Method for solving nonlinear equation set, and related device
The convergence strategy of nonlinear system of equations is automatically determined through neural network models, which solves the problem of inefficient solution in the existing technology, and realizes a more efficient and accurate iterative solution process.
WO2025066155A1 Publication Date: 2025-04-03HUAWEI CLOUD COMPUTING TECHNOLOGIES CO LTD
7 Cites 0 Cited by
Patent Information
- Application Number
- PCT/CN2024/092193
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2024-03-29
- Filing Date
- 2024-05-10
- Publication Date
- 2025-04-03
AI Technical Summary
Technical Problem
When solving nonlinear equation systems, it is difficult to quickly determine the appropriate convergence strategy, resulting in inefficient efficiency of the iterative process and relying on manual experience, affecting processing efficiency.
Method used
The neural network model is used to infer the iterative description information of nonlinear equation systems, automatically determine the appropriate convergence strategy, and iteratively solve it through the target convergence strategy.
Benefits of technology
It improves the iterative solution performance of nonlinear equation systems, improves the solution speed and accuracy, reduces the solution time, and avoids generalization problems based on expert experience.
✦ Generated by Eureka AI based on patent content.
Abstract
The embodiments of the present application disclose a method for solving a nonlinear equation set. In the method, when an ith iteration of a first nonlinear equation set is performed, a neural network model can identify, on the basis of input description information of the ith iteration, equation set feature information of the first nonlinear equation set during the ith iteration, so as to efficiently determine a first convergence policy which is relatively suitable for the first nonlinear equation set during the current iteration process, such that a target convergence policy is determined on the basis of the first convergence policy, and iterative solving is efficiently and accurately performed on the first nonlinear equation set on the basis of the target convergence policy, thereby improving the iterative solving performance of the nonlinear equation set (e.g., increasing the solving speed and reducing the time taken for solving), and avoiding the generalization problem caused by a convergence policy and a default convergence policy determined on the basis of expert experience, and thus solving the problem of convergence policies determined in different scenarios having a poor effect and difficulty in achieving convergence when a nonlinear equation set is solved.
Need to check novelty before this filing date? Find Prior Art