- +30-2410-684522
Ilias K Savvas
University of Thessaly, Department of Digital Systems, Faculty Member
- Cloud Computing, Distributed Algorithms, High Performance Distributed System, P2P, Distributed Data Mining, Dynamic Load Balancing, and 7 moreHigh Performance Scientific Computing, Distributed Computing, Parallel Programming, Parallel Algorithms, Parallel Computing, Clustering and Classification Methods, and Quantum Computingedit
- Ilias K. Savvas received his Ph.D. in Computer Science at the University College of Dublin in Ireland. He is a Profes... moreIlias K. Savvas received his Ph.D. in Computer Science at the University College of Dublin in Ireland. He is a Professor of Computer Science, Department of Digital Systems, University of Thessaly, Greece, and currently Dean of School of Technology, University of Thessaly. He was the head of the Department of Computer Science and Engineering of T.E.I. of Thessaly and Dean of the School of Technological Applications (T.E.I. of Thessaly). His research interests include issues related to Parallel and Distributed Computing, Data Mining, Quantum Computing, and High-Performance Computing. He serves as a program/scientific committee member and also as a reviewer of many international conferences and journals and he was the General Chair of the 19th IEEE WETICE conference and Quantum Computing: Current state and Future trends, PCI 2021, and 20222. He is the author of a lot of research studies published in international journals, conference proceedings as well as many book chapters and a book in Quantum Computing.edit
The drug discovery process is a rigorous and time-consuming endeavor, typically requiring several years of extensive research and development. Although classical machine learning (ML) has proven successful in this field, its computational... more
The drug discovery process is a rigorous and time-consuming endeavor, typically requiring several years of extensive research and development. Although classical machine learning (ML) has proven successful in this field, its computational demands in terms of speed and resources are significant. In recent years, researchers have sought to explore the potential benefits of quantum computing (QC) in the context of ML, leading to the emergence of Quantum Machine Learning (QML) as a distinct research field. The objective of the current study is twofold: first, to present a review of the proposed QML algorithms for application in the drug discovery pipeline, and second, to compare QML algorithms with their classical and hybrid counterparts in terms of their efficiency. A query-based search of various databases took place, and five different categories of algorithms were identified in which QML was implemented. The majority of QML applications in drug discovery are primarily focused on the...
Research Interests:
The process to find the prime factors of a large number, the “factoring problem” is believed to be a very hard problem. For this reason, it is the cornerstone of modern cryptographic schemes, like RSA cryptosystem. In 1994, Professor... more
The process to find the prime factors of a large number, the “factoring problem” is believed to be a very hard problem. For this reason, it is the cornerstone of modern cryptographic schemes, like RSA cryptosystem. In 1994, Professor Peter Shor proposed a new polynomial-time quantum algorithm that finds the prime factors of a number with many digits. This was a bolt from the blue for the security of transactions and electronic communications and became an example of how quantum computing changes our perception of security and safety. In this paper Shor's Algorithm is presented and an implementation, a way to factor number 21 is described. In addition, some reliability issues of quantum devices were considered in order to explore the potentiality of Shor's algorithm.
Research Interests:
Research Interests:
Data warehouse (DW) systems provide the best solution for intelligent data analysis and decision-making. Changes applied to data gradually in real life have to be projected to the DW. Slowly changing dimension (SCD) refers to the... more
Data warehouse (DW) systems provide the best solution for intelligent data analysis and decision-making. Changes applied to data gradually in real life have to be projected to the DW. Slowly changing dimension (SCD) refers to the potential volatility of DW dimension members. The treatment of SCDs has a significant impact over the quality of data analysis. A new SCD type, Type N, is proposed in this research paper, which encapsulates volatile data into historical clusters. Type N preserves complete history of changes, additional tables, columns, and rows are not required, extra join operations are omitted, and surrogate keys are avoided. Type N is implemented and compared to other SCD types. Good candidates for practicing SCDs are spatiotemporal objects (i.e., objects whose shape or geometry evolves slowly over time). The case study used and implemented in this paper concerns shape-shifting constructions (i.e., buildings that respond to changing weather conditions or the way people u...
