Informations importantes
- La demande d'agrément de stage ou de mémoire doit être rendue avant la fin du semestre qui précède le stage/mémoire.
- La convention de stage doit être rendue avant le début du stage.
- Le mémoire doit être remis avant la fin du semestre durant lequel il est réalisé.
- La soutenance du mémoire doit avoir lieu avant la fin de la session d'examen du semestre en question.
Sujets de mémoires
Ci-dessous, nous vous proposons une liste de sujets de mémoire/stage. Pour plus de renseignements sur un sujet, adressez vous au professeur de contact.Vous pouvez également prendre contact avec l'un ou l'autre des professeurs de notre Institut des Systèmes d'Information (ISI), afin de convenir d'un autre sujet de mémoire que ceux proposés ci-dessous. Une liste complète des professeurs de l'ISI est disponible ici.
Stages
Switzernethttp://www.switzernet.com/public/080403-student-projects/
Contact: Yves.Pigneur@unil.ch
Intsoft SA
Génération automatique de code pour des applications de gestion en environnement Microsoft avec une architecture logique organisée en 3 tiers. Exploitation de l'outil CodeSmith (générateur de code basé sur des templates). Génération de code dans les couches suivantes :
- Base de données. Génération des procédures stockées.
- Couche d'accès aux données (DAL). Génération de l'ensemble des classes (custom entities).
- Couche métier (partie basse) (BLL). Génération des classes associées aux objets métiers.
- Couche présentation (UIL). Génération systématique des pages "standard" de l'interface utilisateur.
Contact dans l'entreprise: Tony de Candia, directeur
Contact: Yves.Pigneur@unil.ch - Thibault.Estier@unil.ch
Mémoires académiques
Multinational firms networks and development of citiesThis project is a part of a research project ESPON (European Spatial Planning Organization Network) concerning cities (FOCI : Futur of Cities). The aim of the project is to build Urban indexes using their centralities in Multinational firms networks. Network indexes will concern each network, or will be aggregated by sector activity or by cities. The database is constituted by the 1 472 first worldwide Multinational firms networks composed by 700.000 enterprises. The database is in PHP format and includes 2 parts :
1- information database :
- name of each enterprise (700.000)
- Turn-over
- Employees
- address
- Location (city)
- country
- Activity sector: NACE
- financial links between the 700.000 enterprises (subsidiary links)
- share of these links (Shareholders)
The job will consist to think about several indexes in collaboration with the users (geographers) and to build them in TULIP's Plugins. The work will be articulated with a Urban Geography Master Student already operational. The students will be engaged as Student-Assistant for one year (the share has to be discussed). It will begin in January - February 09.
Contact: Marco.Tomassini@unil.ch
Business/IT alignement using strategy map
Location: Royal Institute of Technology (KTH) Stockholm (Prof. Paul Johannesson)
Contact: Yves.Pigneur@unil.ch
Technology foresight Reconciling actor-issue analysis with multi-criterion decision-making analysis
Location: University of Hawaii (Prof. Tung Bui)
Contact: Yves.Pigneur@unil.ch
Improving knowledge management in a networked context
This master thesis aims at improving knowledge management in a networked context where knowledge is represented as an interconnected graph of indexed learning objects.
This objective could concern one of the following issues:
- Identifying constraints related to learning object insertion and removal
- Defining a way to manage metadata
- Proposing a method that guarantee intellectual property
- Implementing insert and remove services
Document retrieval
Document retrieval depends on the way in which the need is expressed. This need varies from simple keyword to a complex training query. This research can be done in a closed as well as an open environment. The more the environment is open, the greater the probability of having documents irrelevant increases.
This master thesis project aims at describing document retrieval process both in a closed and an open environment.
This objective could concern one of the following issues:
- Identifying personalization sides in each type of query
- Defining and testing a architecture allowing to submit a query in an open environment.
Providing a personalized learning path to a leaner in a social context
This master thesis aims at providing a personalized learning path to a leaner in a social context.
This objective could concern one of the following issues:
- Identifying useful user characteristics
- Applying an adaptation strategy to a set of indexed documents.
- Implementing a tool that generate a suitable learning path
- Evaluating the results
Traçabilité des données sur un réseau de partage des connaissances
OpenAdap.net est un projet "open source" de partage de connaissances au sein de communautés interconnectées, constituées par des contributeurs et des consommateurs. L'incorporation de nouveaux membres est dynamique et dans le contexte du réseau des contributeurs la traçabilité des responsabilités est une tâche essentielle et complexe.
L'objectif de ce projet de master est de doter les messages transférés sur le réseau OpenAdap.net des propriétés suivantes:
- authenticité: identification de l'auteur ;
- intégrité: absence de modification par des tiers ;
- non-répudiation: impossibilité de nier une action ;
- confidentialité (optionnellement): maintien du secret.
