Calculating Device Variable Update Method for Ising Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for solving large-scale optimization problems, such as Ising problems, are inefficient due to high computational requirements and lack of parallel processing capabilities, particularly in updating variables which necessitate extensive calculations.
Innovation Solution
A calculating device and method that repeatedly updates variables using specific functions and parameters, allowing for parallel processing by separating the updates of first, second, and third variables, reducing computational load through optimized function updates and element functions, and utilizing a symplectic Euler method for Hamiltonian dynamical systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current methods are used to solve large-scale optimization problems, then solution accuracy can be maintained, but computational time and complexity increase significantly
Solution Approach 1:
The patent segments the optimization problem into two distinct variable sets: first variables (x) representing Ising spins and second variables (y) representing auxiliary continuous variables. This segmentation allows independent update rules to be applied to each variable type, enabling parallel processing and reducing overall computational time while maintaining solution accuracy through coordinated evolution of both variable sets.
Solution Approach 2:
The patent introduces dynamic update rules where first variables are updated based on second variables, and second variables are subsequently updated based on the new first variable values. This dynamic, sequential update mechanism within each iteration allows the system to evolve toward optimal solutions efficiently, balancing computational speed with solution accuracy.
2Measurement precision
If sequential variable updates are performed, then calculation accuracy is maintained, but parallel processing capability is lost
Solution Approach 1:
By segmenting variables into two distinct sets with different update dependencies, the patent enables parallel processing of first variables while second variables are being updated, and vice versa. This segmentation maintains the necessary sequential logic for accuracy while exploiting parallel computational capabilities across multiple processors or cores.
Solution Approach 2:
The patent implements periodic alternating updates where one iteration focuses on updating first variables while holding second variables constant, and the next iteration updates second variables while holding first variables constant. This periodic action pattern enables parallel processing within each phase while maintaining calculation accuracy through the alternating refinement of both variable sets.
3Measurement precision
If extensive calculations are performed for variable updates, then solution precision is improved, but computational load increases
Solution Approach 1:
The patent divides the computational workload into two separate update processes: one for first variables using a simplified update rule, and another for second variables using a different update rule. This segmentation reduces the computational complexity of each individual update operation compared to a monolithic update approach, thereby reducing overall computational load while maintaining solution precision through the coordinated interaction of both variable sets.
Solution Approach 2:
The dynamic update mechanism allows the system to adapt computational effort across iterations. Early iterations use broader updates to quickly approach the solution space, while later iterations can focus computational resources on refining specific variable sets, thereby optimizing the balance between solution precision and computational load.
Data Source
AI summary
According to one embodiment, a calculating device includes a processor repeating a processing procedure. The processing procedure includes first, second, and third variable updates. The first variable update includes updating an ith entry of a first variable xi by adding an ith entry of a first function to the first variable xi. The second variable update includes updating the second variable yi by adding, to the second variable yi, an arithmetic result of an ith entry of a second function, an ith entry of a third function, and an ith entry of a first element function. The third variable update includes updating the third variable z by adding an ith entry of a second element function to the third variable z. The processor performs at least an output of at least one of the first variable xi or a function of the first variable xi.