Marvin: A Heuristic Search Planner with Online Macro-Action Learning

TL;DR

Marvin is a planning system that employs online macro-action learning and action-sequence memoisation to improve search efficiency, demonstrating effectiveness across multiple planning domains and supporting complex language features.

Contribution

The paper introduces Marvin, a novel heuristic search planner that integrates online macro-action learning and supports advanced planning language features.

Findings

Macro-actions improve planning performance

Marvin's search behavior is effective across domains

Supports ADL and Derived Predicates

Abstract

This paper describes Marvin, a planner that competed in the Fourth International Planning Competition (IPC 4). Marvin uses action-sequence-memoisation techniques to generate macro-actions, which are then used during search for a solution plan. We provide an overview of its architecture and search behaviour, detailing the algorithms used. We also empirically demonstrate the effectiveness of its features in various planning domains; in particular, the effects on performance due to the use of macro-actions, the novel features of its search behaviour, and the native support of ADL and Derived Predicates.