On Learning Action Costs from Input Plans

2024/08/20 by Marianela Morales, Morales, Marianela, Alberto Pozanco +9 · 1 citation
Business, Management and Accounting · #Business Process Modeling and Analysis

paper · pdf · doi:10.48550/arxiv.2408.10889

Abstract

Most of the work on learning action models focus on learning the actions' dynamics from input plans. This allows us to specify the valid plans of a planning task. However, very little work focuses on learning action costs, which in turn allows us to rank the different plans. In this paper we introduce a new problem: that of learning the costs of a set of actions such that a set of input plans are optimal under the resulting planning model. To solve this problem we present LACFIP^k, an algorithm to learn action's costs from unlabeled input plans. We provide theoretical and empirical results showing how LACFIP^k can successfully solve this task.

Cited by

• Planning with Minimal Disruption

Related

• Planning with Dynamically Estimated Action Costs
• Learning Action Models from Disordered and Noisy Plan Traces
• Learning Features and Abstract Actions for Computing Generalized Plans
• Decision-Focused Learning to Predict Action Costs for Planning
• Model-based Reinforcement Learning with Multi-step Plan Value Estimation