Planning for Novelty: Width-Based Algorithms for Common Problems in Control, Planning and Reinforcement Learning

TL;DR

This paper reviews width-based algorithms that leverage state novelty for efficient planning and reinforcement learning, highlighting their theoretical guarantees and potential for cross-disciplinary research.

Contribution

It provides a comprehensive survey of width-based planning algorithms, their theoretical foundations, and discusses future research directions in the field.

Findings

Width-based algorithms achieve state-of-the-art results in classical planning.

They offer polynomial guarantees on runtime and memory based on planning width.

The paper summarizes current research and explores future directions in width-based methods.

Abstract

Width-based algorithms search for solutions through a general definition of state novelty. These algorithms have been shown to result in state-of-the-art performance in classical planning, and have been successfully applied to model-based and model-free settings where the dynamics of the problem are given through simulation engines. Width-based algorithms performance is understood theoretically through the notion of planning width, providing polynomial guarantees on their runtime and memory consumption. To facilitate synergies across research communities, this paper summarizes the area of width-based planning, and surveys current and future research directions.