Available Theses
General topics:
- Mobile Applications ( ToothAgent
, BlueAgent, ANDIAMO
projects)
- Multi-agent Systems and Data Maining (Implicit Culture project)
- XBRL (eXtensible Business Reporting Language) (SMEFIN project)
- P2P Data bases (Hyperion
project)
- Business Intelligence and DataWarehouses (GRAnD
project)
- Software Agents and Multi-agent Systems (see the Implicit Culture
project)
- Agent-oreinted Software Engineering (see the TROPOS project)
- Security and Depandability Patterns (progetto SERENITY)
- Requirements Engineering (see the TROPOS project)
- Security Engineering ( SERENITY
and MOSTRO projects)
- Goal-oriented Requirements Engineering (see the TROPOS project)
Specific topics:
Here you can find theses developed by former students: Theses.
For further information please contact me: paolo.giorgini@dit.unitn.it
Negotiation in
Agent-based Mobile Service Application
Technology has made it clear for everybody; advanced lightweight
devices are ubiquitous and necessary. These devices employ reliable
means of communications such as Bluetooth and recently, Wi-Fi. Long
ago, we watched the entire world moving towards computer-based
application services. Then, the Internet Bubble, which made us observe
services that are extending to include aspects never existed before,
all made remotely through the web. Nowadays, services are taking
different shape, which is mobile-based. Unsurprisingly, novel services
started to emerge on the way to meet new user’s requirements. In
computer Science, Multi-Agent Systems (MAS) have played major role in
delivering services content to lightweight devices. Cooperative
Environments that Distributed Artificial Intelligence (DAI) and Ambient
Intelligence (AmI) scholars are studying is mostly agent-based
applications. Negotiation and coordination among agents differs when
the assembled service is to be delivered to cellular phones. Time,
network traffic and connectivity constraints in mobile environments are
hardly accepting the exiting system negotiation models. We innovatively
focus our research on implementing a dedicated agent’s negotiation
protocol that can be standardized to assist Multi-Agent Systems that
are serving Bluetooth enabled lightweight devices.
Reference:
http://dit.unitn.it/blueagents/
http://dit.unitn.it/blueagents/toothagent/
http://dit.unitn.it/blueagents/andiamo/index.htm
http://dit.unitn.it/~sameh/
Software variability
representation
Description: an important aspect in current software is variability,
namely the capability of software systems to perform their activities
selecting different strategies. Feature models [1] and their extensions
(among which [2][3]) are the widely adopted notation for representing
variability. Recently, a formal semantics [4] was proposed to give a
precise representation of feature diagrams. A complementary approach
proposes the use of goal modeling to represent and analyze high level
(intentional) variability [5]. The objective of the thesis is
todevelop a tool to create, analyze, and formally represent a
particular class of feature diagrams, intended to depict software
configurations. The tool is fundamental step to enable the integration
of self-configuration capabilities in Autonomic Computing [6] systems.
Requirements: Java programming,
basic concepts of logics
Reference:
[1] K. Kang, S. Cohen, J. Hess, W. Novak, S. Peterson. Feature-Oriented
Domain Analysis (FODA) Feasibility Study. Technical Report
CMU/SEI-90-TR-21, Software Engineering Institute, Carnegie Mellon
University, November 1990
[2] K.C. Kang, S. Kim, J. Lee, K. Kim. FORM: A Feature-Oriented Reuse
Method. In Annals of Software Engineering 5, pages 143-168, 1998
[3] M. Riebisch, K. Böllert, D. Streitferdt, I. Philippow.
Extending Feature Diagrams with UML Multiplicities. In Proceedings of
the Sixth Conference on Integrated Design and Process Technology (IDPT
2002), Pasadena, CA, June 2002
[4] P.Y. Schobbens, P. Heymans, J.C. Trigaux. Feature Diagrams: A
Survey and a Formal Semantics. In Proceedings of 14th IEEE
International Requirements Engineering Conference, 2006.
[5] S. Liaskos, L. Jiang, A. Lapouchnian, Y. Wang, Y. Yu, J. do Prado
Leite, J. Mylopoulos. Exploring the Dimensions of Variability: a
Requirements Engineering Perspective. In First Variability Modelling of
Software-intensive Systems (VAMOS) Workshop, Limerick, Ireland, January
2007.
[6] J.O. Kephart, D.M. Chess. The Vision of Autonomic Computing. In
Computer, Vol. 36, No. 1, January 2003.
Algoritmi di
contrattazione per sistemi multi-agente: analisi e sviluppo delle varie
tipologie di asta.
La tesi si colloca nell’ambito del progetto Andiamo. Viene richiesta
un’analisi attenta delle varie tipologie di asta usualmente utilizzate
(first price, second price, etc.) e l’implementazione di algoritmi che
massimizzino la soddisfazione degli utenti che effettuano delle
specifiche richieste al sistema. I vari approcci dovranno essere
testati e confrontati per stabilire quale sia l’approccio migliore per
ogni tipo di situazione.
