Engineering Commons^™

Decision-Analytic Models Using Reinforcement Learning To Inform Dynamic Sequential Decisions In Public Policy, Seyedeh Nazanin Khatami

Doctoral Dissertations

We developed decision-analytic models specifically suited for long-term sequential decision-making in the context of large-scale dynamic stochastic systems, focusing on public policy investment decisions. We found that while machine learning and artificial intelligence algorithms provide the most suitable frameworks for such analyses, multiple challenges arise in its successful adaptation. We address three specific challenges in two public sectors, public health and climate policy, through the following three essays. In Essay I, we developed a reinforcement learning (RL) model to identify optimal sequence of testing and retention-in-care interventions to inform the national strategic plan “Ending the HIV Epidemic in the US”. …

Adaptive Dynamic Programming With Eligibility Traces And Complexity Reduction Of High-Dimensional Systems, Seaar Jawad Kadhim Al-Dabooni

Doctoral Dissertations

"This dissertation investigates the application of a variety of computational intelligence techniques, particularly clustering and adaptive dynamic programming (ADP) designs especially heuristic dynamic programming (HDP) and dual heuristic programming (DHP). Moreover, a one-step temporal-difference (TD(0)) and n-step TD (TD(λ)) with their gradients are utilized as learning algorithms to train and online-adapt the families of ADP. The dissertation is organized into seven papers. The first paper demonstrates the robustness of model order reduction (MOR) for simulating complex dynamical systems. Agglomerative hierarchical clustering based on performance evaluation is introduced for MOR. This method computes the reduced order denominator of the transfer …

Neuron Clustering For Mitigating Catastrophic Forgetting In Supervised And Reinforcement Learning, Benjamin Frederick Goodrich

Doctoral Dissertations

Neural networks have had many great successes in recent years, particularly with the advent of deep learning and many novel training techniques. One issue that has affected neural networks and prevented them from performing well in more realistic online environments is that of catastrophic forgetting. Catastrophic forgetting affects supervised learning systems when input samples are temporally correlated or are non-stationary. However, most real-world problems are non-stationary in nature, resulting in prolonged periods of time separating inputs drawn from different regions of the input space.

Reinforcement learning represents a worst-case scenario when it comes to precipitating catastrophic forgetting in neural networks. …

Quantum Inspired Algorithms For Learning And Control Of Stochastic Systems, Karthikeyan Rajagopal

Doctoral Dissertations

"Motivated by the limitations of the current reinforcement learning and optimal control techniques, this dissertation proposes quantum theory inspired algorithms for learning and control of both single-agent and multi-agent stochastic systems.

A common problem encountered in traditional reinforcement learning techniques is the exploration-exploitation trade-off. To address the above issue an action selection procedure inspired by a quantum search algorithm called Grover's iteration is developed. This procedure does not require an explicit design parameter to specify the relative frequency of explorative/exploitative actions.

The second part of this dissertation extends the powerful adaptive critic design methodology to solve finite horizon stochastic optimal …

Revenue Management For Make-To-Order And Make-To-Stock Systems, Jiao Wang

Doctoral Dissertations

With the success of Revenue Management (RM) techniques over the past three decades in various segments of the service industry, many manufacturing firms have started exploring innovative RM technologies to improve their profits. This dissertation studies RM for make-to-order (MTO) and make-to-stock (MTS) systems.

We start with a problem faced by a MTO firm that has the ability to reject or accept the order and set prices and lead-times to influence demands. The firm is confronted with the problem to decide, which orders to accept or reject and trade-off the price, lead-time and potential for increased demand against capacity constraints, …