Italy

DIMET - Dipartimento di Matematica, Informatica, Elettronica e Trasporti - Universita "Mediterranea" di Reggio Calabria

Membership Details (University/College)

Contact:
Member node: 063
DIMET c/o Facolta di Ingegneria via Graziella, loc. Feo di Vito I-89060 Reggio Calabria ITALY
Italy

>>> Include yourself at this organisation
>>> Update your name or personal details listed below
>>> Submit papers, projects etc to the archive
>>> Update Member Description or Contact details

Description

The research activity performed by the Computer Science Group of the University "Mediterranea" of Reggio Calabria includes many fields. However, Intelligent Agent research plays a prominent role. In particular, the main research streams on:

- Knowledge Representation in Multi-Agent Systems. In a multi-agent environment it is possible to represent agents' requirements or desires as logic programs. In [6, 7] we have introduced the Joint Fixpoint Semantics which can be used in order to model any joint decision reflecting a common agreement among agents by a suitable joint model of the respective logic programs.
Moreover, we have developed a front-end to compromise logic programs (COLPs) representing agents' requirements or desires in a multiagent environment [4]. COLP language enriches classic logic programming with the new predicates "okay" and "okay group" expressing agents' desires and concessions w.r.t. a target joint decision. By ex-ploiting a direct mapping from COLP programs to classic logic programs our approach computes COLPs' joint fixpoints modeling a common agreement among agents. Moreover an option is provided for computing minimal joint fixpoints in order to deal with situations where minimality is an issue to be considered.

-Management of Quality of Service in Telecommunications Networks. [3, 2, 23, 1, 22]. In this context, we have developed an XML-based multi-agent system for facing such a problem. More specifically, we have considered a generic telecommunications network capable of supporting multimedia streaming; each user accessing this network is provided with a user profile describing her/his past interactions with the network and her/his interest towards the various components (e.g. video, audio and data) of multimedia streaming. User profiles are exploited for re-distributing the available network bandwidth in presence of a network fault or a network congestion; the redistribution activity aims to maximize the QoS perceived by the user according to the new bandwidth limitations. We have developed a prototype and we have tested it on a wireless platform.

- Personalized Access to Information Sources. In [29, 30] we propose an approach capable of learning and managing a user profile according to the information sources she/he accessed in the past. Such a profile is thus exploited for supporting the user when she/he would like to visit new information sources. In more detail, each information source is "translated" into a suitable conceptual model called SDR-Network.
Each user profile is learned by means of a suitable integration activity carried out on the information sources the user accessed in the past.

Finally, our approach exploits a complex pattern inferential mechanism derived from Description Logic in order to identify the Web sites which will be the most appealing for her/him. This approach is refined in [8, 9]; here we propose a graph based approach, called c-graph which is obtained by enriching original SDR-Network model by means of suitable coefficients which describe the user behavior when she/he accesses an information source.

- Web Browsing Activities. In [18, 17] we propose X-Compass, an XML-based agent system capable of supporting the user involved in Web browsing activities. X-Compass learns and manages very detailed and rich user profiles. These profiles are exploited for supporting the user when she/he looks for information on the Web. In particular, X-Compass acts as a hybrid Recommender System, i.e. it combines the typical features of content-based as well as collaborative recommender systems. Interestingly enough, X-Compass can run over various kinds of hardware/software platforms.
Moreover, in this context, we have developed an agent system (called DL-WBA) that supports a user in her/his navigation through the Web. DL-WBA associates a profile with each user. Both the user
profile management and its exploitation for supporting user naviga-tion is carried out by means of a particular Description Logic. Our research works belonging to this topic are described in [19].

- E-commerce. In [32, 31] we propose a multi-agent model capable of managing e-commerce activities. In this model, each e-commerce site is associated with an agent which manages the information stored on the site. Moreover, each customer is provided with an agent which
manages her/his profile. The ontologies of involved agents are represented by means of B-SDR network model. The exploitation of B-SDR network makes easy the cooperation between the agents composing our model.
A refinement of the concepts proposed in [32, 31] is discussed in [14]. In [14] we describe an agent clustering technique; the ontologies of involved agents are represented by means of B-SDR network model, which is an adapted version of the c-graph model cited above. More specifically, we propose to hierarchically cluster both the agents associated with e-commerce sites and the agents managing user profiles. In this paper we show how the exploitation of agent hierarchies may effectively support both customers and service/good providers.