Research Interests:
Περιέχει το πλήρες κείμενοΗ Ηλεκτρονική Ανταλλαγή Δεδομένων (EDI) αφορά την από Η/Υ προς Η/Υ ανταλλαγή εμπορικών και διοικητικών πληροφοριών/ παραστατικών μέσω μιας τυποποιημένης μορφής δεδομένων και με την ελάχιστη δυνατή ανθρώπινη... more
Περιέχει το πλήρες κείμενοΗ Ηλεκτρονική Ανταλλαγή Δεδομένων (EDI) αφορά την από Η/Υ προς Η/Υ ανταλλαγή εμπορικών και διοικητικών πληροφοριών/ παραστατικών μέσω μιας τυποποιημένης μορφής δεδομένων και με την ελάχιστη δυνατή ανθρώπινη παρέμβαση. Τα τυποποιημένα EDI μηνύματα είναι βασισμένα σε κοινά επιχειρησιακά έγγραφα όπως τιμολόγια, παραγγελίες αγοράς, φορτωτικές και πιστωτικά και τα οποία διακινούνται από Η/Υ σε Η/Υ μέσω τηλεπικοινωνιακών συνδέσεων χωρίς ανθρώπινη παρέμβαση ή μετάφραση. Οι ακαδημαϊκές βιβλιοθήκες, ενώ δεν δραστηριοποιούνται επιχειρηματικά, λειτουργούν σε ένα καθαρά επιχειρησιακό περιβάλλον. Οι βιβλιοθήκες αγοράζουν βιβλία, περιοδικά, CD-ROMs και άλλο υλικό από πολλούς προμηθευτές, τιμολογούνται για αυτό το υλικό και προβαίνουν σε πληρωμές. Επίσης, εμπλέκονται και σε άλλες εμπορικές συναλλαγές όπως πληρωμές φωτοαντιγράφων ή εκτυπώσεων, παραδόσεις υλικού στους χρήστες τους και διαδικασίες διαδανεισμού. Όλα αυτά δείχνουν, ότι οι βιβλιοθήκες γενικά λειτουργούν σε πολλ...
Research Interests:
Περιέχει το πλήρες κείμενοΗ επίδραση του Ιστού (Web) έχει ως αποτέλεσμα να επιτρέπει όλο και σε περισσότερους χρήστες του Διαδικτύου να έχουν τις ευκαιρίες απομακρυσμένης πρόσβασης, του ηλεκτρονικού ¨ξεφυλλίσματος¨, της αναζήτησης και της... more
Περιέχει το πλήρες κείμενοΗ επίδραση του Ιστού (Web) έχει ως αποτέλεσμα να επιτρέπει όλο και σε περισσότερους χρήστες του Διαδικτύου να έχουν τις ευκαιρίες απομακρυσμένης πρόσβασης, του ηλεκτρονικού ¨ξεφυλλίσματος¨, της αναζήτησης και της απευθείας διάθεσης τεκμηρίων από πληροφοριακές πηγές. Η εξέλιξη των e-prints ηλεκτρονικών αρχείων παρέχει νέες υπηρεσίες αναζήτησης για τους χρήστες σε αντίγραφα ερευνητικών εργασιών για να εντοπίζουν τις πληροφορίες ευκολότερα, γρηγορότερα και χωρίς κόστος. Ένα καινοτόμο παράδειγμα e-print ηλεκτρονικού αρχείου είναι το arXiv.org αρχείο. Παραδοσιακά χρησιμοποιείται από τους φυσικούς που είναι πρωτοπόροι στον τρόπο ηλεκτρονικής επικοινωνίας. Για περισσότερα από 13 έτη ανταλλάσσουν τα ερευνητικά αποτελέσματά τους ηλεκτρονικά και μπορούν να ενημερώνονται για οτιδήποτε νέο κυκλοφορεί στον κλάδο τους άμεσα. Σ΄ αυτή τη μελέτη παρουσιάζονται οι τρόποι ανάκτησης πληροφοριών των φυσικών και ειδικότερα εξετάζεται πώς οι χρήστες επιτυγχάνουν την πρόσβαση στα ...
