arXiv subjects

Matthew Hale

Publications and source records attributed to Matthew Hale.

At least 37 records · Page 2Linked to original sources

arXiv2025

Differentially Private Linear Programming: Reduced Sub-Optimality and Guaranteed Constraint Satisfaction

Linear programming is a fundamental tool in a wide range of decision systems. However, without privacy protections, sharing the solution to a linear program may reveal information about the underlying data used to formulate it, which may be sensitive. Therefore, in this paper we introduce an approach for protecting sensitive data while formulating and solving a linear program. First, we prove that this method perturbs objectives and constraints in a way that makes them differentially private. Then, we show that (i) privatized problems always have solutions, and (ii) their solutions satisfy the constraints in their corresponding original, non-private problems. The latter result solves an open problem in the literature. Next, we analytically bound the expected sub-optimality of solutions that is induced by privacy. Numerical simulations show that, under a typical privacy setup, the solution produced by our method yields a $65\%$ reduction in sub-optimality compared to the state of the art.

math.OC↗

arXiv2025

Asynchronous Nonlinear Sheaf Diffusion for Multi-Agent Coordination

Cellular sheaves and sheaf Laplacians provide a far-reaching generalization of graphs and graph Laplacians, resulting in a wide array of applications ranging from machine learning to multi-agent control. In the context of multi-agent systems, so called coordination sheaves provide a unifying formalism that models heterogeneous agents and coordination goals over undirected communication topologies, and applying sheaf diffusion drives agents to achieve their coordination goals. Existing literature on sheaf diffusion assumes that agents can communicate and compute updates synchronously, which is an unrealistic assumption in many scenarios where communication delays or heterogeneous agents with different compute capabilities cause disagreement among agents. To address these challenges, we introduce asynchronous nonlinear sheaf diffusion. Specifically, we show that under mild assumptions on the coordination sheaf and bounded delays in communication and computation, nonlinear sheaf diffusion converges to a minimizer of the Dirichlet energy of the coordination sheaf at a linear rate proportional to the delay bound. We further show that this linear convergence is attained from arbitrary initial conditions and the analysis depends on the spectrum of the sheaf Laplacian in a manner that generalizes the standard graph Laplacian case. We provide several numerical simulations to validate our theoretical results.

math.OC↗

arXiv2025

Generating Differentially Private Networks with a Modified Erdős-Rényi Model

Differential privacy has been used to privately calculate numerous network properties, but existing approaches often require the development of a new privacy mechanism for each property of interest. Therefore, we present a framework for generating entire networks in a differentially private way. Differential privacy is immune to post-processing, which allows for any network property to be computed and analyzed for a private output network, without weakening its protections. We consider undirected networks and develop a differential privacy mechanism that takes in a sensitive network and outputs a private network by randomizing its edge set. We prove that this mechanism does provide differential privacy to a network's edge set, though it induces a complex distribution over the space of output graphs. We then develop an equivalent privacy implementation using a modified Erdős-Rényi model that constructs an output graph edge by edge, and it is efficient and easily implementable, even on large complex networks. Experiments implement $\varepsilon$-differential privacy with $\varepsilon=2.5$ when computing graph Laplacian spectra, and these results show the proposed mechanism incurs $49.34\%$ less error than the current state of the art.

math.OC↗

arXiv2025

Distributed Asynchronous Time-Varying Quadratic Programming with Asynchronous Objective Sampling

Existing works on multi-agent time-varying optimization allow agents to asynchronously communicate and/or compute, but do not allow asynchronous sampling of objectives. Sampling can be difficult to synchronize, and we therefore present a multi-agent optimization framework that allows asynchrony in sampling, communications, and computations for time-varying quadratic programs. We show that agents have bounded error when tracking the solution to the asynchronously sampled problem, which solves an open problem for quadratic programs. Simulations validate these results.