Contact: Alessandro.Villa@unil.ch
Sécurité du commerce électronique et protection des données personnelles
Signature numérique et infrastructures à clé publique (PKI)
Problématique et éléments de solution de la sécurité des communications « Voix sur IP » (VoIP)
Conformité règlementaire et sécurité de l’information
Sécurité des réseaux sans fil
Etude comparative des méthodologies de gestion de risques
Contact: sgh@unil.ch
Virtual worlds are advancing in their level of realism and are becoming increasingly popular among a broad spectrum of users as internet speeds rise. Second Life, as one exemplar of a virtual world, currently inhabits more than 12 million avatars and generates an economic market exchange that exceeds several million US dollars. As computer-based simulated environments where individuals assume the identity of an avatar, virtual worlds have become an alternative means to the real world for social and economical interactions.
Social Interactions in Virtual Worlds
For any social interaction, the concept of identity plays a crucial role; yet, identities are highly ambiguous in virtual worlds. Consider the simple cues that humans employ in the real world to decipher identity: physical appearance, age, gender, or ethnicity. In virtual worlds, these cues may or may not match reality. In fact, a recent survey of Second Life avatars found that a majority of those sampled do indeed represent themselves inconsistently with respect to their real lives. This raises the question of how much (or how little) real and virtual identities match. Ergo, the focus of this study would be to analyze the discrepancies between real and virtual identities in relation to some dependent variable.
Economical Interactions in Virtual Worlds
Numerous businesses have established a virtual presence in Second Life already, such as Dell, IBM, Toyota, and others. Some of these businesses use their virtual presence for marketing purposes, others for generating revenues, again others as an alternative means to communicate and collaborate within and outside the organizational boundaries. Up to now, however, most businesses do not know how to leverage virtual worlds, and that of Second Life in particular, most effectively and efficiently. Ergo, this study would focus on studying and classifying the various business models and their novelty aspect compared to traditional "real-life" business models.
Contact: Iris.Junglas@unil.ch
Crowd-Sourcing
Crowd-sourcing is a phenomenon that has achieved increasing attention in the recent past. Web sites, such as Facebook, Wiki, or MySpace use contributions of the masses to generate an information pool for other to visit and participate. Individuals supply their own knowledge and personal time to add and edit entries. Even though media has speculated about their authority and credibility, recent studies have shown that entries in Wikipedia, for example, are comparatively accurate. The question remains, however, what motivational factors influence people to contribute their free time in providing content to these sites. Another question would be to study how organizations can leverage this phenomenon for their own advantage.
Contact: Iris.Junglas@unil.ch
| The Calculus of Computation from a Logical Perspective Due to the indefectible reliability required in embarked systems, or simply the ubiquity of computers, implementing correct systems becomes more important every day. Computational logic is a field with many applications as for instance, the verification of software or even hardware systems, as well as artificial intelligence or constraint solving. Modal logics together with first-order logic give rise to steady decision procedures for data structures, combination theories or even arithmetic. Contact: Jaques.Duparc@unil.ch |
| Games, Automata, and Logic Games provide a powerful mathematical framework that captures in a natural way the aspect of interaction. For instance, a software system can be seen as a finite strategy that the "computer" applies while playing a game against the "environment" opponent, where the specifications provide the winning conditions. From this perspective, specifying a software means describing a game, verifying that it corresponds to the specifications means to check that a strategy is winning, and finally synthesizing it means computing a winning strategy. In practice, algorithms are often based on automata which are basic but general models for state-based information processing systems; and they are related to logic. This triptych forms a fast-growing, well-developed area of research. Contact: Jaques.Duparc@unil.ch |
| Designing IT support to risk and compliance management This master thesis aims at improving the information systems management support to Risk and Compliance Management process and/or value chain for financial industry, specifically in an networked context. Compliance Management is the management of all compliance imperatives that impact an organization, including both legal (SOX, Basel II, MiFID …) and strategically self-imposed imperatives (mission statements, codes, standards, guidelines, etc.). This objective could concern one of the following issues:
|
| Repeated Games in Complex Networks The aim of the project is to simulate evolutionary iterated games such as hawks-doves or the prisoner's dilemma in model societies of simple agents. The society will be structured as a network of different types and the conditions under which one observes evolution of cooperative and coordinate behavior will be analyzed. Contact: Marco.Tomassini@unil.ch |
