This study is about the implementation of a reinforcement learning algorithm in the trajectory planning of manipulators. We have a 7-DOF robotic arm to pick and place the randomly placed block at a random target point in an unknown environment. The obstacle is randomly moving which creates a hurdle in picking the object. The objective of the robot is to avoid the obstacle and pick the block with constraints to a fixed timestamp. In this literature, we have applied a deep deterministic policy gradient (DDPG) algorithm and compared the model's efficiency with dense and sparse rewards.

This study explores a learning-based tri-finger robotic arm manipulating task, which requires complex movements and coordination among the fingers. By employing reinforcement learning, we train an agent to acquire the necessary skills for proficient manipulation. To enhance the efficiency and effectiveness of the learning process, two knowledge transfer strategies, fine-tuning and curriculum learning, were utilized within the soft actor-critic architecture. Fine-tuning allows the agent to leverage pre-trained knowledge and adapt it to new tasks. Several variations like model transfer, policy transfer, and across-task transfer were implemented and evaluated. To eliminate the need for pretraining, curriculum learning decomposes the advanced task into simpler, progressive stages, mirroring how humans learn. The number of learning stages, the context of the sub-tasks, and the transition timing were found to be the critical design parameters. The key factors of two learning strategies and corresponding effects were explored in context-aware and context-unaware scenarios, enabling us to identify the scenarios where the methods demonstrate optimal performance, derive conclusive insights, and contribute to a broader range of learning-based engineering applications.

In this paper, we investigate the impact of introducing relative entropy regularization on the Nash Equilibria (NE) of General-Sum $N$-agent games, revealing the fact that the NE of such games conform to linear Gaussian policies. Moreover, it delineates sufficient conditions, contingent upon the adequacy of entropy regularization, for the uniqueness of the NE within the game. As Policy Optimization serves as a foundational approach for Reinforcement Learning (RL) techniques aimed at finding the NE, in this work we prove the linear convergence of a policy optimization algorithm which (subject to the adequacy of entropy regularization) is capable of provably attaining the NE. Furthermore, in scenarios where the entropy regularization proves insufficient, we present a $\delta$-augmentation technique, which facilitates the achievement of an $\epsilon$-NE within the game.

The control of traffic signals is crucial for improving transportation efficiency. Recently, learning-based methods, especially Deep Reinforcement Learning (DRL), garnered substantial success in the quest for more efficient traffic signal control strategies. However, the design of rewards in DRL highly demands domain knowledge to converge to an effective policy, and the final policy also presents difficulties in terms of explainability. In this work, a new learning-based method for signal control in complex intersections is proposed. In our approach, we design a concept of phase urgency for each signal phase. During signal transitions, the traffic light control strategy selects the next phase to be activated based on the phase urgency. We then proposed to represent the urgency function as an explainable tree structure. The urgency function can calculate the phase urgency for a specific phase based on the current road conditions. Genetic programming is adopted to perform gradient-free optimization of the urgency function. We test our algorithm on multiple public traffic signal control datasets. The experimental results indicate that the tree-shaped urgency function evolved by genetic programming outperforms the baselines, including a state-of-the-art method in the transportation field and a well-known DRL-based method.

As machine learning workloads significantly increase energy consumption, sustainable data centers with low carbon emissions are becoming a top priority for governments and corporations worldwide. This requires a paradigm shift in optimizing power consumption in cooling and IT loads, shifting flexible loads based on the availability of renewable energy in the power grid, and leveraging battery storage from the uninterrupted power supply in data centers, using collaborative agents. The complex association between these optimization strategies and their dependencies on variable external factors like weather and the power grid carbon intensity makes this a hard problem. Currently, a real-time controller to optimize all these goals simultaneously in a dynamic real-world setting is lacking. We propose a Data Center Carbon Footprint Reduction (DC-CFR) multi-agent Reinforcement Learning (MARL) framework that optimizes data centers for the multiple objectives of carbon footprint reduction, energy consumption, and energy cost. The results show that the DC-CFR MARL agents effectively resolved the complex interdependencies in optimizing cooling, load shifting, and energy storage in real-time for various locations under real-world dynamic weather and grid carbon intensity conditions. DC-CFR significantly outperformed the industry standard ASHRAE controller with a considerable reduction in carbon emissions (14.5%), energy usage (14.4%), and energy cost (13.7%) when evaluated over one year across multiple geographical regions.

The current thesis aims to explore the reinforcement learning field and build on existing methods to produce improved ones to tackle the problem of learning in high-dimensional and complex environments. It addresses such goals by decomposing learning tasks in a hierarchical fashion known as Hierarchical Reinforcement Learning. We start in the first chapter by getting familiar with the Markov Decision Process framework and presenting some of its recent techniques that the following chapters use. We then proceed to build our Hierarchical Policy learning as an answer to the limitations of a single primitive policy. The hierarchy is composed of a manager agent at the top and employee agents at the lower level. In the last chapter, which is the core of this thesis, we attempt to learn lower-level elements of the hierarchy independently of the manager level in what is known as the "Eigenoption". Based on the graph structure of the environment, Eigenoptions allow us to build agents that are aware of the geometric and dynamic properties of the environment. Their decision-making has a special property: it is invariant to symmetric transformations of the environment, allowing as a consequence to greatly reduce the complexity of the learning task.

