Temporal Planning and Acting in Dynamic Domains

Due to increased demands related to flexible product configurations, frequent order changes, and tight delivery windows, there is a need for flexible production using AI methods. A way of addressing this is the use of temporal planning as it provides the ability to generate plans for complex goals while considering temporal aspects such as deadlines, concurrency, and durations. Drawbacks in applying such methods in dynamic environments are their high and unpredictable planning time as well as the limited ability to react to unexpected delays during the plan execution. In this paper, we address these issues with a temporal planning and execution framework that employs an online goal management strategy to deal with high and variable computational costs, and a plan dispatcher that takes the future temporal feasibility of plans into consideration. The framework has been evaluated using the RoboCup Logistics League simulation framework.