing on Graphics Processing Units) paradigm. These techniques are considered in this research mainly to improve performance and to reduce computing timings. What follows are details of the research activities spread over different topics. As indicated above, researches often have interdisciplinary connotations, and therefore can be placed in more than one of the categories listed. Please refer to scientific publications for further details on the theory, techniques and solutions adopted for each research activity. PARALLEL COMPUTING AND RELATED RESEARCHES From the establishment of the MIUR Centre of Excellence for High Performance Computing (HPCC) at the University of Calabria in 2001, several geological simulation (based on CA) models and optimization (based on the Genetic Algorithms paradigm) ones have been implemented on high-performance parallel computers that are located at the Centre. In both cases – for CA and GA - given their inherent parallel computing nature, the MPI (Message Passing Interface – standard de-facto in parallel programming) library has used, for the message exchange between computing processes. The experience gained over years regarding Message Passing techniques, modeling and optimization has led to the creation of a C ++ open-source parallel library for Scientific Computing: libAuToti. The library, designed specifically for the definition of Macroscopic Cellular Automata and completely portable across different platforms (ie Windows, Linux, MacOS), allows a simple and concise definition of the transition function and of other features of the model and can be used on sequential and parallel architectures (thanks to the adoption of MPI), hiding, in this case, implementation details to the user. At the same time, several viewing modules were also implemented, written using OpenGL, Qt, VTK and wxWidgets, which can be directly used with programs that use libAuToti, allowing an interactive graphic interface within a simulation. In addition, as part of library development, a dynamic load balancing algorithm was developed, available both for multithread and MPI versions of the library, that allows to significantly reduce the execution time of simulations. Recently, the OpenMP (Open Multi Processing) API has been used for the parallelization of CA simulation models. OpenMP is an open standard for parallel programming on shared memory architectures, integrated in most compilers on nearly all software and hardware platforms currently available on the market (eg multi-core processors). In conjunction with this line of research, one regarding Web 2.0 applications for scientific computing has begun with SWII (Sciddica Web Interactive Interface), a first example of application of a Cellular Automata model to be used directly over the Internet. The user interface is based on AJAX (Asynchronous JavaScript And XML), provides a clear separation between the client and server application, and a high degree of interactivity and usability for the user. The perfect application modularity allows the replacement of the simulation model with any other (even not based on the CA paradigm). Another research line concerning the implementation of CA computational models by means of GPGPU (General Purpose Programming with Graphics Processing Units) is currently been carried out, also thanks to the official collaboration with the School of Computing and Mathematics (Faculty of Science and Technology) of the University of Plymouth (UK). GPUs (Graphical Processing Units) are usually adopted for their efficiency for computer graphic purposes (e.g. rendering processes, three dimensional image creation). However, recently, they are also exploited for applications not directly related to the field of Computer Graphics with excellent results. Different implementations of the CA lava flow model SCIARA were conducted on high-end graphic processors (e.g., TESLA C1060 and C2070, specifically designed for High Performance Computing) and on commodity graphic cards (Geforce GT330M, GTX Series, etc), in all cases using the NVIDIA CUDA architecture. The results of simulations demonstrated performance improvements in the order of a factor of 100, depending on the size of the problem and the type of memory optimization, confirming the validity of the GPGPU for the implementation of Cellular Automata models as an alternative to more expensive hardware solutions, such as cluster or multi-core machines. OPTIMIZATION: GENETIC ALGORITHMS As part of the validation and optimization phase of parameters used in the various CA models, from 2004 researches and applications on automatic optimization methods such as Genetic Algorithms (GA) have taken place. GA are general-purpose search algorithms inspired by genetics and natural selection. GA simulate the evolution of a population of candidate solutions to a specific research problem, encouraging the "reproduction" of individuals (genotype) that are most appropriate. The validation of CA simulation models occurred in the past by the classical method of