Research Interests:
Detection of point anomalies is a very important issue in a large scale of fields from Astronomy and Biology to network intrusions. Clustering has been employed by many researchers to solve such problems and DBSCAN seems like the most... more
Detection of point anomalies is a very important issue in a large scale of fields from Astronomy and Biology to network intrusions. Clustering has been employed by many researchers to solve such problems and DBSCAN seems like the most efficient technique. Due to its high computational complexity, this work focused on decreasing it by decreasing the dimensionality of the data points. For this reason, Principal Components Analysis used and then DBSCAN applied on the new data sets provided by PCA. The quality of the experimental results was very promising proving that such an approach can be adapted. In addition, the performance of the combined PCA and DBSCAN was examined. Our analysis shows that a speedup of 25% was achieved while the quality was 80% reducing the dimensionality of data set to half.
Research Interests:
Im theoretischen Rahmen der Systemischen Geopolitik wird der Versuch unternommen, Data-Mining Techniken – vorerst Clusteranalyse – im Bereich der internationalen Beziehungen anzuwenden. Schon die fruhe Entwicklungsphase eines bestandigen... more
Im theoretischen Rahmen der Systemischen Geopolitik wird der Versuch unternommen, Data-Mining Techniken – vorerst Clusteranalyse – im Bereich der internationalen Beziehungen anzuwenden. Schon die fruhe Entwicklungsphase eines bestandigen Analysemodells fuhrτ nicht nur zu verfahrenstechnischen (z.B. Vergleich von Algorithmen) Schlussfolgerungen, sondern auch zur Andeutung geopolitischer Strukturmerkmale, deren Ausbau, Verfeinerung und Verifizierung noch bevorsteht. In der vorliegenden Arbeit wurde versucht, moglichen Strukturanderungen nach dem bevorstehenden Austritt des Vereinigten Konigreichs nachzugehen, indem Indikatoren der kunftigen EU-Konstellation einer Clusteranalyse unterzogen und der entsprechenden 28 EU-Staaten Zusammensetzung gegenubergestellt wurden. Ein besonderer Anstos war Frankreichs Erscheinung am Rand einer haufig auftretenden Deutschland- England–Frankreich Gruppierung und die Frage nach der zukommenden Ordnung an der EU-Spitze. Die neuen Ergebnisse deuten auf F...
Research Interests:
In recent years, machine learning has penetrated a large part of our daily lives, which creates special challenges and impressive progress in this area. Nevertheless, as the amount of daily data is grown, learning time is increased.... more
In recent years, machine learning has penetrated a large part of our daily lives, which creates special challenges and impressive progress in this area. Nevertheless, as the amount of daily data is grown, learning time is increased. Quantum machine learning (QML) may speed up the processing of information and provide great promise in machine learning. However, it is not used in practice yet, because quantum software and hardware challenges are still unsurmountable. This paper provides current research of quantum computing and quantum machine learning algorithms. Also, the quantum vendors, their frameworks, and their platforms are presented. A few fully implemented versions of quantum machine learning are presented, which are easier to be evaluated. Finally, QML's challenges, and problems are discussed.
Research Interests:
The customers' data of Telecommunication Companies represent a powerful tool to explore their behaviour and then to increase their satisfaction. The data produced is too large to extract on time useful information, which could be... more
The customers' data of Telecommunication Companies represent a powerful tool to explore their behaviour and then to increase their satisfaction. The data produced is too large to extract on time useful information, which could be beneficial for both sides, companies and customers. One of the solutions to explore this data is representing it by the adoption of clustering techniques. In this work, two distributed / multi-core versions of clustering algorithms were used, namely DBSCAN and k-means, which both of them cluster data according to its characteristics. While DBSCAN is a density-based spatial clustering algorithm and groups data based on the minimum size of participating objects per cluster and the minimum required distance between them, k-means clusters the data objects according the pre-desired number of groups. Thus, since the two methods use different roads to group the data objects, they form different clusters but each one has its importance depending on the characteristics of the applied method. The experiment results of both proposed distributed / multi-core techniques proved their efficiency.