- Device-Adaptive E-Commerce. In this application context we have developed an XML-based multi-agent system (EC-XAMAS) capable of supporting a customer in her/his e-commerce activities. In particular, EC-XAMAS exploits both a user profile (describing her/his preferences and behaviors) and a device profile, (describing the characteristics of the device she/he is currently exploiting). This information is exploited by EC-XAMAS for selecting the products/services of the e-commerce site that appear the most appealing, according to the profile of the user, and the most suitable, according to the device she/he is currently using for accessing the site. Our research efforts relative to this argument are described in [27, 28].

- Device-Adaptive E-Learning. In this application context we have developed an XML-based multi-agent system (X-Learn) capable of managing e-learning activities. The philosophy underlying the implementation of X-Learn is strongly influenced by EC-XAMAS. In particular, X-Learn stores both user and device profiles; it exploits this information for carrying out curriculum sequencing activity, i.e. to select the educational resources that appear the most appealing for the user and the most suitable for the device she/he exploits. Our research works belonging to this topic are described in [20, 21].

- Recommender System in On-Line Recruitment Service. In this application context we have developed an XML-based multi-agent system capable of supporting the user when she/he looks for an employment on the Internet. Given a set of job proposals J, the system exploits the profiles of the user (storing her/his preferences and past choices) for identifying a subset J1 of J consisting of the job proposals appearing to be the most appealing for her/him. Our research efforts relative to
this argument are described in [26].

- Coalition Formation in Multi-Agent Systems. An emerging application for software agents is serving as proxies for trade. Agents are not limited to the procurement of traditional goods, such as flights, books and anything sold on the Internet, but they are also used in order to substitute human users within electronic auctions. One method for agents to improve their performance is forming coalitions with other agents. One reason why this might occur is because different agents could have been created by the same owner, so an incentive to cooperate naturally exists. Competing agents can also choose to coordinate their actions when there is a mutually beneficial result. Our investigations w.r.t. this field concern how the joint fixpoint semantics is able to model such a scenario [5].

- Cooperation in Multi-Agent Systems. In [12] we define a semantic properties-based cooperation model for multi-agent systems. Such properties are defined as similarity and attractivity between agents and they are determined by adaptive algorithms. In particular, by solving a linear system containing adaptive parameters a novel technique is defined by which every agent can locate other agents possibly providing a fruitful cooperation. Preliminary and corollary results are included in [13, 10, 11].

- Filtering of Multimedia Information in Telecommunications Networks. In this area, we have developed an XML-based multi-agent system (called XICOMASQ). The system stores and handles both user and
device profiles. Given a multimedia stream S consisting of n information contents, XICOMASQ exploits profiles for identifying a subset S1 of S consisting of the information contents appearing to be the most appealing for the user and that can be displayed on the device she/he is currently exploiting. Our research activities with regard to this topic are described in [24, 25].

- Protein Tertiary Structure Prediction. In [15, 16] we propose an XML-based multi-agent expert system supporting the activities of a scientist with regard to the protein structure prediction. Our system performs the following activities: (i) it chooses the set consisting of the best promising predictors to be applied in the context of protein structure problem; (ii) it integrates the results derived by each predictor in order to obtain a global prediction.

Article

Buccafurri, F. and Caminiti, G. (2004) Implementing joint fixpoint semantics on top of DLV. Proc. of Convegno Italiano di Logica Computazionale (CILC 2004).

Araniti, G. and De Meo, P. and Iera, A. and Ursino, D. (2004) WAQM: managing QoS in wireless networks by means of an XML-based multi-agent system. International Journal on Wireless and Mobile Computing.

Araniti, G. and De Meo, P. and Iera, A. and Ursino, D. (2003) Adaptively control the QoS of multimedia wireless applications through user profiling techniques. IEEE Journal On Selected Areas in Communications, 21 (10). pp. 1546-1556.

Conference or Workshop Item

Araniti, G. and De Meo, P. and Iera, A. and Ursino, D. (2004) Enhancing 3G cellular systems with user profiling techniques for multimedia application QoS adaptation. In: Springer Verlag-Lecture Notes in Computer Science for the International Conference on Telecommunications, Fortaleza, Brasil.

This list was generated on Mon Sep 10 16:38:58 BST 2007.

   

AgentLink is the European Commission's IST-funded Coordination Action for Agent-Based Computing
and is coordinated by the
University of Liverpool and University of Southampton
If you encounter any problems with these pages please contact web@agentlink.org.