trial and error. While this methodology is in the case of models with few parameters, the introduction of additional terms of variability in the outcome of models required the adoption of automated (and more sophisticated) optimization techniques. Applications of GA (in particular versions implemented on parallel computers) for the optimization of parameters of cellular automata models for the simulation of lava and debris flows have proven the validity of the method, both in terms of computational efficiency, and in terms of goodness of the produced solutions. Recently, studies on multi-objective evolutionary algorithms are been carried out, in order to perform the calibration of simulation models on the basis of different parameters. COMPLEX PHENOMENA MODELLING: LAVA FLOWS The decision to pursue the modeling of lava flows by Cellular Automata is derived from the fact that the numerical resolution of traditional computational methods, whose solution is related to that of differential equations that typically govern lava flows (the Navier-Stokes equations), is very difficult. These equations are particularly complex, owing that the rheological characteristics of lava can vary, during cooling, from those of a quasi-Newtonian fluid to a Bingham one. The initial model, called SCIARA, has been improved, extended and applied to other real lava flow events. Furthermore, the model was also applied on real events with different chemical-physical characteristics from Etnean lavas. In this regard, the collaboration with researchers of the IPGP (Institute de Physique du Globe de Paris) and Observatoires Volcanologiques of Reunion Island has proven to be very constructive. The overall simulation methodology has achieved important results and has been appreciated by the International Scientific Community thanks to its good response during the 2001 crisis of the My Etna in July 2001, when the research team has been at the forefront of the development scenarios and crisis intervention, participating at the invitation of the INGV (National Institute of Geophysics and Volcanology) and transferring monitoring, simulation and research structures directly on Mt. Etna. The last phase of the use of these models regards applications in high risk Etnean areas, for the aim of creating risk maps for civil protection purposes. Results have been published in the prestigious Journal of Geophysical Research, where a new lava invasion susceptibility map regarding some highly inhabited areas of the volcano has been presented. Several have been, referring to this line of research, participations in both national and European projects. COMPLEX PHENOMENA MODELLING: DEBRIS FLOWS The collaboration with CNR-IRPI of Cosenza (since 1995), has led to the realization of research projects involving the development of Cellular Automata models for simulation debris flows and porous soil rainwater infiltration and motion. Numerous publications and communications to national and international conferences are good examples of the results obtained. The first phenomenon on which the SCIDDICA (Simulation through Computational Innovative methods for the Detection of Debris flow path using Interactive Cellular Automata) model has been validated regards the Mount Ontake (Japan) volcano 1984 event. The SCIDDICA model was further improved within the European Project PROGRAMME ENVIRONMENT AND CLIMATE - Runout Subproject (Partners: University College London, University of Ferrara, CNR-IRPI Padova, University of Las Palmas, Canary Islands, University of Calabria) and applied to other types of landslides (eg Tessina - Italy). The positive results described above, were further validated for other types of landslides, such as mud-flows, which occurred during the tragic events in May 1998 in Sarno (Campania), and tackled within a national project for the Campania Region starting from 1999 to 2002, entitled "DEVELOPMENT OF AN EXPERIMENTAL MODELLING SPACE - TIME OF EVOLUTIONARY PROCESSES fOR tHE TERRITORY OF THE ENVIRONMENT AND RISK MITIGATION". It is in for models, which require the tuning of various parameters for calibration purposes, that automatic optimization techniques based on Genetic Algorithms, as described above, were used for the first time. COMPLEX PHENOMENA MODELLING: PYROCLASTIC FLOWS Since 2001, a new collaboration with the Department of Earth Sciences, University of Calabria, regarding a first study of the application of Cellular Automata for the simulation of pyroclastic flows has been carried out. Thanks to contacts with international researchers (mainly Dutch and Filipino) of the simulation and study of pyroclastic flows fields, the first version of the model was applied to a real case of catastrophic pyroclastic event occurred in the Philippines in 1991, on Mount Pinatubo, and Soufrière Hills, Montserrat in 1996, with satisfactory and encouraging results. APPLICATION OF PARALLEL COMPUTERS FOR THE MODELLING OF COMPLEX PHENOMENA: SOIL BIOREMEDIATION Cellular Automata based models for soil bioremediation were