Research Interests:
Cellular vehicle-to-everything (C-V2X) communication has recently gained attention in industry and academia. Different implementation scenarios have been derived by the 3rd Generation Partnership Project (3GPP) 5th Generation (5G)... more
Cellular vehicle-to-everything (C-V2X) communication has recently gained attention in industry and academia. Different implementation scenarios have been derived by the 3rd Generation Partnership Project (3GPP) 5th Generation (5G) Vehicle-to-Everything (V2X) standard, Release 16. Quality of service (QoS) is important to achieve reliable communication and parameters which have to be considered are reliability, end-to-end latency, data rate, communication range, throughput and vehicle density for an urban area. However, it would be desirable to design a dynamic selecting system (with emphasis on channel coding parameters selection) so that all QoS parameters are satisfied. Having this idea in mind, in this work we examine nine V2X implementation scenarios using Long Term Evolution (LTE) turbo coding with a geometry-based efficient propagation model for vehicle-to-vehicle communication (GEMV), where we consider the above QoS parameters for SOVA, log-MAP and max-log-MAP decoding algorit...
Research Interests:
Big Data explosion is a phenomenon of the 21st century. Nowadays, more and more people are using the internet and creating new data regarding ideas, opinions, feelings or their views on a variety of topics and products. The micro-blogging... more
Big Data explosion is a phenomenon of the 21st century. Nowadays, more and more people are using the internet and creating new data regarding ideas, opinions, feelings or their views on a variety of topics and products. The micro-blogging platform Twitter is very popular and produces massive amount of data every fraction of a second and useful information may be generated from them. Sentiment analysis or opinion mining plays a major role in discovering all this information by analyzing these data and gain better insight of the public opinion on any subject in specific. In this paper a framework is proposed, built on top of the Hadoop ecosystem, for analyzing data from Twitter using a domain-specific Lexicon (in Greek).
Research Interests:
It is our pleasure to welcome you to the 21st Pan-Hellenic Conference on Informatics - PCI 2017. PCI 2017 is a unique event since it offers to the Greek informatics community, researchers and practitioners, a forum to strengthen their... more
It is our pleasure to welcome you to the 21st Pan-Hellenic Conference on Informatics - PCI 2017. PCI 2017 is a unique event since it offers to the Greek informatics community, researchers and practitioners, a forum to strengthen their relationships, to explore new trends in the broad area of informatics and to present and discuss papers on relevant important and timely topics.
Research Interests:
The last years, huge masses of data are produced or extracted by computational systems and independent electronic devices. To exploit this resource, novel methods must be employed or the established ones may be altered in order to... more
The last years, huge masses of data are produced or extracted by computational systems and independent electronic devices. To exploit this resource, novel methods must be employed or the established ones may be altered in order to confront the issues that arise. One of the most fruitful techniques, in order to locate and use information from data sources is clustering, and k-means is a successful representative algorithm which clustersdata according specific characteristics. However, its main disadvantage is the computational complexity which proves the techniques very unproductive to apply on big datasets. Although k-means is a very well studied technique, a fullyoperational distributed version combining the multi-core power of today machines, have not been accepted yet by the scientific community. In this work, a three phase distributed / multi-core version of k-means is presented. The obtained experimental results are very promising and prove the correctness, the scalability, and the effectiveness of the proposed technique.
Research Interests:
The primary goal of this paper is to develop a distributed ontology-based knowledge representation approach useful for data warehouses design in the security applications area. The paper proposes a novel database design for registering... more
The primary goal of this paper is to develop a distributed ontology-based knowledge representation approach useful for data warehouses design in the security applications area. The paper proposes a novel database design for registering security incidents in critical infrastructure on railways. We propose an approach based on the data warehouse architecture that consists of distributed smart database micro services patterns, which are represented by the distributed ontology. This representation is using novel distributed dynamic description logic for knowledge representation. We give the base of distributed dynamic description logic and forming the queries to the designed distributed knowledge bases.
Research Interests:
This work presents a new area of application for clustering techniques in industrial and transport applications. The main aim of the research is to propose the technique for detection of point anomalies in telecommunication traffic... more
This work presents a new area of application for clustering techniques in industrial and transport applications. The main aim of the research is to propose the technique for detection of point anomalies in telecommunication traffic produced by network subsystems of railway intelligent control system. The central idea behind is to apply enhanced DBSCAN algorithms for finding the outliers in traffic which are associated with unintended erroneous events or deliberated attacks targeted to infrastructure malfunction. The traffic flows in a part of the railway intelligent control system has been described in detail. Point anomaly detection in IP-networks data using distributed DBSCAN has been proposed. Series of computation experiments for outlier detection in network traffic has been implemented. The experiments showed the applicability of distributed DBSCAN technique to the robust detection of point anomalies caused by various incidents in the network infrastructure of railway intellige...
