Self-Organizing Systems: Cellular Automata Problem description:

Self-organization is ubiquitous in nature. Self-organizing systems are highly distributed, composed of large numbers of units that execute tasks with no direct control or guidance from an external supervisor. These systems exhibit complex global behaviors that emerge from the simple local interactions between their constituent units. In order to get a better understanding over the complex processes involved in the development of such systems a bottom-up analysis is needed. An interesting formal model of high interest are the cellular automata, capable to model biological pattern formation. Cellular automata are mathematical models for complex natural systems containing large number of identical components obeying local interactions. These systems can be seen also as discrete dynamical systems with simple construction but complex self-organizing behavior. Much as in the development of natural organisms, the localized construction process is a function of its surrounding structure and the description residing in every unit. The interactions, constructions principles and development into functional structures can me mapped into technical implementations which will be able to accommodate dynamical changes in their interactions with the environment. Furthermore, the cellular space will provide a representation of the environment such that spatio-temporal correlations can be learned and used in subsequent stages of self-creation and adaptation to accommodate changes in the environment. This advanced seminar aims to provide a survey of principles of self-organization in cellular automata, applicable to engineered systems (e.g. mobile robots) in order exhibit self-organizing behavior. Tasks: • Investigate the computational aspects of cellular automata. • Investigate the self-reproducing capabilities of cellular automata. • Investigate implementations of cellular automata in robotics