arXiv subjects

Shivam Bajaj

Publications and source records attributed to Shivam Bajaj.

12 recordsLinked to original sources

arXiv2026

Comparison of Non-Deterministic Nonlinear Systems

We characterize a notion of system comparison, termed as $(T_e,γ,δ)$-similarity, for non-deterministic nonlinear systems. Building on a similar notion recently proposed for stable linear systems, the proposed notion characterizes the dissimilarity between the outputs, measured using the $L_2$ norm, of two nonlinear dynamical systems in terms of their inputs and disturbances. By establishing a relationship between $(T_e,γ,δ)$-similarity and differential dissipativity, we establish equivalence between $(T_e,γ,δ)$-similarity of nonlinear systems and the $(T_e,γ,δ)$-similarity of their differential dynamics. We characterize the $(T_e,γ,δ)$-similarity for nonlinear systems as a Linear Matrix Inequality feasibility problem and also provide necessary and sufficient conditions for solving this feasibility problem. We demonstrate the utility of the proposed notion through its use in two applications: (i) robust hierarchical control applied to a planar aircraft and (ii) the improvement (or design) of abstract models applied to the Moore-Greitzer model and an electronic circuit.

eess.SY↗

arXiv2025

Harnessing Data from Clustered LQR Systems: Personalized and Collaborative Policy Optimization

It is known that reinforcement learning (RL) is data-hungry. To improve sample-efficiency of RL, it has been proposed that the learning algorithm utilize data from 'approximately similar' processes. However, since the process models are unknown, identifying which other processes are similar poses a challenge. In this work, we study this problem in the context of the benchmark Linear Quadratic Regulator (LQR) setting. Specifically, we consider a setting with multiple agents, each corresponding to a copy of a linear process to be controlled. The agents' local processes can be partitioned into clusters based on similarities in dynamics and tasks. Combining ideas from sequential elimination and zeroth-order policy optimization, we propose a new algorithm that performs simultaneous clustering and learning to output a personalized policy (controller) for each cluster. Under a suitable notion of cluster separation that captures differences in closed-loop performance across systems, we prove that our approach guarantees correct clustering with high probability. Furthermore, we show that the sub-optimality gap of the policy learned for each cluster scales inversely with the size of the cluster, with no additional bias, unlike in prior works on collaborative learning-based control. Our work is the first to reveal how clustering can be used in data-driven control to learn personalized policies that enjoy statistical gains from collaboration but do not suffer sub-optimality due to inclusion of data from dissimilar processes. From a distributed implementation perspective, our method is attractive as it incurs only a mild logarithmic communication overhead.

cs.LG↗

arXiv2025

Leveraging Offline Data from Similar Systems for Online Linear Quadratic Control

``Sim2real gap", in which the system learned in simulations is not the exact representation of the real system, can lead to loss of stability and performance when controllers learned using data from the simulated system are used on the real system. In this work, we address this challenge in the linear quadratic regulator (LQR) setting. Specifically, we consider an LQR problem for a system with unknown system matrices. Along with the state-action pairs from the system to be controlled, a trajectory of length $S$ of state-action pairs from a different unknown system is available. Our proposed algorithm is constructed upon Thompson sampling and utilizes the mean as well as the uncertainty of the dynamics of the system from which the trajectory of length $S$ is obtained. We establish that the algorithm achieves $\tilde{\mathcal{O}}({f(S,M_δ)\sqrt{T/S}})$ Bayes regret after $T$ time steps, where $M_δ$ characterizes the \emph{dissimilarity} between the two systems and $f(S,M_δ)$ is a function of $S$ and $M_δ$. When $M_δ$ is sufficiently small, the proposed algorithm achieves $\tilde{\mathcal{O}}({\sqrt{T/S}})$ Bayes regret and outperforms a naive strategy which does not utilize the available trajectory.

