AI & ML
Data-driven control of polynomial systems using implicit methods
arXiv cs.SY0 views
arXiv:2611.24600v1 Announce Type: new Abstract: While conventional behavioral data-driven control represents linear time-invariant (LTI) systems directly from data, extending these data representations to nonlinear systems remains an open challenge. We address this gap for general polynomial systems, characterizing their finite-length trajectories as a real algebraic variety defined by implicit polynomial relations rather than explicit equations. First, we provide conditions on measured data under which these data can be used to represent the system. Building on this representation, we formulate data-driven simulation and predictive control problems that are equivalent to their model-based counterparts, and solve the resulting polynomial systems globally using homotopy continuation. Numerical case studies confirm the effectiveness of this approach by successfully simulating a stiff Van der Pol oscillator in situations where explicit models fail. Furthermore, these data-driven simulation and control methods are illustrated on a Duffing oscillator, demonstrating that our method outperforms Koopman-based approaches in predicting long trajectories.
Source:https://arxiv.org/abs/2611.24600PreviousQuantifying the hosting capacity value of workload flexibility for AI data centers: A two-stage electricity-computation co-optimization frameworkNextData-Driven Factored Abstractions of Stochastic Systems
Related posts
A Matching Lower Bound for Online Smooth Calibration
0arXiv stat.ML
How Path Integral Diffusion Decides a Finite Spin Glass
0arXiv stat.ML
Prescriptive Learning of Linear Policies with Feasibility Guarantees for Binary Decision-Making
0arXiv stat.ML
IQS-BO: In-Context Query Selection for Bayesian Optimisation
0arXiv stat.ML
Swiss-Knife: A Framework for Reconfigurable Externalised Multi-Objective Alignment at Decode Time
0arXiv stat.ML
Inverse Design for Conditional Distribution Matching
0arXiv stat.ML