Reasoning about Uncertainty in Agent Control

It is paramount for agent-based systems to adapt to the dynamics of open environments. The agents need to adapt their processing to available resources, deadlines, the goal criteria specified by the clients as well their current problem solving context in order to survive in these environments. Design-to-Criteria scheduling is the soft real-time process of custom building a schedule to meet real-time performance goals specified by dynamic client goal criteria (including real-time deadlines). Problem solving tasks are represented using a task model that lays out alternate ways to achieve tasks and subtasks. Recent extensions to this technology has spawned a post-scheduling contingency analysis step that can be employed in deadline critical situations where the added computational cost is worth the expense. This is based on the evaluation of available schedules that can be used to recover from a situation in which partially executed schedules cannot be completed successfully. We describe how this analysis improves the schedule evaluation from the uncertainty perspective, and provide empirical evidence to support our claims.