Research Interests:
In a cloud based data warehouse (DW), business users can access and query data from multiple sources and geographically distributed places. Business analysts and decision makers are counting on DWs especially for data analysis and... more
In a cloud based data warehouse (DW), business users can access and query data from multiple sources and geographically distributed places. Business analysts and decision makers are counting on DWs especially for data analysis and reporting. Temporal and spatial data are two factors that affect seriously decision-making and marketing strategies and many applications require modelling and special treatment of these kinds of data since they cannot be treated efficiently within a conventional multidimensional database. One main application domain of spatiotemporal data warehousing is telecommunication industry, which is rapidly dominated by massive volume of data. In this paper, a DW schema modelling approach is proposed which integrate in a unified manner temporal and spatial data in a general data warehousing framework. Temporal and spatial data integration becomes more important as the volume and sharing of data grows. The aim of this research work is to facilitate the understanding, querying and management of spatiotemporal data for on-line analytical processing (OLAP). The proposed new spatiotemporal DW schema extends OLAP queries for supporting spatial and temporal queries. A case study is developed and implemented for the telecommunication industry.
Research Interests:
Learning analytics have proved promising capabilities and opportunities to many aspects of academic research and higher education studies. Data-driven insights can significantly contribute to provide solutions for curbing costs and... more
Learning analytics have proved promising capabilities and opportunities to many aspects of academic research and higher education studies. Data-driven insights can significantly contribute to provide solutions for curbing costs and improving education quality. This paper adopts a two-phase machine learning approach, which utilizes both unsupervised and supervised learning techniques for predicting outcomes of students following Higher Education programs of studies. The approach has been applied in a case-study which has been performed in the context of an undergraduate Computer Science curriculum offered by the University of Thessaly in Greece. Students involved in the case study were initially grouped based on the similarity of specific education-related factors and metrics. Using the K-Means algorithm, our clustering experiments revealed the presence of three coherent clusters of students. Subsequently, the discovered clusters were utilized to train prediction models for addressing each particular cluster of students individually. In this regard, two machine learning models were trained for every cluster of students in order to predict the time to degree completion and student enrollment in the offered educational programs. The developed models are claimed to produce predictions with relatively high accuracy. Finally, the paper discusses the potential usefulness of the clustering-aided approach for learning analytics in Higher Education.
Research Interests:
Research Interests:
Nowadays, huge quantities of data are generated by billions of machines and devices. Numerous of methods have been employed, in order to make use of this valuable resource, some of them are altered versions of established known... more
Nowadays, huge quantities of data are generated by billions of machines and devices. Numerous of methods have been employed, in order to make use of this valuable resource, some of them are altered versions of established known algorithms. One of the most seminal methods, in order to mine from data sources is clustering, and k-means is a key algorithm which forms clusters of data according to a set of attributes. However, its main shortcoming is the high computational complexity which proves the k-means very inecient to perform on big datasets. Although k-means is a very well utilized algorithm, a functional distributed variant combining the multi-core power of contemporary machines, have not been accepted yet by researchers. In this work, a three phase distributed/multi-core version of k-means and the analysis of its results are presented. The obtained experimental results are in line with the theoretical outcomes and prove the correctness, eciency, and scalability of the proposed technique.
Research Interests:
Research Interests:
As a geographical method of analyzing power redistribution, Systemic Geopolitical Analysis (according to Ioannis Th. Mazis theoretical basis) proposes a multi-dimensional, interdisciplinary research pattern, which embraces economic,... more
As a geographical method of analyzing power redistribution, Systemic Geopolitical Analysis (according to Ioannis Th. Mazis theoretical basis) proposes a multi-dimensional, interdisciplinary research pattern, which embraces economic, cultural, political and defensive facts. The amount of data produced combining these attributes is extremely large and complex. One of the solutions to explore and analyze this data is clustering it and one of the most popular and useful techniques in order to group data within appropriate sets is k-means algorithm which clusters data according to its characteristics. The main disadvantage is its computational complexity which makes the technique very difficult to apply on big and dynamic data sets. In this study, a parallel version of k-means is used in order to cluster the European Union countries according to their attributes and the results obtained prove the importance of this research.
