Identifying Desirable Game Character Behaviours through the Application of Evolutionary Algorithms to Model-Driven Engineering Metamodels
This paper describes a novel approach to the derivation of model-driven engineering (MDE) models using metaheuristic search, and illustrates it using a specific engineering problem: that of deriving computer game characters with desirable properties. The character behaviour is defined using a human-readable domain-specific language (DSL) that is interpreted using MDE techniques. We apply the search to the underlying MDE metamodels, rather than the DSL directly, and as a result our approach is applicable to a wide range of MDE models. An implementation developed using the Eclipse Modeling Framework, the most widely-used toolset for MDE, is evaluated. The results demonstrate not only the derivation of characters with the desired properties, but also the identification of unexpected features of the behavioural description language and the game itself.
Unable to display preview. Download preview PDF.