On Design Mining: Coevolution and Surrogate Models

Abstract

Design mining is the use of computational intelligence techniques to iteratively search and model the attribute space of physical objects evaluated directly through rapid prototyping to meet given objectives. It enables the exploitation of novel materials and processes without formal models or complex simulation. In this paper, we focus upon the coevolutionary nature of the design process when it is decomposed into concurrent sub-design threads due to the overall complexity of the task. Using an abstract, tuneable model of coevolution we consider strategies to sample sub-thread designs for whole system testing, how best to construct and use surrogate models within the coevolutionary scenario, and the effects of access to multiple whole system (physical) testing equipment on performance. Drawing on our findings, the paper then describes the effective design of an array of six heterogeneous vertical-axis wind turbines.