Research Interests:
Research Interests:
ABSTRACT Nowadays, colossal amount of information is produced by computational systems and electronic instruments such as telescopes, medical devices and so on. To explore these petabytes of data, new fast algorithms must be discovered or... more
ABSTRACT Nowadays, colossal amount of information is produced by computational systems and electronic instruments such as telescopes, medical devices and so on. To explore these petabytes of data, new fast algorithms must be discovered or old ones may be redesigned. One of the most popular and useful techniques in order to discover and extract information from data pools is clustering, and k-means is an algorithm which clusters data according its characteristics. Its main disadvantage is its computational complexity which makes the technique very difficult to apply on big data sets. Although k-means is a very well studied technique, a fully parallel version of it has not been explored yet. In this work, a parallel version of the k-means is presented for 1-d objects. The experimental results obtained are inline with the theoretical outcome and prove both the correctness and the effectiveness of the technique.
Research Interests:
We present a process-algebraic approach for the specification of agent systems where agents participate in joint activities, extending previous work by the first author in (8, 9). While related to existing work on teamwork, such as (15,... more
We present a process-algebraic approach for the specification of agent systems where agents participate in joint activities, extending previous work by the first author in (8, 9). While related to existing work on teamwork, such as (15, 16, 17, 19, 20, 21), our focus here is not on discussing notions bearing on joint intentions or abilities. Rather, we focus on providing a specification language for agent systems and behaviors of agents. The language of behaviors we propose can also serve as a coordination language for specifying the flow of control in joint agent activity. We then provide an operational semantics for the language of social agents that we present, based on a notion of social structure that simplifies the semantics of social plan execution. The formation of the required social structure is part of the process of social plan formation. In this report we assume social plans are already available (pre-compiled or already formed) and focus on providing semantics for their execution.
Research Interests:
A growing interest has been shown recently, concerning buildings as well as different constructions that use transformative and mobile attributes for adapting their shape, size and position in response to different environmental factors,... more
A growing interest has been shown recently, concerning buildings as well as different constructions that use transformative and mobile attributes for adapting their shape, size and position in response to different environmental factors, such as humidity, temperature, wind and sunlight. Responsive architecture as it is called, can exploit climatic conditions and changes for making the most of them for the economy of energy, heating, lighting and much more. In this paper, a data warehouse has been developed for supporting and managing spatiotemporal objects such as shape-shifting constructions. Spatiotemporal data collected from these transformations are good candidates for analysis by data warehouses for decision making and business intelligence. The approach proposed in this research work is based on the integration of space and time dimensions for the management of these kinds of data. A case study is presented where a shape-shifting buildings data warehouse is developed and imple...
Research Interests:
Nowadays, when the data size grows exponentially, it becomes more and more difficult to extract useful information in reasonable time. One very important technique to exploit data is clustering and many algorithms have been proposed like... more
Nowadays, when the data size grows exponentially, it becomes more and more difficult to extract useful information in reasonable time. One very important technique to exploit data is clustering and many algorithms have been proposed like k-means and its variations (k-medians, kernel k-means etc.), DBSCAN, OPTICS and others. The time complexity of all these methods is prohibitive (NP hard) in order to make decisions on time and the solution is either new faster algorithms to be invented, or increase the performance of the old well tested ones. Distributed, parallel, and multi-core GPU computing or even combination of these platforms consist a very promising method to speed up clustering techniques. In this paper, parallel versions of the above mentioned algorithms were used and implemented in order to increase their performance and consequently, their perspectives in several fields like industry, political/social sciences, telecommunications businesses, and intrusion detection in big...
