A comparative analysis on genome pleiotropy for evolved soft robots
Dries Marzougui, Matthijs Biondina, Francis wyffels · Proceedings of the Genetic and Evolutionary Computation Conference Companion · 2022
Biological evolution shapes the body and brain of living creatures together over time. By contrast, in evolutionary robotics, the co-optimization of these subsystems remains challenging. Conflicting mutations cause dissociation between morphology and control, which leads to premature convergence. Recent works have proposed algorithmic modifications to mitigate the impact of conflicting mutations. However, the importance of genetic design remains underexposed. Current approaches are divided between a single, pleiotropic genetic encoding and two isolated encodings representing morphology and control. This design choice is commonly made ad hoc, causing a lack of consistency for practitioners. To standardize this design, we performed a comparative analysis between these two configurations on a soft robot locomotion task. Additionally, we incorporated two currently unexplored alternatives that drive these configurations to their logical extremes. Our results demonstrate that pleiotropic representations yield superior performance in fitness and robustness towards premature convergence. Moreover, we showcase the importance of shared structure in the pleiotropic representation of robot morphology and control to achieve this performance gain. These findings provide valuable insights into genetic encoding design, which supply practitioners with a theoretical foundation to pursue efficient brain-body co-optimization.