Novelty search was proposed as a means of circumventing deception and providing selective pressure towards novel behaviours to provide a path towards open-ended evolution. Initial implementations relied on neuro-evolution approaches which increased network complexity over time. However, although many studies have reported impressive results, it is still not clear whether the benefits of evolving topologies are outweighed by the overall complexity of the approach. Given that novelty search can also be combined with evolutionary methods that utilise fixed topologies, we undertake a systematic comparison of evolving topologies, using two types of fixed topology networks in conjunction with novelty search on two test-beds. We show that evolving topologies do not systematically help, and discuss the practical consequences of these results and the research perspectives opened up.

In evolutionary robot systems where morphologies and controllers of real robots are simultaneously evolved, it is clear that there is likely to be requirements to refine the inherited controller of a ‘newborn’ robot in order to better align it to its newly generated morphology. This can be accomplished via a learning mechanism applied to each individual robot: for practical reasons, such a mechanism should be both sample and time-efficient. In this paper, We investigate two ways to improve the sample and time efficiency of the well-known learner CMA-ES on navigation tasks. The first approach combines CMA-ES with Novelty Search, and includes an adaptive restart mechanism with increasing population size. The second bootstraps CMA-ES using Bayesian Optimisation, known for its sample efficiency. Results using two robots built with the ARE project's modules and four environments show that novelty reduces the number of samples needed to converge, as does the custom restart mechanism; the latter also has better sample and time efficiency than the hybridised Bayesian/Evolutionary method.