Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


Computer Science

Committee Chair

Mikel D. Petty

Committee Member

Letha H. Etzkorn

Committee Member

Daniel M Rochowiak

Committee Member

Harry S Delugach

Committee Member

James J Swain


Reinforcement learning, Materials handling, Discrete-time systems--Simulation methods


Applying machine learning methods to improve the efficiency of complex manufacturing processes, such as material handling, can be challenging. The interconnectedness of the multiple components that make up such processes and the typically large number of variables required to specify procedures and plans within them combine to make it very difficult to map the details of a real-world manufacturing process to an abstract mathematical representation suitable for machine learning methods. This work comprises both applied and theoretical results. On the applied side, machine learning, in particular reinforcement learning, was used to generate increasingly efficient plans for material handling to satisfy temporally varying product demands in a representative manufacturing facility. The essential steps in the research included defining a formal representation of a realistically complex material handling plan, specifying a set of suitable plan change operators as reinforcement learning actions, implementing a simulation-based multi-objective reward function that considers multiple components of material handling costs, and abstracting the many possible material handling plans into a state set small enough to enable reinforcement learning. Extensive experimentation with multiple starting plans showed that reinforcement learning could consistently reduce the material handling plans’ costs over time. This work may be among the first applications of reinforcement learning with a multi-objective reward function to a realistically complex material handling process. On the theoretical side, this work examines the effectiveness of balancing exploration and exploitation in a multi-objective plan-based reinforcement learning problem with different ε values. It also shows that with careful selection of actions and state input, a neural network-based state function approximator can learn action values of a continuous reinforcement learning material handling problem.



To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.