Keywords: Metareasoning, Situated Planning, Concurrent Planning and Execution
TL;DR: A formal metareasoning model for concurrent planning and execution. This is a hard problem so greedy algorithms for solving model instances are developed,
Abstract: If a planning agent is considering taking a bus, for example, the time that passes during its planning can affect the feasibility of its plans, as the bus may depart before the agent has found a complete plan. Previous work on this situated temporal planning setting proposed an abstract deliberation scheduling scheme for minimizing the expected cost of a plan that is still feasible at the time it is found. In this paper, introduce the CoPEC model, extending the deliberation scheduling approach to allow for executing actions before the solution plan is fully specified. While allowing additional time to plan when timing is very tight, that entails the risk of performing incorrect actions, a tradeoff formally treated in our model. We develop algorithms to solve CoPEC instances. Our empirical evaluation shows that the greedy scheme performs well in practice on problem instances generated from weighted 15-puzzle instances.