addressed in two ESPRIT projects funded by the European Commission: CABOTO-CAPRI (1995-96) and the subsequent COLOMBO (1998-2000). The two projects involved the development and implementation of a portable simulation environment on distributed memory computers, using the standard MPI (Message Passing Interface) library. Project results have been excellent, thanks to the constant and valuable collaboration with researchers from, among other partners, the Centre for Environmental Research Montecatini di Ravenna and Edinburgh Parallel Computing Centre (UK). COMPLEX PHENOMENA MODELLING: HIGHWAY TRAFFIC In conjunction with researchers from the Department of Urban Planning at the University of Calabria, Rende, and the Department of Electrical Engineering and Applied Mathematics, University of Reggio Calabria, a model for the microscopic simulation of highway traffic has been developed, with the aim to cover a whole variety of the main features of the phenomenon in comparison with traditional ones, which usually describe only partial aspects. Starting from the partial implementation of the model of sequential computers, the simulations have achieved a good agreement with experimental data of highway traffic, relative to highways to two lanes in congested traffic conditions. Subsequently, the model has been refined, considering three-lane highways and high traffic. The results of the simulations have reproduced correctly the experimental data for a wide range of traffic conditions of the highway. RESEARCH PROJECTS A. Project: PON “RICERCA E COMPETITIVITÀ” 2007-2013 - n. 01_01503 – (CUP B31H11000370005): Landslides Early Warning - Integrated systems for monitoring, early warning and mitigation of landslide risk along main roads - Period 2011 – 2013. B. Project: INGV – Etna 2008-2010 – Financed by the Italian Civil Protection Agency - Definition of lava flow invasion susceptibility maps for Mt Etna - Period 2008 – 2010. C. Project: PRIN 2009 - Development of a unified model for the simulation of flow-type and granular flow landslides by the Cellular Automata computational paradigm, its implementation on parallel platforms for mitigation and prevention in risk areas (Positively evaluated but not funded) D. Project: PRIN 2007: - PROMETEO - Process Management for Support to Operators and Emergency Teams. (Positively evaluated but not funded) E. Project: INGV V3/6 – Etna 2005-2007 - Financed by the Italian Civil Protection Agency - Definition of lava flow invasion susceptibility maps for Mt Etna - Period 2005 – 2007 F. Project: PRIN 2006- prot. 2006042319 - Integration of innovative techniques for geological and geophysical exploration of onshore and offshore for the study of coastal processes of gravitational instability – Period: 2007-2009. G. Project: MIUR - FIRB (Fund for Investment in Basic Research) - Lava flow simulation by cellular automata - Protocol: RBAU01RMZ4 – Period: 2003-2006. H. Project: Development of a Space-Time Modeling Experimental Evolutionary Processes Environment and Spatial Planning for Hazard Mitigation Financed by the Campania Region, Period: 2002 - 2005. I. European Project: Programme EnviroNment and Climate- FLOW Subproject (Sponsor: European Commision).– Period 1999 – 2001. L. Project: HPCN-COLOMBO (Sponsor: European Commision). Period 1998 – 2000. M. European project: European Programme EnviroNment and Climate Project – RUNOUT subproject. (Sponsor: European Commision). Period 1998 – 2000. N. Project: HPCN-ESPRIT-CAPRI (Calcolo ad Alte PRestazioni per le Imprese)-sub-project CABOTO (Sponsor: European Commision). Period – 1995 – 1996. EDITORIAL BOARD, ETC Guest Editor – "Special Issue on Cellular Automata" on The Journal of Supercomputing (ISSN: 0920-8542) – To appear in 2012. PEER REVIEWING He has referred several papers for the following international journals: ACM Transactions on Modeling and Computer Simulation, Applied Mathematics and Computation, Journal of SuperComputing, Fundamenta Informaticae, Mathematical and Computer Modelling, Computer and Geosciences, Annals of Geophysics. He has referred several papers for the following international conferences: International Conference on Cellular Automata for Research and Industry (ACRI), Parallel Processing and Applied Mathematics (PPAM), International Conference on Scientific Computing (CSC). He is member of the Scientific Committee for the International Conference on Scientific Computing (CSC), the European Modeling and Simulation Symposium (EMSS) and the International Conference on Parallel Processing and Applied Mathematics (PPAM). TEACHING ACTIVITIES - Lecturer of “Parallel Algorithms and Distributed Systems”, Master Computer Science degree course, University of Calabria, Italy. - Lecturer of “Computer Science”, Natural Scinces and Culturage Heritage Bachelor degree course, University of Calabria, Italy. SUPERVISOR OF THESES He is advisor of over 30 theses for the Master of Computer Science course at Department of Mathematics, University of Calabria.
No items here
Back to previous page
Contacts