Finding Options that Minimize Planning Time
This addresses the computational challenge of option discovery in planning for AI systems, though it is incremental as it builds on existing option-based planning frameworks.
The paper tackles the problem of selecting an optimal set of options to minimize planning time, showing it is NP-hard even for deterministic tasks and presenting a polynomial-time approximation algorithm with bounded suboptimality, evaluated empirically in grid-based domains like four-rooms.
We formalize the problem of selecting the optimal set of options for planning as that of computing the smallest set of options so that planning converges in less than a given maximum of value-iteration passes. We first show that the problem is NP-hard, even if the task is constrained to be deterministic---the first such complexity result for option discovery. We then present the first polynomial-time boundedly suboptimal approximation algorithm for this setting, and empirically evaluate it against both the optimal options and a representative collection of heuristic approaches in simple grid-based domains including the classic four-rooms problem.