Ising Model Data Processing Device Group Switching Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As the number of state variables increases in combinatorial optimization problems, the large number of weighting factors required for the Ising model becomes impractical to store in a storage unit, leading to inefficiencies in switching between groups of state variables during solution finding, particularly due to the time-consuming process of reading and updating local fields.
Innovation Solution
A data processing device that divides the problem into subproblems and switches between groups of state variables, using change information to selectively read and update weighting factors, reducing the overhead of reading and updating local fields by only considering changes in state variables from other groups.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the number of state variables is increased to solve larger combinatorial optimization problems, then the problem-solving capability is improved, but the storage capacity required for weighting factors becomes impractical
Solution Approach 1:
The patent divides the set of all state variables into multiple groups, where each group can be processed independently. By segmenting the problem into smaller subsets, the system only needs to load weighting factors for the current group into storage, rather than storing all weighting factors simultaneously. This reduces the required storage capacity while maintaining the ability to solve larger optimization problems through iterative group processing.
2Reliability
If all weighting factors are stored in the storage unit, then the completeness of the Ising model is maintained, but the switching time between groups of state variables increases
Solution Approach 1:
The patent pre-calculates and stores the local field values for each group of state variables before switching between groups. By performing this calculation in advance and storing the results, the system avoids time-consuming recalculation during group switching. This preliminary action maintains the reliability of the Ising model while significantly reducing the time required to switch between different groups of state variables.
3Measurement precision
If the local field is updated for all state variables, then the accuracy of the search is maintained, but the processing time increases
Solution Approach 1:
The patent updates the local field only for the currently active group of state variables rather than all state variables. This localized update approach maintains search accuracy for the current group while reducing overall processing time. The system selectively applies computational resources to the relevant subset of variables, achieving a balance between precision and efficiency in the optimization search process.
Data Source
AI summary
A data processing method including: upon switching an object to a first group among groups of an Ising model when a search for a solution to a problem represented by the Ising model is performed by switching to each of the groups obtained by dividing the state variables, reading first weighting factors corresponding to pairs of the state variables whose values have changed and each first state variable belonging to the first group; updating a local field of the first state variable; executing the search on the first group by using second weighting factors and the local field of the first state variable; and after ending the search on the first group, updating the change information according to presence or absence of a change in the values for the first state variable by the search at a current time, to switch the object to be searched to a next group.