| Adaptive Resources-aware Data Replication The aim of this pro ject is to compare the impact, in terms of performance, of different multicast primitives when applied to the data replication system defined in [3]. The multicast primitives to apply are adaptive to the environment constraints. These constraints could be related to the unreliable behaviour of the environment components (processes and links) and/or to the limited bandwidth. They could also degrade the scalability when processes have a limited memory. Based on tree-like propagation graphs, the multicast primitives to apply use different tree construction techniques depending on the constraint to deal with [2, 1]. Both the tree construction techniques and the data replication are described in previous works. The pro ject consists first; in implementing them using the simulation tool Sinalgo (based on Java). Then, from the resulting measures, compare different impacts and make conclusion. As a starting point, we provide with a Sinalgo implementation of the data replication system defined in [3] that is based on a tree taking the unreliable behaviour of the environment components. You will have to change the implementation to use different tree construction techniques to take into account other constraints. [1] M. Allani, B. Garbinato, F. Pedone, and M. Stamenkovic. Scalable and reliable stream diffusion: A gambling resource-aware approach. In Proceedings of 26th IEEE Symposium on Reliable Distributed Systems (SRDS’2007), pages 288 – 297, Beijing, CHINA, October 2007. [2] B. Garbinato, F. Pedone, and R. Schmidt. An adaptive algorithm for efficient message diffusion in unreliable environments. In Proceedings of IEEE International Conference on Dependable Systems and Networks conference (DSN’2004), June 2004. Accepted for publication. [3] O. Wolfson, S. Ja jodia, and Y. Huang. An adaptive data replication algorithm. ACM Transactions on Database Systems, 22(2):255–314, June 1997. Contact: Benoit.Garbinato@unil.ch Resource-Aware Multicast Algorithm The aim of this pro ject is to extend a solution of the reliable multicast taking into account the environment constraints, e.g., the limited bandwidth, the process and link crash. Indeed, many Internet systems rely on streaming multicast, e.g., Video or Music streaming; consequently, their performance depends on the performance of the underlying streaming mechanism. On the other hand, environment conditions are constrained by the reliability and the capacity (usually limited) of its components. For example, nodes and communication links can fail, unexpectedly ceasing their operation and dropping messages, respectively. Moreover, real-world deployment does not offer infinite bandwidth. Therefore, realistic solutions should use inter-node communication sparingly, and account for node crashes and message losses. This makes the reliable multicast under constrained environment conditions a fundamental problem in distributed computing. In this pro ject, we aim to develop an adaptive solution for the reliable multicast. The adaptive aspect of this solution relies on taking into account the different constraints that could exist in the environment. As a starting point of this pro ject, we provide an initial solution taking into account the unreliable behaviours of processes and links as well as the limited outgoing bandwidth of processes [1]. We ask you to extend this solution by taking into account some additional constraints. For instance, we consider the limited bandwidth related to links and the limited incoming bandwidth at processes, i.e., the maximum number of messages a process can deliver at a time. To do so, we provide a Java program simulating the initial solution and the paper describing it. [1] M. Allani, B. Garbinato, F. Pedone, and M. Stamenkovic. Scalable and reliable stream diffusion: A gambling resource-aware approach. In Proceedings of 26th IEEE Symposium on Reliable Distributed Systems (SRDS’2007), pages 288 – 297, Beijing, CHINA, October 2007. Contact: Benoit.Garbinato@unil.ch |
| Déploiement Agile : méthodologie de déploiement ERP pour PME’s Les éditeurs traditionnels de progiciels de gestion intégrés (en anglais ERP – Enterprise Resource Planning) peinent à adresser efficacement les besoins des PME. Les projets de déploiement d’ERP sont souvent trop lourds à porter pour une PME, trop longs à faire vivre avant le premier retour sur investissement et conduisent le plus souvent à des implantations rigides qui n’accompagnent pas bien l’évolution des modèles d’affaires et la réactivité nécessaire aux PME. L’objectif de se mémoire est d’appliquer les principes du développement agile (itérations courtes et nombreuses, objectifs par paliers, peer-review, strong user involvment) à une méthodologie de déploiement de logiciels ERP. Contexte technologique : ERP Open Source TinyERP (tinyerp.com) et les ressources de la communauté associée. Etapes principales du travail de recherche :
|
| Développement/Intégration d’un éditeur de modèle de processus à un ERP open source L’objectif principal de ce mémoire est de permettre la pilotage par les modèles (Model Driven Configuration) de la parmétrisation de l’ERP TinyERP. Ce sujet est supposé se concentrer sur les modèles de processus qui sont implémentés et servis par l’ERP, dont la représentation actuelle est notamment faite à l’aide de graphes d’états définis dans des fichiers XML particuliers. Cette représentation de niveau implémentation est très difficile à manipuler par les spécialistes « métiers » pour contrôler le comportement de l’ERP (règles de gestion, règles d’ordonnancement, etc.). Le but est d’introduire un éditeur supportant la notation BPMN (Businness Process Management Notation 1.1 – OMG) comme interface utilisateur pour remplacer la manipulation des fichiers XML du moteur de workflows (graphes d’états, etc.). Contexte technologique :
|
