Observational emergence in evolutionary fuzzy robotics

摘要

Emergence is a universal phenomenon that can be seen frequently in many fields such as evolutionary computation, artificial life and complex systems. However, many experimental results have argued that certain behaviours or properties are emergent without relying on formal definitions. In this paper, we analyse the effect of individual rules and that of interacted rules of a Fuzzy Logic Controller (FLC) evolved by a Genetic Algorithm (GA) from the perspective of observational emergence. The analysis shows that the FLC has acquired emergent behaviours through the interactions of the underlying fuzzy rules.