Differences

This shows you the differences between two versions of the page.

--- teaching:mfe:ia [2011/03/18 11:55]
bersini
+++ teaching:mfe:ia [2011/03/23 17:06]
mdorigo
@@ Line 20: / Line 20: @@
   * Contact : [[http://code.ulb.ac.be/iridia.people.php?id=1|Hugues Bersini (IRIDIA)]]
-===== > Développer un programme informatique permettant une analyse statistique en vue de  l'évaluation d'un module psychothérapeutique. =====
+===== Développer un programme informatique permettant une analyse statistique en vue de  l'évaluation d'un module psychothérapeutique. =====
 Ce mémoire se fera en collaboration avec l'équipe médicale du centre pour l'anorexie et la boulimie de l'hôpital Erasme. Il consistera en l'analyse informatisée des données récoltées lors d'entretiens avec le patient et sa famille au cours du traitement.
@@ Line 169: / Line 169: @@
 Among the possible generalisations of the problem there is the Vertex Weighted and Edge Weighted Maximum Clique which asks to find the clique of maximum weight. Being generalisations they are also NP-hard. The goal of the project is to devise heuristic algorithms or adapt existing algorithms of the Maximum Clique for weighted version.
+Required skills: good knowledge of C or C++ programming.
+  * Contacts :
+    * [[http://iridia.ulb.ac.be/~stuetzle|Thomas Stuetzle (IRIDIA)]]
+    * [[http://iridia.ulb.ac.be/~fmascia|Franco Mascia (IRIDIA)]]
+===== Stochastic Local Search heuristics for solving NP-complete puzzles. ======
+This project is about single player games (puzzles) and the design of algorithms for tackling hard combinatorial optimisation problems.
+Example puzzles are: [[http://en.wikipedia.org/wiki/Light_Up|Light Up]], [[http://en.wikipedia.org/wiki/Mastermind_(board_game)|Mastermind]], [[http://en.wikipedia.org/wiki/Minesweeper_(video_game)|Minesweeper]], etc.
+The student will learn how to design and implement a Stochastic Local Search algorithm to solve NP-complete puzzles. The student will also learn how to analyse the performaces of the algorithm and perform statistically sound comparisons with the other algorithms available in literature.
 Required skills: good knowledge of C or C++ programming.
@@ Line 199: / Line 215: @@
 */
-===== Self-organized task allocation in swarm robotics =====
+===== Experiments with the e-puck robot and the IRIDIA TAM =====
-Swarm robotics is an innovative branch of collective robotics that aims at designing robot behaviors by taking inspiration from social animals, such as ants and bees. "Task allocation" in such robotic swarms is the problem of "who is doing what job and when?" Obviously, this problem of assigning jobs to a whole swarm robots can be very difficult, especially when using many robots that cannot communicate with each other on a global level.
+At IRIDIA, we are conducting many experiments with the e-puck
+robot and a task abstraction device, the IRIDIA TAM. The topic of the
+master thesis would be integrate the TAM with the e-puck robot and our
+simulation environment, ARGoS. The final goal is to have the TAM tested
+in real-robot experiments.
-The goal of the project is to implement new algorithms for solving this problem on the e-Puck robot and run extensive experiments with this robot in various environments. The project will involve experimentation with about 30 real e-Pucks. The project is tightly connected to the research in swarm robotics carried out at IRIDIA.
+The subject is practical and requires a dedicated student that is able
+to program in C++. A possible candidate should be willing to work with
-Required skills: The candidates should be acquainted with C/C++ programming and have a working knowledge of the English language.
+hardware and real robots. Additionally, the candidate must be very
+motivated and creative. The working language is English.
   * Contacts: [[http://iridia.ulb.ac.be/~mbiro|Mauro Birattari]], Marco Dorigo, Arne Brutschy, Giovanni Pini (IRIDIA)
-===== Studying collaboration between flying robots and ground-based robots =====
+===== Collaboration between flying robots and ground-based robots =====
-In previous studies, it has been shown that multiple ground-based robots can autonomously form various patterns by attaching to each other. These robots used simple rule sets and local communication to form pre-defined or random patterns. In this thesis, the student will study how flying robots can collaborate with ground-based robots to select and control the pattern formation process. The student will implement the results of his study and various other algorithms that would facilitate such a collaboration. In order to gain a sound understanding of the matter, the student will first study and benchmark collaboration techniques used in existing robotic systems including flying and ground-based robots.
+Current research in self-assembling robots mainly focuses on systems composed of identical (i.e., homogeneous) robots. In this thesis, however, we consider a system composed of robots with varying capabilities and different sensors. In particular, we consider a heterogeneous self-assembling system composed of both ground-based robots and flying robots. The ground-based robots can respond to various task contingencies by autonomously connecting to each other and forming collective structures. The flying robots can use their large field of view (from their elevated positions) to assist the ground-based robots in their tasks.
-A possible candidate student must be very motivated, ready to invest extra hours into the thesis, and have a good grasp of C++.  The working  language is English.
+In this thesis, the student will focus on the flying robots in the system. The student will explore how the flying robots can i) run internal simulations on possible connections between the ground-based robots to determine the response structure to a task and ii) apply machine learning techniques to let the flying robot use previous, successful experiences to learn about tasks and their possible response structures.  The results of the study can be tested on real flying and ground-based robots.
+Concrete ideas will be developed together with the student. A candidate student must be very motivated, independent, have a good knowledge of machine learning techniques, and have a good grasp of C++. The working language is English.
