Merge-and-Shrink Heuristics for Classical Planning: Efficient Implementation and Partial Abstractions

Abstract

Merge-and-shrink heuristics are a successful class of abstraction heuristics used for optimal classical planning. With the recent addition of generalized label reduction, merge-and-shrink can be understood as an algorithm framework that repeatedly applies transformations to a factored representation of a given planning task to compute an abstraction. In this paper, we describe an efficient implementation of the framework and its transformations, comparing it to its previous implementation in Fast Downward. We further discuss partial merge-and-shrink abstractions that do not consider all aspects of the concrete state space. To compute such partial abstractions, we stop the merge-and-shrink computation early by imposing simple limits on the resource consumption of the algorithm. Our evaluation shows that the efficient implementation indeed improves over the previous one, and that partial merge-and-shrink abstractions further push the efficiency of merge-and-shrink planners.

Iterative-Deepening Uniform-Cost Heuristic Search

“…elevators08, transport11, and transport14 are undirected graphs, while the other 8 domains are directed graphs. We tested three different heuristic functions and picked the best one for each domain: the iPDB heuristic with the default configuration in Fast Downward (Haslum et al 2007;Sievers, Ortlieb, and Helmert 2012), the merge-and-shrink heuristic (M&S) with the recommended configuration in Fast Downward Helmert 2014, 2016;Sievers 2018), and the landmark-cut heuristic (LM-cut, Helmert and Domshlak 2009). We used LM-cut for floortile11, M&S for agricola18, barman11, and woodworking08, and iPDB for the rest.…”