Requirements: conoscenza Java.
References: http://dit.unitn.it/blueagents/andiamo/
http://www.andiamocarpooling.it/
Progetto “Andiamo
Carpooling”: analisi, sviluppo e test del software per i
terminali mobili
La tesi si colloca nell’ambito del progetto Andiamo. Viene
richiesta la re-implementazione del software esistente per i cellulari
e l’analisi delle performances. La comunicazione via sms esistente,
dovrà essere affiancata dalla comunicazione via Bluetooth e/o
GPRS.
Requirements: Java, J2ME
Reference: http://dit.unitn.it/blueagents/andiamo/
http://www.andiamocarpooling.it/
Learning how to play
board/card games
The aim of the thesis is to evaluate effect of recommendations on the
people learning to play games. Forza4 (ConnectFour) – a board game for
two players is considered and a JADE-based simulator of the game has
been developed already. The players in the simulator can estimate their
situations after each turn (e.g. what is the probability of win), and
therefore, they can measure the usefulness of suggestions (whether the
situation became better or worse).
Your task will be to add the use of suggestions into this simulator,
using a java library called the IC-Service. The suggestions will be
based on past moves performed by players in similar situations. The
further idea is to choose another game (e.g. card came such as Magic
the Gathering, vampire: the eternal struggle, etc.) and to develop
simulator for it, again using suggestions about moves.
Requirements: Java. Knowledge
of JADE and basics of game theory is a plus.
Reference:
A system that
facilitates scientific publications search
Description: The goal of the thesis is to develop a system that helps
to search for a paper on a specific topic. The system architecture
consists of the JADE-based multi-agent platform on the server side and
web-based user interface used on the client side. The main focus will
be on user interface: how to search papers, how to specify topics, how
to display recommendations about papers and record feedback.
Requirements: web programming
(something from this list: php/html/jsp/servlets/…). The knowledge of
java and JADE platform is a plus.
Reference: you can see the
description of the system here: http://dit.unitn.it/~birukou/publications/papers/200605aamas.pdf
Self-configuring
multi-agent systems: implementation and analysis
The goal of the project is to develop an architecture of a multi-agent
platform that is able to self-configure, that is to evolve dynamically
in response to changes in its environment. A runtime reconfiguration
mechanism will be based on AI planning for generating possible system
configurations, and game theory inspired procedure for their
evaluation. Your task is to provide an architecture of such a system,
and implement it for further simulations.
Requirements: Java; some
knowledge of agents and multi-agent systems, AI basics, and Jade/Jadex
would be appreciated; your English should be good enough for reading
and work discussions.
Reference: For the high-level
description of the approach see http://dit.unitn.it/~bryl/bryl-giorgini-06-SOAS.pdf
Tool-supported
requirements analysis: exploring and evaluating alternatives
The goal of the project is to develop a tool to support a requirements
engineer in looking for, evaluating and comparing alternative
requirements configurations. A starting point would be an existing
prototype tool which adopts i*/Tropos modeling notation to represent
requirements graphically, and uses a planner to generate alternative
requirements structures. Your task is to revise the architecture and to
complete the implementation for the further validation.
Requirements: Java; some
knowledge of requirements engineering, AI basics, agents and
multi-agent systems would be appreciated; your English should be good
enough for reading and work discussions.
Reference: For the early
version of the work see http://dit.unitn.it/~bryl/bryl-giorgini-06-CoopIS.pdf
Design of Robust
Coordination Protocols for Open Agent Multi-Agent Systems
Open Multi-Agent Systems are formed by collections of autonomous
software agents that interact with one another in order to
automatically negotiate services, for instance through the Internet.
The openness of such environments requires the development of
mechanisms that help make MAS operate more robustly by detecting and
handling exceptions. This goal requires to learn how we can develop MAS
for open contexts where the constituent agents can come from anywhere,
may be buggy or even malicious, and must run in the dynamic and
potentially failure-prone environments at hand.
The main goal of this thesis is to design a fault tolerant
auction-based coordination protocol suitable for open Multi-Agent
Systems. This requires at least two main tasks:
1) identify all the protocol-specific exceptions that can occur in an
open environment
2) define an exception handling architecture for handling failures.
Possible plan of the work:
1) pick an auction-based agent coordination protocol
2) perform an exception analysis of the protocol, i.e. analyze it to
identify the characterize ways that protocol can fail
3) analyze all the fault tolerant agent-based architectures present in
literature in order to identify an approach suitable for handling
failures or define a new one
4) design the exception handling architecture that can detect and
handle the failures identified in 2) while making minimal assumptions
about the agents
(- implement and test the architecture)
NOTES: Project in collaboration with Mark Klein (MIT Center for
Collective Intelligence)