Several goal-oriented problems in the real-world can be naturally expressed as Stochastic Shortest Path problems (SSPs). However, the computational complexity of solving SSPs makes nding solutions to even moderately sized problems intractable. State-of-the-art SSP solvers are unable to learn generalized solutions or policies that would solve multiple problem instances with different object names and/or quantities. This paper presents an approach for learning Generalized Policy Automata (GPAs): non-deterministic partial policies that can be used to catalyze the solution process. GPAs are learned using relational, feature-based abstractions, which makes them applicable on broad classes of related problems with different object names and quantities. Theoretical analysis of this approach shows that it guarantees completeness and hierarchical optimality. Empirical analysis shows that this approach effectively learns broadly applicable policy knowledge in a few-shot fashion and signicantly outperforms state-of-the-art SSP solvers on test problems whose object counts are far greater than those used during training.

Many high-level multi-agent planning problems, including multi-robot navigation and path planning, can be effectively modeled using deterministic actions and observations. In this work, we focus on such domains and introduce the class of Deterministic Decentralized POMDPs (Det-Dec-POMDPs). This is a subclass of Dec-POMDPs characterized by deterministic transitions and observations conditioned on the state and joint actions. We then propose a practical solver called Iterative Deterministic POMDP Planning (IDPP). This method builds on the classic Joint Equilibrium Search for Policies framework and is specifically optimized to handle large-scale Det-Dec-POMDPs that current Dec-POMDP solvers are unable to address efficiently.

Deterministic partially observable Markov decision processes (DetPOMDPs) often arise in planning problems where the agent is uncertain about its environmental state but can act and observe de-terministically. In this paper, we propose DetM-CVI, an adaptation of the Monte Carlo V alue Iteration (MCVI) algorithm for DetPOMDPs, which builds policies in the form of finite-state controllers (FSCs). DetMCVI solves large problems with a high success rate, outperforming existing baselines for DetPOMDPs. We also verify the performance of the algorithm in a real-world mobile robot forest mapping scenario.

We present an A*-based algorithm to compute policies for finite-horizon Dec-POMDPs. Our goal is to sacrifice optimality in favor of scalability for larger horizons. The main ingredients of our approach are (1) using clustered sliding window memory, (2) pruning the A* search tree, and (3) using novel A* heuristics. Our experiments show competitive performance to the state-of-the-art. Moreover, for multiple benchmarks, we achieve superior performance. In addition, we provide an A* algorithm that finds upper bounds for the optimum, tailored towards problems with long horizons. The main ingredient is a new heuristic that periodically reveals the state, thereby limiting the number of reachable beliefs. Our experiments demonstrate the efficacy and scalability of the approach.

When agents that are independently trained (or designed) to complete their individual tasks are deployed in a shared environment, their joint actions may produce negative side effects (NSEs). As their training does not account for the behavior of other agents or their joint action effects on the environment, the agents have no prior knowledge of the NSEs of their actions. We model the problem of mitigating NSEs in a cooperative multi-agent system as a Lexicographic Decentralized Markov Decision Process with two objectives. The agents must optimize the completion of their assigned tasks while mitigating NSEs. We assume independence of transitions and rewards with respect to the agents' tasks but the joint NSE penalty creates a form of dependence in this setting. To improve scalability, the joint NSE penalty is decomposed into individual penalties for each agent using credit assignment, which facilitates decentralized policy computation. Our results in simulation on three domains demonstrate the effectiveness and scalability of our approach in mitigating NSEs by updating the policies of a subset of agents in the system.

Contract scheduling is a widely studied framework for designing real-time systems with interruptible capabilities. Previous work has showed that a prediction on the interruption time can help improve the performance of contract-based systems, however it has relied on a single prediction that is provided by a deterministic oracle. In this work, we introduce and study more general and realistic learning-augmented settings in which the prediction is in the form of a probability distribution, or it is given as a set of multiple possible interruption times. For both prediction settings, we design and analyze schedules which perform optimally if the prediction is accurate, while simultaneously guaranteeing the best worst-case performance if the prediction is adversarial. We also provide evidence that the resulting system is robust to prediction errors in the distributional setting. Last, we present an experimental evaluation that confirms the theoretical findings, and illustrates the performance improvements that can be attained in practice.

Applications such as robot control and wireless communication require planning under uncertainty. Partially observable Markov decision processes (POMDPs) plan policies for single agents under uncertainty and their decentralized versions (DEC-POMDPs) find a policy for multiple agents. The policy in infinite-horizon POMDP and DEC-POMDP problems has been represented as finite state controllers(FSCs). We introduce a novel class of periodic FSCs, composed of layers connected only to the previous and next layer. Our periodic FSC method finds a deterministic finite-horizon policy and converts it to an initial periodic infinitehorizon policy. This policy is optimized by a new infinite-horizon algorithm to yield deterministic periodic policies, and by a new expectation maximization algorithm to yield stochastic periodic policies. Our method yields better results than earlier planning methods and can compute larger solutions than with regular FSCs.

Mobile robotic agents often suffer from localization uncertainty which grows with time and with the agents' movement. This can hinder their ability to accomplish their task. In some settings, it may be possible to perform assistive actions that reduce uncertainty about a robot's location. For example, in a collaborative multi-robot system, a wheeled robot can request assistance from a drone that can fly to its estimated location and reveal its exact location on the map or accompany it to its intended location. Since assistance may be costly and limited, and may be requested by different members of a team, there is a need for principled ways to support the decision of which assistance to provide to an agent and when, as well as to decide which agent to help within a team. For this purpose, we propose Value of Assistance (VOA) to represent the expected cost reduction that assistance will yield at a given point of execution. We offer ways to compute VOA based on estimations of the robot's future uncertainty, modeled as a Gaussian process. We specify conditions under which our VOA measures are valid and empirically demonstrate the ability of our measures to predict the agent's average cost reduction when receiving assistance in both simulated and real-world robotic settings.

Contract scheduling is a general technique that allows the design of systems with interruptible capabilities, given an algorithm that is not necessarily interruptible. Previous work on this topic has assumed that the interruption is a worst-case deadline that is unknown to the scheduler. In this work, we study new settings in which the scheduler has access to some imperfect prediction in regards to the interruption. In the first setting, which is inspired by recent advances in learning-enhanced algorithms, the prediction describes the time that the interruption occurs. The second setting introduces a new model in which predictions are elicited as responses to a number of binary queries. For both settings, we investigate trade-offs between the robustness (i.e., the worst-case performance of the schedule if the prediction is generated adversarially) and the consistency (i.e., the performance assuming that the prediction is error-free). We also establish results on the performance of the schedules as a function of the prediction error.