This paper offers a detailed investigation of switchback designs in A/B testing, which alternate between baseline and new policies over time. Our aim is to thoroughly evaluate the effects of these designs on the accuracy of their resulting average treatment effect (ATE) estimators. We propose a novel "weak signal analysis" framework, which substantially simplifies the calculations of the mean squared errors (MSEs) of these ATEs in Markov decision process environments. Our findings suggest that (i) when the majority of reward errors are positively correlated, the switchback design is more efficient than the alternating-day design which switches policies in a daily basis. Additionally, increasing the frequency of policy switches tends to reduce the MSE of the ATE estimator. (ii) When the errors are uncorrelated, however, all these designs become asymptotically equivalent. (iii) In cases where the majority of errors are negative correlated, the alternating-day design becomes the optimal choice. These insights are crucial, offering guidelines for practitioners on designing experiments in A/B testing. Our analysis accommodates a variety of policy value estimators, including model-based estimators, least squares temporal difference learning estimators, and double reinforcement learning estimators, thereby offering a comprehensive understanding of optimal design strategies for policy evaluation in reinforcement learning.

We revisit the problem of offline reinforcement learning with value function realizability but without Bellman completeness. Previous work by Xie and Jiang (2021) and Foster et al. (2022) left open the question whether a bounded concentrability coefficient along with trajectory-based offline data admits a polynomial sample complexity. In this work, we provide a negative answer to this question for the task of offline policy evaluation. In addition to addressing this question, we provide a rather complete picture for offline policy evaluation with only value function realizability. Our primary findings are threefold: 1) The sample complexity of offline policy evaluation is governed by the concentrability coefficient in an aggregated Markov Transition Model jointly determined by the function class and the offline data distribution, rather than that in the original MDP. This unifies and generalizes the ideas of Xie and Jiang (2021) and Foster et al. (2022), 2) The concentrability coefficient in the aggregated Markov Transition Model may grow exponentially with the horizon length, even when the concentrability coefficient in the original MDP is small and the offline data is admissible (i.e., the data distribution equals the occupancy measure of some policy), 3) Under value function realizability, there is a generic reduction that can convert any hard instance with admissible data to a hard instance with trajectory data, implying that trajectory data offers no extra benefits over admissible data. These three pieces jointly resolve the open problem, though each of them could be of independent interest.

Neural network actor-critic algorithms are one of the most popular methods in deep reinforcement learning. A neural network model is trained to select the policy (the "actor") while another neural network (the "critic") is simultaneously trained to learn the value function given the actor's policy. Specifically, the actor selects an action and, given the action, a new state transition occurs according to a Markov stochastic process and a reward (a measurement of the success/failure) is observed. The critic must learn to approximate the value function - the solution to the Bellman equation - given the actor's policy. Given the critic's estimate for the value function of the current policy, the actor must be updated to improve the value function (i.e., increase the expected reward). Actor-critic algorithms are well-established methods in reinforcement learning [17, 15]; the key recent advance is using (deep) neural networks as the actor/critic and training their parameters using gradient descent methods [26, 10, 25, 2, 29]. Analysis of neural network actor-critic algorithms is challenging due to: (1) the non-convexity of the neural networks, (2) the complex feedback loop between the actor and critic (the actor determines the sequence of data samples which are used to train the critic and the critic is used to train the actor), and (3) the simultaneous online updates of both the actor and critic which lead to (3A) the distribution of the data, which depends upon the actor, constantly evolving in time and (3B) the actor being updated with a noisy, biased estimate of the value function.

Accurate risk quantification and reachability analysis are crucial for safe control and learning, but sampling from rare events, risky states, or long-term trajectories can be prohibitively costly. Motivated by this, we study how to estimate the long-term safety probability of maximally safe actions without sufficient coverage of samples from risky states and long-term trajectories. The use of maximal safety probability in control and learning is expected to avoid conservative behaviors due to over-approximation of risk. Here, we first show that long-term safety probability, which is multiplicative in time, can be converted into additive costs and be solved using standard reinforcement learning methods. We then derive this probability as solutions of partial differential equations (PDEs) and propose Physics-Informed Reinforcement Learning (PIRL) algorithm. The proposed method can learn using sparse rewards because the physics constraints help propagate risk information through neighbors. This suggests that, for the purpose of extracting more information for efficient learning, physics constraints can serve as an alternative to reward shaping. The proposed method can also estimate long-term risk using short-term samples and deduce the risk of unsampled states. This feature is in stark contrast with the unconstrained deep RL that demands sufficient data coverage. These merits of the proposed method are demonstrated in numerical simulation.