Evolving Morphologies and Neural Controllers Based on the Same Underlying Principle: Specific Ligand-Receptor Interactions

摘要: Our research is motivated by the fact that in Nature macroscopic forms are often shaped with high precision (the shape of cornea, bones an articulation etc.), this paper investigates possible use biological mechanisms to evolve shapes and functions for a given task. The concurrent evolution neural controller agent creates new kind problem: As nets not independent body structure (often sensors distributed over body, effectors have their positions), change will decrease overall fitness. This flaw any artificial evolutionary system, unless process robust against changes positions cells. Therefore, proposes system able explore networks independently based on specific receptor-ligand interactions.