Computer Science ETDs

Publication Date

Spring 5-16-2026

Abstract

Reinforcement learning (RL) excels at solving complex tasks, but training times can become prohibitively large for challenging motion-planning problems. Methods that address this cost often require additional training or tuning, counteracting the goal of reducing training time. A more effective approach is to exploit inherent task equivalences: many elements of the state space, dynamics, or structure are functionally interchangeable, enabling simplification or knowledge reuse. We present learning solutions that leverage these equivalences to enhance the RL process. First, we leverage the symmetry of homogeneous multi-agent teams to simplify the task to a single strategy. Second, we map correspondences between distinct tasks to reuse model representations and hyperparameters,minimizing tuning overhead. Third, we introduce a novel inference-time enhancement that leverages geometric task symmetry without incurring any additional training cost. Extensive tests for each method demonstrate that they effectively leverage different forms of equivalence to enhance the RL process at different stages.

Language

English

Keywords

Reinforcement Learning, Machine Learning, Motion planning, robotics, artificial intelligence

Document Type

Dissertation

Degree Name

Computer Science

Level of Degree

Doctoral

Department Name

Department of Computer Science

First Committee Member (Chair)

Dr. Lydia Tapia

Second Committee Member

Dr. Evan C. Carter

Third Committee Member

Dr. Xin Chen

Fourth Committee Member

Dr. Rafael Fierro

Download

Share

COinS