math.OC↗

arXiv2025

Distributed Multi-agent Coordination over Cellular Sheaves

Techniques for coordination of multi-agent systems are vast and varied, often utilizing purpose-built solvers or controllers with tight coupling to the types of systems involved or the coordination goal. In this paper, we introduce a general unified framework for heterogeneous multi-agent coordination using the language of cellular sheaves and nonlinear sheaf Laplacians, which are generalizations of graphs and graph Laplacians. Specifically, we introduce the concept of a nonlinear homological program encompassing a choice of cellular sheaf on an undirected graph, nonlinear edge potential functions, and constrained convex node objectives, which constitutes a standard form for a wide class of coordination problems. We use the alternating direction method of multipliers to derive a distributed optimization algorithm for solving these nonlinear homological programs. To demonstrate the applicability of this framework, we show how heterogeneous coordination goals including combinations of consensus, formation, and flocking can be formulated as nonlinear homological programs and provide numerical simulations showing the efficacy of our distributed solution algorithm.

math.OC↗

arXiv2025

Time-Constrained Model Predictive Control for Autonomous Satellite Rendezvous, Proximity Operations, and Docking

This paper presents a time-constrained model predictive control strategy for the six degree-of-freedom autonomous rendezvous, proximity, operations and docking problem between a controllable "deputy" satellite and an uncontrolled "chief" satellite. The objective is to achieve a docking configuration defined by both the translational and attitudinal states of the deputy relative to the chief, whose dynamics are respectively governed by both the Clohessy-Wiltshire equations and Euler's second law of motion. The proposed control strategy explicitly addresses computational time constraints that are common to state-of-the-art space vehicles. Thus, a time-constrained model predictive control strategy is implemented on a space-grade processor. Although suboptimal with regards to energy consumption when compared to conventional optimal RPO trajectories, it is empirically demonstrated via numerical simulations that the deputy spacecraft still achieves a successful docking configuration while subject to computational time constraints.

eess.SY↗

arXiv2024

A Dissipativity Approach to Analyzing Composite Spreading Networks

The study of spreading processes often analyzes networks at different resolutions, e.g., at the level of individuals or countries, but it is not always clear how properties at one resolution can carry over to another. Accordingly, in this work we use dissipativity theory from control system analysis to characterize composite spreading networks that are comprised by many interacting subnetworks. We first develop a method to represent spreading networks that have inputs and outputs. Then we define a composition operation for composing multiple spreading networks into a larger composite spreading network. Next, we develop storage and supply rate functions that can be used to demonstrate that spreading dynamics are dissipative. We then derive conditions under which a composite spreading network will converge to a disease-free equilibrium as long as its constituent spreading networks are dissipative with respect to those storage and supply rate functions. To illustrate these results, we use simulations of an influenza outbreak in a primary school, and we show that an outbreak can be prevented by decreasing the average interaction time between any pair of classes to less than 79% of the original interaction time.

physics.soc-ph↗

arXiv2024

Modeling Epidemic Spread: A Gaussian Process Regression Approach

Modeling epidemic spread is critical for informing policy decisions aimed at mitigation. Accordingly, in this work we present a new data-driven method based on Gaussian process regression (GPR) to model epidemic spread through the difference on the logarithmic scale of the infected cases. We bound the variance of the predictions made by GPR, which quantifies the impact of epidemic data on the proposed model. Next, we derive a high-probability error bound on the prediction error in terms of the distance between the training points and a testing point, the posterior variance, and the level of change in the spreading process, and we assess how the characteristics of the epidemic spread and infection data influence this error bound. We present examples that use GPR to model and predict epidemic spread by using real-world infection data gathered in the UK during the COVID-19 epidemic. These examples illustrate that, under typical conditions, the prediction for the next twenty days has 94.29% of the noisy data located within the 95% confidence interval, validating these predictions. We further compare the modeling and prediction results with other methods, such as polynomial regression, k-nearest neighbors (KNN) regression, and neural networks, to demonstrate the benefits of leveraging GPR in disease spread modeling.

