Swarm Intelligence (SI) is a new computational and behavioral paradigm for solving distributed problems based on self-organization. While its main principles are similar to those underlying the behavior of natural systems consisting of many individuals, such as ant colonies and flocks of birds, SI is continuosly incorporating new ideas, algorithms, and principles from the engineering and basic science communities. The student will be able to understand the underlying principles of collective behavior in natural systems through mathematical models and study their extension with engineering knowledge and application to concrete engineering and computer science difficult problems. In particular, the course will cover original combinatorial optimization algorithms, multi-robot coordination strategies, and distributed intelligent control in sensor networks. The course is a well-balanced mixture of theory and laboratory exercises using simulation and real hardware platforms.
The course will involve:
- Introduction to key concepts (e.g., self-organization, stigmergy) and tools (e.g., simulation, robots, sensor nodes).
- Collective movements, foraging, trail-laying and -following, task allocation and division of labor, aggregation and segregation, and self-assembling in natural and artificial societies.
- Multi-level modeling methodologies for collective systems.
- Machine-learning methodologies for automatic design and optimization of collective systems.
- SI-based optimization algorithms (Ant Colony Optimization and Particle Swarm Optimization)
- Applications in robotics, telecommunication, and operational research.
- Selected topics in swarm robotics and sensor networks.