Information Processing for Sign-Guided Combinatorial Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle with obtaining accurate solutions for large-scale combinatorial optimization problems or require excessive processing time.
Innovation Solution
An information processing apparatus employing a first search unit to find a provisional solution using a first search machine and a second search unit to output a final solution based on the comparison between the signs of the provisional solution and the partial differential value of the objective function, utilizing Ising machines like simulated annealing or bifurcation machines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a first search machine solves the overall combinatorial optimization problem, then the solution is obtained, but the accuracy is insufficient and processing time is excessive
Solution Approach 1:
The patent divides the combinatorial optimization problem into two parts: an overall problem solved by a first search machine and subproblems solved by a second search machine. The overall problem involves optimizing all dimensions, while subproblems focus on specific incomplete dimensions where the sign of the provisional solution matches the sign of the partial differential value. This segmentation allows each search machine to handle appropriately sized problems, improving overall accuracy while reducing processing time through parallel computation.
Solution Approach 2:
The patent applies partial action by identifying and solving only the subproblems corresponding to incomplete dimensions rather than repeatedly solving the entire problem. The second search machine is invoked only for dimensions where the sign mismatch occurs, performing optimization actions only where necessary. This partial approach avoids redundant computations and reduces total processing time while maintaining solution accuracy.
2Measurement precision
If the first search machine solves the overall problem, then a provisional solution is obtained, but the solution accuracy is insufficient
Solution Approach 1:
The patent implements feedback by comparing the sign of the provisional solution with the sign of the partial differential value for each dimension. Based on this feedback, the system identifies which dimensions require further optimization (where signs match) and which are sufficiently optimized (where signs mismatch). This feedback mechanism guides the second search machine to focus only on relevant subproblems, improving solution accuracy without sacrificing processing efficiency.
3Measurement precision
If the second search machine solves subproblems for incomplete dimensions, then solution accuracy is improved, but device complexity increases
Solution Approach 1:
The patent applies universality by designing the second search machine to handle multiple subproblems with different dimensions and constraints. Rather than creating separate specialized machines for each subproblem, the second search machine is designed as a universal solver that can process any subproblem by receiving the appropriate parameters (objective function, constraints, and dimension specifications). This multi-functional approach improves solution accuracy while avoiding the complexity of multiple specialized machines.
Data Source
AI summary
According to one embodiment, an information processing apparatus comprising a processor. The processor is configured to execute: first search processing of searching for a first provisional solution of a combinatorial optimization problem in which an objective function is partially differentiable, using a first search machine that solves an overall problem of the combinatorial optimization problem; and a second search processing of outputting a final solution of the combinatorial optimization problem based on a comparison between signs of the first provisional solution and a partial differential value of the objective function.