RWTH Aachen · Bachelor's thesis

Learning to make sequential decisions

Learning decision strategies that account for the consequences of actions over time.

Supervised by Max Faßbender at RWTH Aachen.

The optimization problem

A decision changes both the immediate outcome and the choices available later. Reinforcement learning frames this as learning a policy that maximizes expected cumulative reward over a sequence of decisions.

My work

In my bachelor's thesis, I investigated hierarchical and multi-agent RL: organizing decisions across levels and coordinating multiple agents in simulation. The focus was on learning operating policies with a long-term objective.

The application

Mobile charging robots provided the setting, with decisions about task selection and operation. The broader questions concern policy structure, coordination, and how individual actions contribute to a shared objective.

Related dissertation · RWTH Aachen · 2026

Intelligent management of mobile robot fleets for electric vehicle charging

Max Faßbender

Section 1.4 lists my bachelor's thesis among the supervised student works that contributed to this dissertation.