stat.ML↗

arXiv2024

Distributed Asynchronous Mixed-Integer Linear Programming with Feasibility Guarantees

In this paper we solve mixed-integer linear programs (MILPs) via distributed asynchronous saddle point computation. This work is motivated by the MILPs being able to model problems in multi-agent autonomy, such as task assignment problems and trajectory planning with collision avoidance constraints in multi-robot systems. To solve a MILP, we relax it with a linear program approximation. We first show that if the linear program relaxation satisfies Slater's condition, then relaxing the problem, solving it, and rounding the relaxed solution produces a point that is guaranteed to satisfy the constraints of the original MILP. Next, we form a Lagrangian saddle point problem that is equivalent to the linear program relaxation, and then we regularize the Lagrangian in both the primal and dual spaces. Doing so gives a regularized Lagrangian that is strongly convex-strongly concave. We then develop a parallelized algorithm to compute saddle points of the regularized Lagrangian, and we show that it is tolerant to asynchrony in the computations and communications of primal and dual variables. Suboptimality bounds and convergence rates are presented for convergence to a saddle point. The suboptimality bound accounts for (i) the error induced by regularizing the Lagrangian and (ii) the suboptimality gap between the solution to the original MILP and the solution to its relaxed form. Simulation results illustrate these theoretical developments in practice, and show that relaxation and regularization combined typically have only a mild impact on the suboptimality of the solution obtained.

math.OC↗

arXiv2024

Guaranteed Feasibility in Differentially Private Linearly Constrained Convex Optimization

Convex programming with linear constraints plays an important role in the operation of a number of everyday systems. However, absent any additional protections, revealing or acting on the solutions to such problems may reveal information about their constraints, which can be sensitive. Therefore, in this paper, we introduce a method for solving convex programs while keeping linear constraints private. First, we prove that this method is differentially private and always generates a feasible optimization problem (i.e., one whose solution exists). Then we show that the solution to the privatized problem also satisfies the original, non-private constraints. Next, we bound the expected loss in performance from privacy, which is measured by comparing the cost with privacy to that without privacy. Simulation results apply this framework to constrained policy synthesis in a Markov decision process, and they show that a typical privacy implementation induces only an approximately $9\%$ loss in solution quality.

math.OC↗

arXiv2024

Differentially Private Reward Functions in Policy Synthesis for Markov Decision Processes

Markov decision processes often seek to maximize a reward function, but onlookers may infer reward functions by observing the states and actions of such systems, revealing sensitive information. Therefore, in this paper we introduce and compare two methods for privatizing reward functions in policy synthesis for multi-agent Markov decision processes, which generalize Markov decision processes. Reward functions are privatized using differential privacy, a statistical framework for protecting sensitive data. The methods we develop perturb either (1) each agent's individual reward function or (2) the joint reward function shared by all agents. We show that approach (1) provides better performance. We then develop a polynomial-time algorithm for the numerical computation of the performance loss due to privacy on a case-by-case basis. Next, using approach (1), we develop guidelines for selecting reward function values to preserve ``goal" and ``avoid" states while still remaining private, and we quantify the increase in computational complexity needed to compute policies from privatized rewards. Numerical simulations are performed on three classes of systems and they reveal a surprising compatibility with privacy: using reasonably strong privacy ($ε=1.3$) on average induces as little as a~$5\%$ decrease in total accumulated reward and a $0.016\%$ increase in computation time.

eess.SY↗

arXiv2024

Technical Report: A Totally Asynchronous Nesterov's Accelerated Gradient Method for Convex Optimization

We present a totally asynchronous algorithm for convex optimization that is based on a novel generalization of Nesterov's accelerated gradient method. This algorithm is developed for fast convergence under "total asynchrony," i.e., allowing arbitrarily long delays between agents' computations and communications without assuming any form of delay bound. These conditions may arise, for example, due to jamming by adversaries. Our framework is block-based, in the sense that each agent is only responsible for computing updates to (and communicating the values of) a small subset of the network-level decision variables. In our main result, we present bounds on the algorithm's parameters that guarantee linear convergence to an optimizer. Then, we quantify the relationship between (i) the total number of computations and communications executed by the agents and (ii) the agents' collective distance to an optimum. Numerical simulations show that this algorithm requires 28% fewer iterations than the heavy ball algorithm and 61% fewer iterations than gradient descent under total asynchrony.

