Abstract

A new line of research concerns the transformation of knowledge about how social insects collectively solve problems into artificial problem-solving techniques, producing a new form of Artificial Intelligence, in which the underlying model of intelligence is the Collective Intelligence of a social insect colony. Ethologists have used modelling to understand animal behaviour, and their recent research in social insect behaviour suggests that models based on self-organization can help explain how complex colony-level behaviour emerges out of interactions among individual insects. Engineers have explored those models beyond the biologically plausible and have provided novel problem-solving techniques. The main objective of this specialized seminar is to invite three leading researchers, who will describe the models and principles developed by biologists, and the corresponding problem-solving techniques developed by engineers, including present and future applications.

Program

Monday 25th February 2002

09:00 - 12:00 Marco Dorigo

An Introduction to Ant Algorithms and Swarm Intelligence

Ant colonies, and more generally social insect societies, are distributed systems that, in spite of the simplicity of their individuals, present a highly structured social organization. As a result of this organization, ant colonies can accomplish astonishingly complex tasks that in some cases far exceed the individual capacities of a single ant. The study of ant colonies behavior and of their self-organizing capacities is interesting for computer scientists because it provides models of distributed organization which are useful to solve difficult optimization and distributed control problems. This is particularly true in application environments in which rapid and autonomous adaptation to environmental changes, as well as robustness to system failures, are important features. In this tutorial I will present some models derived from the observation of real ants and other insect societies, and explain how these models can be used to design multi-agent systems for the solution of problems like distributed and adaptive routing in Internet-like networks, combinatorial optimization, and optimal allocation of resources.

12:15-13:45 Lunch

14:00-17:00 Alcherio Martinoli

Embedded Swarm Intelligence: Towards Methodologies for Engineering Large-Scale Distributed, Mobile Systems

In this seminar, I will give an overview of the research currently carried out worldwide in the field of Embedded Swarm Intelligence. I will focus on three main research thrusts: system design, modeling, and machine-learning optimization of control with particular emphasis on my main problem domain, collective mobile robotics. First, I will introduce the general principles of Swarm Intelligence, why they are appealing from an engineering point of view, and how they can be adapted to the currently available technologies of sensing, acting, computing, and communicating. I will then present a set of case studies, which differ in the collective task performed (collective manipulation, distributed sensing, collective movements) and in the type of the mechatronic platform used. In each of these experiments, I will outline particularly interesting research issues within one of the thrusts mentioned above. After having discussed results, advantages, and limitations of the current methods, I will finally describe the future challenges that will have to be faced in order to deploy large-scale, distributed, mobile, embedded systems for real-world applications.

Tuesday 26th February 2002

09:00 - 12:00 Laurent Keller

De la fourmi à la robotique.

Les fourmis composent prés de 10% de la biomasse animale et jouent un rôle écologique primordial. La raison principale de leur succès est leur structure sociale complexe. Chez de nombreuses espèces seule la reine se reproduit dans la colonie alors que plusieurs millions d'ouvrières sont totalement stériles et s'occupent à d'autres tâches comme l'entretient du nid et le soin au couvain. Une telle division des tâches au sein de la sociét´ est possible gràce au système de communication complexe et performant des fourmis. Des expériences ont montré que le mode de nos connaissances sur les sociétés de fourmis peuvent être utilisées pour programmer de facon efficace des robots et leur permettre de cooperer pour effectuer des tâches diverses et complexes. Ces expériences et leurs implications seront présentées dans cet exposé.

12:15 - 13:45 lunch

14:00 - 17:00 Workshop

Invited Lecturers

Marco Dorigo was born in Milan, Italy, in 1961. He received his doctorate in information and systems electronic engineering in 1992 from Politecnico di Milano, Italy, and the title of Agrégé de l'Enseignement Supérieur in Artificial Intelligence, from the Université Libre de Bruxelles, Belgium, in 1995. Currently, he is a Senior Researcher of the FNRS, the Belgian National Fund for Scientific Research. His main current research interest is in ant algorithms and swarm intelligence, a research field he contributed to create. Other research interests include metaheuristics, autonomous robotics, and reinforcement learning. He is the author of a book on learning robots published by MIT Press and of a book on swarm intelligence published by Oxford University Press.

Alcherio Martinoli has a B.Sc. and a M.Sc. in Electrical Engineering from the Swiss Federal Institute of Technology in Zurich (ETHZ), and a Ph.D. in Computer Science from the Swiss Federal Institute of Technology in Lausanne (EPFL). He spent one year as Research Scientist at the Institute of Biomedical Engineering of the ETHZ and one year at the Institute of Industrial Automation of the Spanish Research Council in Madrid, Spain. His Ph.D. focused on modeling and performance prediction of distributed robotics systems, evolutionary methods for designing distributed control algorithms, and system design of miniature robots and collective-specific mechatronic modules. He is currently research faculty and head of the Collective Robotics Group at the California Institute of Technology, a group which is part of the Center for Neuromorphic Systems Engineering. His research interests include swarm intelligence, collective robotics, distributed control, networks of sensors and actuators, and machine-learning techniques applied to distributed problems.

Laurent Keller was born in 1961 in Switzerland and received a B.Sc and a M.Sc in biology as well as a Ph.D. in Zoology from the university of Lausanne.He worked as a postdoctoral fellow ar universities of Lausanne and Hardward and is now full professor of Evolutionary Ecology and head of the Institute of Ecology in Lausanne. Among his research interests are the evolution of animal societies and the ecological and evolutionary consequences of social life, the study of the role of ecological factors on the evolution of social organization, and the development of models to account for the evolution of sociality and the dynamics of cooperation and conflict within animal societies.

Organization

Contact:

Béat Hirsbrunner, Alessio Gaspar
Institute of Informatics, University of Fribourg
Chemin du Musée 3, CH-1700 Fribourg, Switzerland
Phone: +41 (0)26 300 8464 and +41 (0)26 300 8473, Fax +41 (0)26 300 9731

E-mail:

Béat.Hirsbrunner@unifr.ch, Alessio.Gaspar@unifr.ch

Dates:

- 25 February 2002 (09:00 - 17:00)

- 26 February 2002 (09:00 - 17:00)

Conference Location:

Auditorium
Institute of Informatics, University of Fribourg
Site Pérolles, Chemin du musée 3
CH-1700 Fribourg, Switzerland.
Access plan.

Trains: Timetable of Swiss Federal Railways.

Registration: Electronic registration are encouraged : Deadline 18 February