eess.SY↗

arXiv2025

Online Model Order Reduction of Linear Systems via $(γ,δ)$-Similarity

Model order reduction aims to determine a low-order approximation of high-order models with least possible approximation errors. For application to physical systems, it is crucial that the reduced order model (ROM) is robust to any disturbance that acts on the full order model (FOM) -- in the sense that the output of the ROM remains a good approximation of that of the FOM, even in the presence of such disturbances. In this work, we present a framework for online model order reduction for a class of continuous-time linear systems that ensures this property for any $\mathcal{L}_2$ disturbance. Apart from robustness to disturbances in this sense, the proposed framework also displays other desirable properties for model order reduction: (1) a provable bound on the error defined as the $L_2$ norm of the difference between the output of the ROM and FOM, (2) preservation of stability, (3) compositionality properties and a provable error bound for arbitrary interconnected systems, (4) a provable bound on the output of the FOM when the controller designed for the ROM is used with the FOM, and finally, (5) compatibility with existing approaches such as balanced truncation and moment matching. Property (4) does not require computation of any gap metric and property (5) is beneficial as existing approaches can also be equipped with some of the preceding properties. The theoretical results are corroborated on numerical case studies, including on a building model.

eess.SY↗

arXiv2024

QueryExplorer: An Interactive Query Generation Assistant for Search and Exploration

Formulating effective search queries remains a challenging task, particularly when users lack expertise in a specific domain or are not proficient in the language of the content. Providing example documents of interest might be easier for a user. However, such query-by-example scenarios are prone to concept drift, and the retrieval effectiveness is highly sensitive to the query generation method, without a clear way to incorporate user feedback. To enable exploration and to support Human-In-The-Loop experiments we propose QueryExplorer -- an interactive query generation, reformulation, and retrieval interface with support for HuggingFace generation models and PyTerrier's retrieval pipelines and datasets, and extensive logging of human feedback. To allow users to create and modify effective queries, our demo supports complementary approaches of using LLMs interactively, assisting the user with edits and feedback at multiple stages of the query formulation process. With support for recording fine-grained interactions and user annotations, QueryExplorer can serve as a valuable experimental and research platform for annotation, qualitative evaluation, and conducting Human-in-the-Loop (HITL) experiments for complex search tasks where users struggle to formulate queries.

cs.IR↗

arXiv2024

Shortest Trajectory of a Dubins Vehicle with a Controllable Laser

We formulate a novel planar motion planning problem for a Dubins-Laser system that consists of a Dubins vehicle with an attached controllable laser. The vehicle moves with unit speed and the laser, having a finite range, can rotate in a clockwise or anti-clockwise direction with a bounded angular rate. From an arbitrary initial position and orientation, the objective is to steer the system so that a given static target is within the range of the laser and the laser is oriented at it in minimum time. We characterize multiple properties of the optimal trajectory and establish that the optimal trajectory for the Dubins-laser system is one out of a total of 16 candidates. Finally, we provide numerical insights that illustrate the properties characterized in this work.

eess.SY↗

arXiv2024

Rationality of Learning Algorithms in Repeated Normal-Form Games

Many learning algorithms are known to converge to an equilibrium for specific classes of games if the same learning algorithm is adopted by all agents. However, when the agents are self-interested, a natural question is whether agents have a strong incentive to adopt an alternative learning algorithm that yields them greater individual utility. We capture such incentives as an algorithm's rationality ratio, which is the ratio of the highest payoff an agent can obtain by deviating from a learning algorithm to its payoff from following it. We define a learning algorithm to be $c$-rational if its rationality ratio is at most $c$ irrespective of the game. We first establish that popular learning algorithms such as fictitious play and regret matching are not $c$-rational for any constant $c\geq 1$. We then propose and analyze two algorithms that are provably $1$-rational under mild assumptions, and have the same properties as (a generalized version of) fictitious play and regret matching, respectively, if all agents follow them. Finally, we show that if an assumption of perfect monitoring is not satisfied, there are games for which $c$-rational algorithms do not exist, and illustrate our results with numerical case studies.

cs.GT↗

arXiv2023

Perimeter Defense using a Turret with Finite Range and Service Times

We consider a perimeter defense problem in a planar conical environment comprising a single turret that has a finite range and non-zero service time. The turret seeks to defend a concentric perimeter against $N\geq 2$ intruders. Upon release, each intruder moves radially towards the perimeter with a fixed speed. To capture an intruder, the turret's angle must be aligned with that of the intruder's angle and must spend a specified service time at that orientation. We address offline and online versions of this optimization problem. Specifically, in the offline version, we establish that in general parameter regimes, this problem is equivalent to solving a Travelling Repairperson Problem with Time Windows (TRP-TW). We then identify specific parameter regimes in which there is a polynomial time algorithm that maximizes the number of intruders captured. In the online version, we present a competitive analysis technique in which we establish a fundamental guarantee on the existence of at best $(N-1)$-competitive algorithms. We also design two online algorithms that are provably $1$ and $2$-competitive in specific parameter regimes.

