Unifying and Certifying Top-Quality Planning
Michael Katz, Junkyu Lee, Shirin Sohrabi
TL;DR
The paper introduces dominance top-quality planning, a unifying framework that uses a plan-dominance relation $R$ to capture multiple top-quality planning variants (top-quality, unordered, subset, loopless). It shows how existing problems are special cases under specific $R$ and demonstrates that top-quality solutions can be certified by transforming the planning task to forbid the extended solution set or by certifying optimality on transformed tasks, including a novel loopless transformation. The key contributions include (i) a formal definition of dominance top-quality planning, (ii) demonstration that common variants align with particular relations, and (iii) an efficient loopless forbidding transformation that enables certification without exhaustive enumeration. Practically, this work enables robust certification of top-quality planning solutions and provides a path to certify planning algorithms in real-world planning systems.
Abstract
The growing utilization of planning tools in practical scenarios has sparked an interest in generating multiple high-quality plans. Consequently, a range of computational problems under the general umbrella of top-quality planning were introduced over a short time period, each with its own definition. In this work, we show that the existing definitions can be unified into one, based on a dominance relation. The different computational problems, therefore, simply correspond to different dominance relations. Given the unified definition, we can now certify the top-quality of the solutions, leveraging existing certification of unsolvability and optimality. We show that task transformations found in the existing literature can be employed for the efficient certification of various top-quality planning problems and propose a novel transformation to efficiently certify loopless top-quality planning.