Marginal Utility for Planning in Continuous or Large Discrete Action Spaces

TL;DR

This paper introduces a novel approach for generating candidate actions in sample-based planning by learning a generator optimized for marginal utility, improving planning efficiency in large or continuous action spaces.

Contribution

The paper proposes a new marginal utility objective for training action generators, outperforming traditional methods and domain knowledge in complex planning tasks.

Findings

Generator trained with marginal utility outperforms hand-coded schemes.

Approach effective in both stochastic continuous and large discrete action spaces.

Method improves sample-based planning efficiency.

Abstract

Sample-based planning is a powerful family of algorithms for generating intelligent behavior from a model of the environment. Generating good candidate actions is critical to the success of sample-based planners, particularly in continuous or large action spaces. Typically, candidate action generation exhausts the action space, uses domain knowledge, or more recently, involves learning a stochastic policy to provide such search guidance. In this paper we explore explicitly learning a candidate action generator by optimizing a novel objective, marginal utility. The marginal utility of an action generator measures the increase in value of an action over previously generated actions. We validate our approach in both curling, a challenging stochastic domain with continuous state and action spaces, and a location game with a discrete but large action space. We show that a generator trained with the marginal utility objective outperforms hand-coded schemes built on substantial domain knowledge, trained stochastic policies, and other natural objectives for generating actions for sampled-based planners.