基于量子近似优化算法（QAOA）的液压约束下天然气输气网络输气量量子优化
Quantum-Based Optimization of Gas Throughput in Natural Gas Transmission Networks Under Hydraulic Constraints Using QAOA
浏览论文内容
中文总结 AI 辅助
本研究针对液压约束下天然气输气网络最大化输气量的组合优化问题，将其适配为QUBO形式并通过QAOA求解，在IonQ Forte-1量子处理器上以仅2层QAOA实现了有效优化，为量子辅助输气网络优化提供了概念验证。
中文摘要 AI 辅助
在液压与运行约束下最大化输气网络的输气量是一个组合问题，其复杂度随网络规模呈指数增长，使得精确求解的计算强度极大。本文通过优化Panhandle-B液压方程下的节点压力分配来解决该基于图的优化问题。将问题构建为对离散化节点压力分配的搜索，结合编码输送目标与物理约束惩罚的代价哈密顿量，建立适用于量子近似优化算法（QAOA）的统一公式。该数学模型被适配为二次无约束二元优化（QUBO）公式，并使用Classiq量子软件平台实现。在基于模拟器的实验中，QAOA恢复了最大输气量的有效运行点，与经典穷举评估及经典液压仿真参考解一致。本研究的一个独特贡献是在IonQ Forte-1阱离子量子处理器上对简化问题实例的端到端执行。值得注意的是，该硬件实现仅使用p=2层QAOA，远少于基于模拟器研究中使用的p=30层。尽管电路深度大幅降低，量子处理单元（QPU）仍产生了物理上有效且可解释的候选解，这些解限定了经典连续最优解的范围，且每个解都位于其一个压力离散化步长内。这些结果表明，使用比最初预期浅得多的QAOA电路即可获得有意义的输气网络优化行为，并为近期量子辅助输气网络优化提供了端到端的概念验证。
英文摘要
Maximizing gas throughput in transmission networks under hydraulic and operational constraints is a combinatorial problem whose complexity grows exponentially with network size, making it computationally intensive to solve exactly. This paper addresses the graph-based optimization problem by optimizing nodal-pressure assignments under the Panhandle-B hydraulic equation. By framing the problem as a search over discretized nodal-pressure assignments coupled with a cost Hamiltonian that encodes both the delivery objective and physical-constraint penalties, we establish a unified formulation suitable for the Quantum Approximate Optimization Algorithm (QAOA). The mathematical model is adapted to a Quadratic Unconstrained Binary Optimization (QUBO) formulation and implemented using the Classiq quantum software platform. In simulator-based experiments, QAOA recovered the maximum-throughput valid operating point, consistent with classical exhaustive evaluation and classical hydraulic simulation reference solutions. A distinctive contribution of this work is the end-to-end execution of a reduced problem instance on the IonQ Forte-1 trapped-ion quantum processor. Remarkably, the hardware implementation used only $p=2$ QAOA layers, substantially fewer than the $p=30$ layers used in the simulator-based study. Despite this significant reduction in circuit depth, the QPU produced physically valid and interpretable candidate solutions that bracketed the continuous classical optimum, with each located within one pressure-discretization step of it. These results demonstrate that meaningful gas-network optimization behavior can be obtained using considerably shallower QAOA circuits than initially expected and provide an end-to-end proof of concept for near-term quantum-assisted gas-network optimization.
发表机构
- INGL – Israel Natural Gas Lines Ltd.（以色列天然气管道有限公司）
- Classiq Technologies（Classiq技术公司）
机构由 AI 辅助整理，请以论文原文为准。
