Hybrid classical-quantum optimization algorithms in electromagnetic applications

Lim, Qi Jian

Permalink

https://hdl.handle.net/2142/132520 Copy

Description

Title

Hybrid classical-quantum optimization algorithms in electromagnetic applications

Author(s)

Lim, Qi Jian

Issue Date

2025-11-24

Director of Research (if dissertation) or Advisor (if thesis)

Peng, Zhen

Doctoral Committee Chair(s)

Peng, Zhen

Committee Member(s)

• Jin, Jian-Ming
• Schutt-Aine, Jose
• Soltanaghai, Elahé

Department of Study

Electrical & Computer Eng

Discipline

Electrical & Computer Engr

Degree Granting Institution

University of Illinois Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Reconfigurable Intelligent Surfaces
• Time-Modulated Simulated Bifurcation
• Ising-based optimization
• Antenna arrays and beamforming
• Hybrid classical–quantum optimization

Abstract

Physics-based computing based on Ising models has emerged as a powerful paradigm for tackling large-scale combinatorial optimization problems. This dissertation develops a unified Ising-based framework for electromagnetic design, with a particular focus on reconfigurable intelligent surfaces (RISs) and antenna array synthesis. A broad class of electromagnetic tasks—including discrete-phase RIS configuration, antenna beamforming and null steering for multi-user links and wireless power transfer, and planar or conformal array synthesis under quantized amplitude and phase constraints—is reformulated as Ising Hamiltonians. On the algorithmic side, the work investigates both quantum and classical physics-based Ising solvers. Quantum annealing is implemented on D-Wave quantum processing units (QPUs) and used to solve medium-scale instances, and a hybrid classical–quantum divide-and-conquer scheme is developed that decomposes large dense problems into QPU-sized subproblems, which are solved on the annealer and reconciled by classical post-processing. These studies clarify the strengths of quantum annealing and hybrid workflows, as well as current limitations in qubit count and sparse hardware connectivity for very large electromagnetic problems. Complementing the quantum approaches, the dissertation also considers classical physics-inspired Ising machines, in which the dynamics of nonlinear oscillators are emulated on conventional CPUs and GPUs. Among these, I focus on the simulated bifurcation (SB) algorithm, which searches for low-energy Ising states by driving a network of coupled oscillators through a controlled bifurcation process. My core contribution is a detailed nonlinear-dynamics analysis of SB based on an eigenvalue stability framework, which explains how pump schedules govern mode bifurcation and reveals bottlenecks such as retarded bifurcation. Guided by these insights, I introduce time-modulated simulated bifurcation (TM-SB), in which system parameters that are fixed in the classical formulation are deliberately modulated over time, together with a momentum-aware pumping strategy. This time modulation makes the dynamics much more likely to settle into high-quality minima, reducing trapping in shallow local optima and reducing the number of runs needed to obtain good solutions. Numerical studies demonstrate that TM-SB and related Ising-based solvers offer robust performance and strong scalability across these scenarios, often matching or surpassing the best-known classical baselines. Overall, the results position physics-based, Ising-model computing—and TM-SB in particular—as a practical and scalable pathway toward adaptive, energy-efficient smart radio environments and intelligent electromagnetic systems.

Graduation Semester

2025-12

Type of Resource

Thesis

Handle URL

https://hdl.handle.net/2142/132520

Copyright and License Information

© 2025 Qi Jian Lim

Owning Collections