講演情報
[1H03]Gradient-free Optimizations in the Variational Quantum Algorithm for solving the Poisson Equation
*Pei LUO1, Haifeng LANG1, Takeshi SATO1 (1. UTokyo)
キーワード：
Variational Quantum Algorithm、Gradient-free Sequential Optimization、Barren Plateau Problem、Poisson Equation
This research investigates gradient-free optimization methods to improve the Variational Quantum Algorithm (VQA) for solving the Poisson equation. Previous VQAs rely on gradient-based optimizers, which often suffer from the barren plateau problem, especially in large systems or deep circuits. To address this, the study implements analytic gradients, sampling-based simulations, and a gradient-free sequential optimization method. Results show that sequential optimization provides stronger resilience to barren plateaus and achieves better fidelity and lower relative error than BFGS, particularly for larger qubit systems and deeper ansatz layers. The research also compares Hardware-Efficient Ansatz and QAOA, demonstrating improved scalability. This work suggests sequential optimization as an efficient and robust approach for quantum Poisson solvers.