Research Interests:
From Pharmacology to Cryptography and from Geology to Astronomy are some of the scientific fields in which Quantum Computing potentially will take off and fly high. Big Quantum Computing vendors invest a large amount of money in improving... more
From Pharmacology to Cryptography and from Geology to Astronomy are some of the scientific fields in which Quantum Computing potentially will take off and fly high. Big Quantum Computing vendors invest a large amount of money in improving the hardware and they claim that soon enough a quantum program will be hundreds of thousands of times faster than a typical one we know nowadays. But still the reliability of such systems is the main obstacle. In this work, the reliability of real quantum devices is tested and techniques of noise and error correction are presented while measurement error mitigation is explored. In addition, a well-known string matching algorithm (Bernstein-Vazirani) was applied to the real quantum computing device in order to measure its accuracy and reliability. Simulated environments were also used in order to evaluate the results. The results obtained, even if these were not 100% accurate, are very promising which proves that even these days a quantum computer w...
Research Interests:
We are living in a world of heavy data bombing and the term Big Data is a key issue these days. The variety of applications, where huge amounts of data are produced (can be expressed in PBs and more), is great in many areas such as:... more
We are living in a world of heavy data bombing and the term Big Data is a key issue these days. The variety of applications, where huge amounts of data are produced (can be expressed in PBs and more), is great in many areas such as: Biology, Medicine, Astronomy, Geology, Geography, to name just a few. This trend is steadily increasing. Data Mining is the process for extracting useful information from large data-sets. There are different approaches to discovering properties of datasets. Machine Learning is one of them. In Machine Learning, unsupervised learning deals with unlabeled datasets. One of the primary approaches to unsupervised learning is clustering which is the process of grouping similar entities together. Therefore, it is a challenge to improve the performance of such techniques, especially when we are dealing with huge amounts of data. In this work, we present a survey of techniques which increase the efficiency of two well-known clustering algorithms, k-means and DBSCAN.
Research Interests:
The present research work proposes the development of an integrated framework for the personalization and parameterization of learning pathways, aiming at optimizing the quality of the offered services by the Higher Educational... more
The present research work proposes the development of an integrated framework for the personalization and parameterization of learning pathways, aiming at optimizing the quality of the offered services by the Higher Educational Institutions (HEI). In order to achieve this goal, in addition to the educational part, the EDUC8 framework encloses the set of parameters that cover both the technical and the financial dimensions of a learning pathway, thus providing a complete tool for the optimization and calculation of the offered services by the HEIs in combination with the minimization of respective costs. Moreover, the proposed framework incorporates simulation modeling along with machine learning for the purpose of designing learning pathways and evaluating quality assurance indicators and the return on investment of implementation. The study presents a case study in relation to tertiary education in Greece, with a particular focus on Computer Science programs. Data clustering is spe...
Research Interests:
Research Interests:
Page 1. Ent?(atpqotetn/1 "Epguva / Operational Research. An International Journal. Vol.4,, No.3 (2004), pp.291-303 Performance Study of a Dynamic Task Scheduling for Heterogeneous Distributed Systems Ilias K. Savvas I and M-Tahar... more
Page 1. Ent?(atpqotetn/1 "Epguva / Operational Research. An International Journal. Vol.4,, No.3 (2004), pp.291-303 Performance Study of a Dynamic Task Scheduling for Heterogeneous Distributed Systems Ilias K. Savvas I and M-Tahar Kechadi 2 Dept. ...
Research Interests:
ABSTRACT This paper presents a Recurrent Neural Network approach for the multipurpose machines Job Shop Scheduling Problem. This case of JSSP can be utilized for the modelling of project portfolio management besides the well known... more
ABSTRACT This paper presents a Recurrent Neural Network approach for the multipurpose machines Job Shop Scheduling Problem. This case of JSSP can be utilized for the modelling of project portfolio management besides the well known adoption in factory environment. Therefore, each project oriented organization develops a set of projects and it has to schedule them as a whole. In this work, we extended a bi-objective system model based on the JSSP modelling and formulated it as a combination of two recurrent neural networks. In addition, we designed an example within its neural networks that are focused on the Makespan and the Total Weighted Tardiness objectives. Moreover, we present the findings of our approach using a set of well known benchmark instances and the discussion about them and the singularity that arises.
