To evolve or not to evolve? That is the question: whether ‘tis nobler in the mind to suffer the slings and arrows of grey goo, or to deny evolution to a sea of self-replicators and by prevention control them? We have been developing a physical self-replicating machine concept for deployment on the Moon built from local resources on the Moon. Here, we are concerned with architectural issues - we specifically address the problem of uncontrolled replication. We propose a multitiered approach to prevent this: (i) denial of self-replication through the implementation of centralised mass manufacturing of replicators; (ii) denial of scarce sodium and chlorine from Earth acts as an Earth-controlled kill switch in preventing further replication; (iii) denial of centralised supplies of asteroidal metals (tungsten-nickel-cobalt-selenium) at the lunar south pole acts as a Moon-controlled kill switch; (iv) denial of online learning capacity through fixed neural weights; (v) denial of extended computing resources through the elimination of transmit communications between self-replicators; (vi) denial of evolutionary capacity by implementing error detection and correction (EDAC) coding. Two kill switches and EDAC provide the backbone to our approach that maintain self-replication capability.

Evolving robot morphologies implies the need for lifetime learning so that newborn robots can learn to manipulate their bodies. An individual’s morphology will obviously combine traits of all its parents; it must adapt its own controller to suit its morphology, and cannot rely on the controller of any one parent to perform well without adaptation. This paper investigates the practicability and benefits of Lamarckian evolution in this setting. Implementing lifetime learning by means of on-line evolution, we first establish the suitability of an indirect encoding scheme that combines Compositional Pattern Producing Networks (CPPNs) and Central Pattern Generators (CPGs) as a relevant learner and controller for open-loop gait controllers. We then analyze a Lamarckian set-up and the effect of the parental genetic material on the early convergence to good locomotion performance.