   * Contacts: [[http://iridia.ulb.ac.be/~mbiro|Mauro Birattari]], Marco Dorigo, Nithin Mathews (IRIDIA)
-===== Adaptive collective alignment with a swarm of e-puck robots =====
+===== Recruitment strategies for collective decision making in swarm robotics =====
-Flocking is a fascinating behavior that birds are able to achieve without a leader or a common frame of reference. Moreover, in some cases, the group goes in the correct direction even if only a small proportion of the group knows the goal direction. This allows birds to avoid a predator even if only a subset of the flock sees it. We want to study one of the most interesting aspects of this mechanism, that is how a group can align collectively to a common direction and change this direction over time according to some stimuli perceived only by a small minority of individuals.
+Studies of ants and bees have led to different models of collective
+decision making methods in social insects. Swarms of cooperating
+robots also have to find consensus decisions and thus face similar
+problems as social insects. It is an interesting research question if
+the biological models can be applied to create decentralized and
+robust decision making methods for swarms of robots. More precisely,
+we assume that robots are able to estimate their confidence
+about their own decision. Thus, if a group of robots is unsure about a
+decision they shall recruit more robots into the decision process to
+assure a certain quality in the overall decision.
-The goal of this project is to apply a methodology, so far studied only in simulation, to the e-puck robots, in order to tackle the adaptive collective alignment problem. A group of e-pucks has to reach consensus and turn to a common random heading direction, using a common light source as reference point. Furthermore, when an obstacle is perceived by a small minority of the group, consensus should be achieved in order to align to a new direction which allows them to avoid the obstacle.
+The goal of this master thesis project is to study different
+recruitment strategies for decision making in swarms of robots. The
+following application scenario will be implemented. A group
+of robots need to classify an object in order to operate on it.
+Through its sensors the single robots can classify an object with a
+certain accuracy. This opinion can then be shared in a group to reach
+consensus. If the individual robot's opinions differ strongly from the
+one of other
+robots or the robots do not have the necessary skills/sensors they
+might not be able to reach a final decision. In this case they can
+recruit other robots and involve them in the decision making process.
-Required skills: The candidates should be acquainted with C/C++ programming and have a working knowledge of the English language.
+Required skills: the candidates should be acquainted with C/C++
+programming and have a working knowledge of the English language.
+* Contact: [[http://iridia.ulb.ac.be/~mbiro|Mauro Birattari]], Marco Dorigo, Manuele Brambilla, Alexander Scheidler (IRIDIA)
-  * Contact: [[http://iridia.ulb.ac.be/~mbiro|Mauro Birattari]], Marco Dorigo, Eliseo Ferrante, Ali Emre Turgut (IRIDIA)
 ===== Scalable aggregation in swarm robotics without global information or environmental clues =====
@@ Line 238: / Line 280: @@
   * Contact: [[http://iridia.ulb.ac.be/~mbiro|Mauro Birattari]], Marco Dorigo, Eliseo Ferrante, Ali Emre Turgut (IRIDIA)
-===== A comparison of decision-making strategies for adaptive foraging in swarm robotics =====
-Group of social insects are able to efficiently find the (shortest) path to the a food source and even to differentiate between the quality of two food sources. Studies with ants showed that this mechanism is driven by the perception of stimuli from chemical substances like pheromone. Moreover ants are able to collectively modify their choices if there are changes in the environment, that is, if a source becomes better than another. These ideas have been a source of inspiration for several algorithms in swarm robotics which solves a similar problem (retrieval of objects) by using different types of stimuli such as the encounter rate of objects.
-The goal of this project is to perform a study on how to solve a foraging task in which robots have to choose between staying at the nest or go foraging for different energy sources. The optimal strategy might change over time. What happens if all the robots go to the best source? Will these "traffic jams" slow the process? Is it possible to avoid this problem? What if source quality changes over time? The study will be conducted only in simulation and will concern comparing different approaches and different metrics to measure stimuli.
-Required skills: The candidates should be acquainted with C++ programming and have a working knowledge of the English language.
-  * Contact: [[http://iridia.ulb.ac.be/~mbiro|Mauro Birattari]], Marco Dorigo, Eliseo Ferrante, Manuele Brambilla (IRIDIA)
 ===== Kaleidoscope: Creating temporal motion patterns in a swarm of robots =====
@@ Line 258: / Line 291: @@
+===== Automatic fitness function definition in evolutionary robotics =====
+Evolutionary robotics is a fascinating approach to the design of robot controllers that takes inspiration from natural evolution.
+In order to obtain a robot that is able to perform a desired task, the evolutionary robotics approach considers a population of robots that evolves in time. Each robot is characterized by a genotype that defines somehow its behavior. Each robot is evaluated according to a fitness function that measures the ability of the robot to perform the desired task. Robots with a low fitness are eliminated. Robots with a high fitness remain in the population and generate offsprings -- e.g., robots with a similar genotype obtained via mutation and/or cross-over. Through this process, generation by generation, the evolutionary robotics approach is able to obtain robots that present higher and higher fitness and that are therefore able to perform the desired task more and more effectively.
+One of the main open problems in evolutionary robotics is that the definition of an appropriate fitness function is a very complex, labor-intensive, and time-consuming activity that requires the attention of an expert researcher.
+The goal of this master thesis is to devise an automatic method to define a fitness function in order to obtain a robot that is able to perform a desired task. This automatic method will be based on machine learning and metaheuristic algorithms. In particular, it will draw ideas from the fields of reinforcement learning and of on-line adaptation of parameters in optimization algorithms.
+Required skills: The candidates should be acquainted with C/C++ programming and have a working knowledge of the English language.
+* Contact: [[http://iridia.ulb.ac.be/~mbiro|Mauro Birattari]], Marco Dorigo, Vito Trianni (IRIDIA)
 ===== Simulation et optimisation de trafic routier =====