Abstract:

We describe a theory from Herbert Simon that links the structure of complex systems to increased speed of evolution, and argue the position that this theory can be beneficial to evolutionary swarm robotic research. We propose a way of applying this theory to evolutionary swarm robotic systems by manually designing the robot to robot communication mechanisms and keeping these constant, whilst evolving the rest of the robots' behaviours. This allows for robots to evolve independently of each other without breaking any inter-dependencies that may exist between robots in the swarm. Finally we address potential criticisms of our suggested approach, and outline a course of future research in this area in order to verify our proposal.