Abstract:

We present a novel evolutionary computing framework, Wallace, that achieves ease-of-use and genericity, via a domain-specific language, and simultaneously achieves efficiency via meta-programming, as well as supporting parallelism. Wallace also includes a novel multiple representation model of individual development, realised using meta-programming. We describe the Wallace framework, illustrating it with a number of example problems from the literature. We compare the performance of this framework to existing EC frameworks; early results show improvements in both conciseness and speed over popular alternatives. Finally, we discuss the future of EC frameworks, and the ongoing developments to the Wallace framework.