eess.SY↗

arXiv2022

Competitive Perimeter Defense of Conical Environments

We consider a perimeter defense problem in a planar conical environment in which a single vehicle, having a finite capture radius, aims to defend a concentric perimeter from mobile intruders. The intruders are arbitrarily released at the circumference of the environment and they move radially toward the perimeter with fixed speed. We present a competitive analysis approach to this problem by measuring the performance of multiple online algorithms for the vehicle against arbitrary inputs, relative to an optimal offline algorithm that has information about entire input sequence in advance. In particular, we establish two necessary conditions on the parameter space to guarantee (i) finite competitiveness of any algorithm and (ii) a competitive ratio of at least 2 for any algorithm. We then design and analyze three online algorithms and characterize parameter regimes in which they have finite competitive ratios. Specifically, our first two algorithms are provably 1, and 2-competitive, respectively, whereas our third algorithm exhibits different competitive ratios in different regimes of problem parameters. Finally, we provide a numerical plot in the parameter space to reveal additional insights into the relative performance of our algorithms.

cs.DS↗

arXiv2021

Cooperative Evasion by Translating Targets with Variable Speeds

We consider a problem of cooperative evasion between a single pursuer and multiple evaders in which the evaders are constrained to move in the positive Y direction. The evaders are slower than the vehicle and can choose their speeds from a bounded interval. The pursuer aims to intercept all evaders in a given sequence by executing a Manhattan pursuit strategy of moving parallel to the X axis, followed by moving parallel to the Y axis. The aim of the evaders is to cooperatively pick their individual speeds so that the total time to intercept all evaders is maximized. We first obtain conditions under which evaders should cooperate in order to maximize the total time to intercept as opposed to each moving greedily to optimize its own intercept time. Then, we propose and analyze an algorithm that assigns evasive strategies to the evaders in two iterations as opposed to performing an exponential search over the choice of evader speeds. We also characterize a fundamental limit on the total time taken by the pursuer to capture all evaders when the number of evaders is large. Finally, we provide numerical comparisons against random sampling heuristics.

eess.SY↗

arXiv2021

Competitive Perimeter Defense on a Line

We consider a perimeter defense problem in which a single vehicle seeks to defend a compact region from intruders in a one-dimensional environment parameterized by the perimeter size and the intruder-to-vehicle speed ratio. The intruders move inward with fixed speed and direction to reach the perimeter. We provide both positive and negative worst-case performance results over the parameter space using competitive analysis. We first establish fundamental limits by identifying the most difficult parameter combinations that admit no $c$-competitive algorithms for any constant $c\geq 1$ and slightly easier parameter combinations in which every algorithm is at best $2$-competitive. We then design three classes of algorithms and prove they are $1$, $2$, and $4$-competitive, respectively, for increasingly difficult parameter combinations. Finally, we present numerical studies that provide insights into the performance of these algorithms against stochastically generated intruders.

eess.SY↗

arXiv2019

Dynamic Boundary Guarding Against Radially Incoming Targets

We introduce a dynamic vehicle routing problem in which a single vehicle seeks to guard a circular perimeter against radially inward moving targets. Targets are generated uniformly as per a Poisson process in time with a fixed arrival rate on the boundary of a circle with a larger radius and concentric with the perimeter. Upon generation, each target moves radially inward toward the perimeter with a fixed speed. The aim of the vehicle is to maximize the capture fraction, i.e., the fraction of targets intercepted before they enter the perimeter. We first obtain a fundamental upper bound on the capture fraction which is independent of any policy followed by the vehicle. We analyze several policies in the low and high arrival rates of target generation. For low arrival, we propose and analyze a First-Come-First-Served and a Look-Ahead policy based on repeated computation of the path that passes through maximum number of unintercepted targets. For high arrival, we design and analyze a policy based on repeated computation of Euclidean Minimum Hamiltonian path through a fraction of existing targets and show that it is within a constant factor of the optimal. Finally, we provide a numerical study of the performance of the policies in parameter regimes beyond the scope of the analysis.

eess.SY↗