math.OC↗

arXiv2024

Modeling Model Predictive Control: A Category Theoretic Framework for Multistage Control Problems

Model predictive control (MPC) is an optimal control technique which involves solving a sequence of constrained optimization problems across a given time horizon. In this paper, we introduce a category theoretic framework for constructing complex MPC problem formulations by composing subproblems. Specifically, we construct a monoidal category - called Para(Conv) - whose objects are Euclidean spaces and whose morphisms represent constrained convex optimization problems. We then show that the multistage structure of typical MPC problems arises from sequential composition in Para(Conv), while parallel composition can be used to model constraints across multiple stages of the prediction horizon. This framework comes equipped with a rigorous, diagrammatic syntax, allowing for easy visualization and modification of complex problems. Finally, we show how this framework allows a simple software realization in the Julia programming language by integrating with existing mathematical programming libraries to provide high-level, graphical abstractions for MPC.

math.OC↗

arXiv2024

A Compositional Framework for First-Order Optimization

Optimization decomposition methods are a fundamental tool to develop distributed solution algorithms for large scale optimization problems arising in fields such as machine learning and optimal control. In this paper, we present an algebraic framework for hierarchically composing optimization problems defined on hypergraphs and automatically generating distributed solution algorithms that respect the given hierarchical structure. The central abstractions of our framework are operads, operad algebras, and algebra morphisms, which formalize notions of syntax, semantics, and structure preserving semantic transformations respectively. These abstractions allow us to formally relate composite optimization problems to the distributed algorithms that solve them. Specifically, we show that certain classes of optimization problems form operad algebras, and a collection of first-order solution methods, namely gradient descent, Uzawa's algorithm (also called gradient ascent-descent), and their subgradient variants, yield algebra morphisms from these problem algebras to algebras of dynamical systems. Primal and dual decomposition methods are then recovered by applying these morphisms to certain classes of composite problems. Using this framework, we also derive a novel sufficient condition for when a problem defined by compositional data is solvable by a decomposition method. We show that the minimum cost network flow problem satisfies this condition, thereby allowing us to automatically derive a hierarchical dual decomposition algorithm for finding minimum cost flows on composite flow networks. We implement our operads, algebras, and algebra morphisms in a Julia package called AlgebraicOptimization.jl and use our implementation to empirically demonstrate that hierarchical dual decomposition outperforms standard dual decomposition on classes of flow networks with hierarchical structure.

math.OC↗

arXiv2023

Distributed Asynchronous Discrete-Time Feedback Optimization

In this article, we present an algorithm that drives the outputs of a network of agents to jointly track the solutions of time-varying optimization problems in a way that is robust to asynchrony in the agents' operations. We consider three operations that can be asynchronous: (1) computations of control inputs, (2) measurements of network outputs, and (3) communications of agents' inputs and outputs. We first show that our algorithm converges to the solution of a time-invariant feedback optimization problem in linear time. Next, we show that our algorithm drives outputs to track the solution of time-varying feedback optimization problems within a bounded error dependent upon the movement of the minimizers and degree of asynchrony in a way that we make precise. These convergence results are extended to quantify agents' asymptotic behavior as the length of their time horizon approaches infinity. Then, to ensure satisfactory network performance, we specify the timing of agents' operations relative to changes in the objective function that ensure a desired error bound. Numerical experiments confirm these developments and show the success of our distributed feedback optimization algorithm under asynchrony.

math.OC↗

arXiv2023

Differentially Private Computation of Basic Reproduction Numbers in Networked Epidemic Models