Research Interests:
The last years, huge masses of data are produced or extracted by computational systems and independent electronic devices. To exploit this resource, novel methods must be employed or the established ones may be altered in order to... more
The last years, huge masses of data are produced or extracted by computational systems and independent electronic devices. To exploit this resource, novel methods must be employed or the established ones may be altered in order to confront the issues that arise. One of
the most fruitful techniques, in order to locate and information from data sources is clustering, and k-means is a successful representative algorithm which clusters data according specific characteristics. However, its main disadvantage is the computational complexity which proves the techniques very unproductive to apply on big datasets. Although k-means is a very well studied technique, a fully operational distributed version combining the multi-core power of today machines, have not been accepted yet by the scientific community. In this work, a three phase distributed / multi-core version of k-means is presented. The obtained experimental results are very promising and prove the correctness, the scalability, and the effectiveness
of the proposed technique.
the most fruitful techniques, in order to locate and information from data sources is clustering, and k-means is a successful representative algorithm which clusters data according specific characteristics. However, its main disadvantage is the computational complexity which proves the techniques very unproductive to apply on big datasets. Although k-means is a very well studied technique, a fully operational distributed version combining the multi-core power of today machines, have not been accepted yet by the scientific community. In this work, a three phase distributed / multi-core version of k-means is presented. The obtained experimental results are very promising and prove the correctness, the scalability, and the effectiveness
of the proposed technique.
Research Interests:
ABSTRACT The project portfolio scheduling problem has become very popular in recent years. Current project oriented organisations have to design a plan in order to execute a set of projects sharing common resources such as personnel... more
ABSTRACT The project portfolio scheduling problem has become very popular in recent years. Current project oriented organisations have to design a plan in order to execute a set of projects sharing common resources such as personnel teams. These projects must, therefore, be handled concurrently. This problem can be seen as an extension of the job shop scheduling problem; the multi-purpose job shop scheduling problem. In this paper, we propose a hybrid approach to deal with a biobjective optimisation problem; Makespan and Total Weighted Tardiness. The approach consists of three phases; in the first phase we utilise a Genetic Algorithm (GA) to generate a set of initial solutions, which are used as inputs to recurrent neural networks (RNNs) in the second phase. In the third phase we apply adaptive learning rate and a Tabu Search like algorithm with the view to improve the solutions returned by the RNNs. The proposed hybrid approach is evaluated on some well-known benchmarks and the experimental results are very promising.
ABSTRACT Nowadays, the growth of data is exponential leading to colossal amount of information which is produced by computational systems and electronic instruments such as telescopes, medical devices and so on. To explore this huge... more
ABSTRACT Nowadays, the growth of data is exponential leading to colossal amount of information which is produced by computational systems and electronic instruments such as telescopes, medical devices and so on. To explore this huge amount of data, new fast algorithms must be discovered or old ones may be redesigned. One of the most popular and useful techniques in order to discover and extract information from data pools is clustering, and k-means is an algorithm which clusters data according its characteristics. Its main disadvantage is its computational complexity, which makes the technique very difficult to apply on big data sets. Although k-means is a very well studied technique, a fully parallel version of it has not been explored yet. In this study, a fully parallel version of the k-means for 1-dimensional objects is presented, and in addition, a near-parallel approach for n-dimensional objects is explored. The experimental results obtained for 1- dimensional data are inline with the theoretical outcome and prove both the correctness and the effectiveness of the technique while for n-dimensional objects, are so close with the outcome of the original sequential k-means so these either could be accepted as they are, or could be used as the initial solution for it.
Research Interests:
ABSTRACT This report gives a brief overview of the main concerns addressed by the authors at the first international workshop on Cooperative Knowledge Discovery & Data Mining (CKDD), held at WETICE 2010. A presentation of the main... more
ABSTRACT This report gives a brief overview of the main concerns addressed by the authors at the first international workshop on Cooperative Knowledge Discovery & Data Mining (CKDD), held at WETICE 2010. A presentation of the main topics is given and then a summary of each paper accepted by the workshop is reported.
Research Interests:
Page 1. Synergy of Task Allocation Techniques for Large Computational Grids Ilias K. Savvas Dept. of Computer Science and Telecommunications TEI of Larissa Larissa, Greece. savvas@teilar.gr Abstract The problem of scheduling ...