Quantum-accelerated coverage optimization for weather and climate satellite constellations
Priyank
Dubey
Texas A&M University
Talk
Satellite constellation configuration design is a complex multidisciplinary optimization problem in which coverage serves as the central driver of mission performance and system cost. Coverage is governed primarily by the number, geometry, and orbital characteristics of the satellites comprising the constellation. Among coverage metrics, Maximum Revisit Time (MRT) is particularly critical for weather and climate monitoring missions, as observation gaps directly degrade the fidelity of numerical weather prediction models, atmospheric data assimilation pipelines, and climate data continuity records. Minimizing MRT across globally distributed Earth observation targets requires solving large-scale constrained combinatorial problems that can get computationally intractable with traditional classical methods (MILP) at operational scales demanded by modern multi-orbit constellation architectures. While quantum computing offers a promising pathway to address such combinatorial challenges through its exponentially larger computational space, current NISQ hardware remains fundamentally constrained by limited qubit counts, shallow circuit depth budgets, and susceptibility to gate errors and decoherence, making direct quantum formulation of large-scale constellation problems equally infeasible without a principled dimensionality reduction strategy.
We present a hybrid classical-quantum pipeline that resolves this intractability through a two-stage approach. First, we use the Laplacian Renormalization Group (LRG), a spectral graph method rooted in renormalization group theory, to compress the solution domain into the qubit scale manifolds. The LRG exploits the natural astrodynamic structure of orbital coverage complementarity, achieving variable reduction while preserving the orbital diversity critical for global Earth observation coverage. Second, the compressed problem is then reformulated as a Quadratic Unconstrained Binary Optimization (QUBO) and solved using the Quantum Approximate Optimization Algorithm (QAOA) on IBM quantum hardware. A physics-informed warm-start initialization biased by per-satellite coverage utility reduces the required circuit depth, while marginal probability decoding applied across a large ensemble of hardware shots provides noise-resilient solution recovery despite gate errors and decoherence. Initial experiments across multiple heterogeneous orbital scenarios verify exact recovery of MRT-optimal constellations against classical ground truth; however, we observe that compression quality and solution recovery degrade for spectrally degenerate uniform cases, where the absence of diversity collapses the LRG spectral gap. These findings collectively demonstrate that LRG-compressed quantum optimization is hardware-feasible today for heterogeneous Earth observation architectures, establishing a scalable and principled framework for next-generation weather and climate observing satellite infrastructure.
We present a hybrid classical-quantum pipeline that resolves this intractability through a two-stage approach. First, we use the Laplacian Renormalization Group (LRG), a spectral graph method rooted in renormalization group theory, to compress the solution domain into the qubit scale manifolds. The LRG exploits the natural astrodynamic structure of orbital coverage complementarity, achieving variable reduction while preserving the orbital diversity critical for global Earth observation coverage. Second, the compressed problem is then reformulated as a Quadratic Unconstrained Binary Optimization (QUBO) and solved using the Quantum Approximate Optimization Algorithm (QAOA) on IBM quantum hardware. A physics-informed warm-start initialization biased by per-satellite coverage utility reduces the required circuit depth, while marginal probability decoding applied across a large ensemble of hardware shots provides noise-resilient solution recovery despite gate errors and decoherence. Initial experiments across multiple heterogeneous orbital scenarios verify exact recovery of MRT-optimal constellations against classical ground truth; however, we observe that compression quality and solution recovery degrade for spectrally degenerate uniform cases, where the absence of diversity collapses the LRG spectral gap. These findings collectively demonstrate that LRG-compressed quantum optimization is hardware-feasible today for heterogeneous Earth observation architectures, establishing a scalable and principled framework for next-generation weather and climate observing satellite infrastructure.