The basic reproduction number of a networked epidemic model, denoted $R_0$, can be computed from a network's topology to quantify epidemic spread. However, disclosure of $R_0$ risks revealing sensitive information about the underlying network, such as an individual's relationships within a social network. Therefore, we propose a framework to compute and release $R_0$ in a differentially private way. First, we provide a new result that shows how $R_0$ can be used to bound the level of penetration of an epidemic within a single community as a motivation for the need of privacy, which may also be of independent interest. We next develop a privacy mechanism to formally safeguard the edge weights in the underlying network when computing $R_0$. Then we formalize tradeoffs between the level of privacy and the accuracy of values of the privatized $R_0$. To show the utility of the private $R_0$ in practice, we use it to bound this level of penetration under privacy, and concentration bounds on these analyses show they remain accurate with privacy implemented. We apply our results to real travel data gathered during the spread of COVID-19, and we show that, under real-world conditions, we can compute $R_0$ in a differentially private way while incurring errors as low as $7.6\%$ on average.

cs.SI↗

arXiv2023

Privacy-Engineered Value Decomposition Networks for Cooperative Multi-Agent Reinforcement Learning

In cooperative multi-agent reinforcement learning (Co-MARL), a team of agents must jointly optimize the team's long-term rewards to learn a designated task. Optimizing rewards as a team often requires inter-agent communication and data sharing, leading to potential privacy implications. We assume privacy considerations prohibit the agents from sharing their environment interaction data. Accordingly, we propose Privacy-Engineered Value Decomposition Networks (PE-VDN), a Co-MARL algorithm that models multi-agent coordination while provably safeguarding the confidentiality of the agents' environment interaction data. We integrate three privacy-engineering techniques to redesign the data flows of the VDN algorithm, an existing Co-MARL algorithm that consolidates the agents' environment interaction data to train a central controller that models multi-agent coordination, and develop PE-VDN. In the first technique, we design a distributed computation scheme that eliminates Vanilla VDN's dependency on sharing environment interaction data. Then, we utilize a privacy-preserving multi-party computation protocol to guarantee that the data flows of the distributed computation scheme do not pose new privacy risks. Finally, we enforce differential privacy to preempt inference threats against the agents' training data, past environment interactions, when they take actions based on their neural network predictions. We implement PE-VDN in StarCraft Multi-Agent Competition (SMAC) and show that it achieves 80% of Vanilla VDN's win rate while maintaining differential privacy levels that provide meaningful privacy guarantees. The results demonstrate that PE-VDN can safeguard the confidentiality of agents' environment interaction data without sacrificing multi-agent coordination.

cs.MA↗

arXiv2023

DOMINO++: Domain-aware Loss Regularization for Deep Learning Generalizability

Out-of-distribution (OOD) generalization poses a serious challenge for modern deep learning (DL). OOD data consists of test data that is significantly different from the model's training data. DL models that perform well on in-domain test data could struggle on OOD data. Overcoming this discrepancy is essential to the reliable deployment of DL. Proper model calibration decreases the number of spurious connections that are made between model features and class outputs. Hence, calibrated DL can improve OOD generalization by only learning features that are truly indicative of the respective classes. Previous work proposed domain-aware model calibration (DOMINO) to improve DL calibration, but it lacks designs for model generalizability to OOD data. In this work, we propose DOMINO++, a dual-guidance and dynamic domain-aware loss regularization focused on OOD generalizability. DOMINO++ integrates expert-guided and data-guided knowledge in its regularization. Unlike DOMINO which imposed a fixed scaling and regularization rate, DOMINO++ designs a dynamic scaling factor and an adaptive regularization rate. Comprehensive evaluations compare DOMINO++ with DOMINO and the baseline model for head tissue segmentation from magnetic resonance images (MRIs) on OOD data. The OOD data consists of synthetic noisy and rotated datasets, as well as real data using a different MRI scanner from a separate site. DOMINO++'s superior performance demonstrates its potential to improve the trustworthy deployment of DL on real clinical data.

cs.CV↗