Some Orders Are Important: Partially Preserving Orders in Top-Quality Planning
Michael Katz, Junkyu Lee, Jungkoo Kang, Shirin Sohrabi
TL;DR
This work introduces partially ordered top-quality planning (P$_X$), a middle ground between fully ordered and unordered top-quality planning by designating an action subset $X$ whose ordering is important. It proposes three computational approaches—post-processing a top-quality planner, extending reduced successors, and modifying stubborn-set-based POR to PO-GSSS—each with theoretical safety guarantees for preserving $P_X$-consistent solutions. The authors prove safety results for both extending reduced successors and PO-GSSS, and validate the approaches experimentally on a large set of benchmarks, showing improved any-time coverage over baseline unordered/top-quality planning. The method enables practical planning with soft or domain-specific order constraints, improving planner efficiency while retaining meaningful control over action sequences in real-world domains.
Abstract
The ability to generate multiple plans is central to using planning in real-life applications. Top-quality planners generate sets of such top-cost plans, allowing flexibility in determining equivalent ones. In terms of the order between actions in a plan, the literature only considers two extremes -- either all orders are important, making each plan unique, or all orders are unimportant, treating two plans differing only in the order of actions as equivalent. To allow flexibility in selecting important orders, we propose specifying a subset of actions the orders between which are important, interpolating between the top-quality and unordered top-quality planning problems. We explore the ways of adapting partial order reduction search pruning techniques to address this new computational problem and present experimental evaluations demonstrating the benefits of exploiting such techniques in this setting.