Roadmap for the next generation of dynamic function allocation theories and strategies

Dynamic function allocation theories have matured, but not to the extent that they can be implemented in actual environments. The authors identify a roadmap to help to get to that stage, after a brief review of relevant function allocation theories and the gaps in knowledge in adaptive allocation theories. Past research has focused on short-term implementations of dynamic function allocation, which last for minutes (but never more than a few hours). Stable and enduring implementations of dynamic function allocation, lasting over days or possibly years, require a better and scientific understanding of the interaction among trigger mechanisms, adaptation strategies, and interruption strategies. The possibility of an unintentional consequence of this adaptive allocation, that the automated agent becomes confined to limited levels of automation, has to be ruled out after experimental evidence.