Showing 1–5 of 5 results for author: Glunt, J J

Sharp Hybrid Zonotopes: Set Operations and the Reformulation-linearization Technique

Authors: Jonah J. Glunt, Joshua A. Robbins, Daniel Silvestre, Herschel C. Pangborn

Abstract: Mixed integer set representations, and specifically hybrid zonotopes, have enabled new techniques for reachability and verification of nonlinear and hybrid systems. Mixed-integer sets which have the property that their convex relaxation is equal to their convex hull are said to be sharp. This property allows the convex hull to be computed with minimal overhead, and is known to be important for imp… ▽ More Mixed integer set representations, and specifically hybrid zonotopes, have enabled new techniques for reachability and verification of nonlinear and hybrid systems. Mixed-integer sets which have the property that their convex relaxation is equal to their convex hull are said to be sharp. This property allows the convex hull to be computed with minimal overhead, and is known to be important for improving the convergence rates of mixed-integer optimization algorithms that rely on convex relaxations. This paper examines methods for formulating sharp hybrid zonotopes and provides sharpness-preserving methods for performing several key set operations. The paper then shows how the reformulation-linearization technique can be applied to create a sharp realization of a hybrid zonotope that is initially not sharp. A numerical example applies this technique to find the convex hull of a level set of a feedforward ReLU neural network. △ Less

Submitted 21 March, 2025; originally announced March 2025.

Automated Functional Decomposition for Hybrid Zonotope Over-approximations with Application to LSTM Networks

Authors: Jonah J. Glunt, Jacob A. Siefert, Andrew F. Thompson, Justin Ruths, Herschel C. Pangborn

Abstract: Functional decomposition is a powerful tool for systems analysis because it can reduce a function of arbitrary input dimensions to the sum and superposition of functions of a single variable, thereby mitigating (or potentially avoiding) the exponential scaling often associated with analyses over high-dimensional spaces. This paper presents automated methods for constructing functional decompositio… ▽ More Functional decomposition is a powerful tool for systems analysis because it can reduce a function of arbitrary input dimensions to the sum and superposition of functions of a single variable, thereby mitigating (or potentially avoiding) the exponential scaling often associated with analyses over high-dimensional spaces. This paper presents automated methods for constructing functional decompositions used to form set-based over-approximations of nonlinear functions, with particular focus on the hybrid zonotope set representation. To demonstrate these methods, we construct a hybrid zonotope set that over-approximates the input-output graph of a long short-term memory neural network, and use functional decomposition to represent a discrete hybrid automaton via a hybrid zonotope. △ Less

Submitted 19 March, 2025; originally announced March 2025.

Error Bounds for Compositions of Piecewise Affine Approximations

Authors: Jonah J. Glunt, Jacob A. Siefert, Andrew F. Thompson, Herschel C. Pangborn

Abstract: Nonlinear expressions are often approximated by piecewise affine (PWA) functions to simplify analysis or reduce computational costs. To reduce computational complexity, multivariate functions can be represented as compositions of functions with one or two inputs, which can be approximated individually. This paper provides efficient methods to generate PWA approximations of nonlinear functions via… ▽ More Nonlinear expressions are often approximated by piecewise affine (PWA) functions to simplify analysis or reduce computational costs. To reduce computational complexity, multivariate functions can be represented as compositions of functions with one or two inputs, which can be approximated individually. This paper provides efficient methods to generate PWA approximations of nonlinear functions via functional decomposition. The key contributions focus on intelligent placement of breakpoints for PWA approximations without requiring optimization, and on bounding the error of PWA compositions as a function of the error tolerance for each component of that composition. The proposed methods are used to systematically construct a PWA approximation for a complicated function, either to within a desired error tolerance or to a given level of complexity. △ Less

Submitted 23 February, 2024; originally announced February 2024.

Set-valued State Estimation for Nonlinear Systems Using Hybrid Zonotopes

Authors: Jacob A. Siefert, Andrew F. Thompson, Jonah J. Glunt, Herschel C. Pangborn

Abstract: This paper proposes a method for set-valued state estimation of nonlinear, discrete-time systems. This is achieved by combining graphs of functions representing system dynamics and measurements with the hybrid zonotope set representation that can efficiently represent nonconvex and disjoint sets. Tight over-approximations of complex nonlinear functions are efficiently produced by leveraging specia… ▽ More This paper proposes a method for set-valued state estimation of nonlinear, discrete-time systems. This is achieved by combining graphs of functions representing system dynamics and measurements with the hybrid zonotope set representation that can efficiently represent nonconvex and disjoint sets. Tight over-approximations of complex nonlinear functions are efficiently produced by leveraging special ordered sets and neural networks, which enable computation of set-valued state estimates that grow linearly in memory complexity with time. A numerical example demonstrates significant reduction of conservatism in the set-valued state estimates using the proposed method as compared to an idealized convex approach. △ Less

Submitted 16 September, 2023; v1 submitted 16 April, 2023; originally announced April 2023.

Reachability Analysis Using Hybrid Zonotopes and Functional Decomposition

Authors: Jacob A. Siefert, Trevor J. Bird, Andrew F. Thompson, Jonah J. Glunt, Justin P. Koeln, Neera Jain, Herschel C. Pangborn

Abstract: This paper proposes methods for reachability analysis of nonlinear systems in both open loop and closed loop with advanced controllers. The methods combine hybrid zonotopes, a construct called a state-update set, functional decomposition, and special ordered set approximations to enable linear growth in reachable set memory complexity with time and linear scaling in computational complexity with t… ▽ More This paper proposes methods for reachability analysis of nonlinear systems in both open loop and closed loop with advanced controllers. The methods combine hybrid zonotopes, a construct called a state-update set, functional decomposition, and special ordered set approximations to enable linear growth in reachable set memory complexity with time and linear scaling in computational complexity with the system dimension. Facilitating this combination are new identities for constructing nonconvex sets that contain nonlinear functions and for efficiently converting a collection of polytopes from vertex representation to hybrid zonotope representation. Benchmark numerical examples from the literature demonstrate the proposed methods and provide comparison to state-of-the-art techniques. △ Less

Submitted 22 February, 2024; v1 submitted 13 April, 2023; originally announced April 2023.