Towards Strategy Repair for Adjustable Autonomy

Strategy Repair is the problem of finding a minimal amount of modifications to turn a strategy for a game on graphs from losing into winning. This allows for minimally changing the programmed behavior of an agent, in response to unexpected changes or additional requirements one might need to fulfill at execution time. In this paper, we report on an extension of previous work concerning Strategy Repair for Reachability Games, and discuss the usefulness of Strategy Repair in addressing some issues related to Adjustable Autonomy, i.e., the idea that agents, in order to be autonomous, might also need to stop execution and ask for external (typically human) intervention.