Annealing and Constraint Logic Solvers for Binary Quadratic Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for solving binary quadratic optimization problems with complex logical constraints, such as CMOS annealing, often fail to guarantee constraint satisfaction and can lead to solutions that break constraints, and existing solutions like ADMM-based methods are inefficient for handling complicated constraints.
Innovation Solution
An optimization system that includes an annealing unit and a constraint logic unit, using iterative processing to search for an annealing solution and a constraint satisfaction solution respectively, with solvers to optimize the objective function while ensuring constraint satisfaction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the coefficient of the penalty term is excessively increased to prioritize constraint satisfaction, then constraint satisfaction is improved, but the value of the objective function is not improved and it becomes difficult to obtain an optimum solution
Solution Approach 1:
The patent segments the solution process into two distinct phases: a first search phase that optimizes the objective function without strict constraint enforcement, and a second search phase that enforces constraints. This segmentation allows each phase to focus on its specific goal without the trade-off present in single-phase methods.
Solution Approach 2:
The patent performs preliminary optimization of the objective function in the first search phase before applying constraint enforcement in the second phase. This preliminary action allows the system to explore the solution space and find promising regions before restricting the search to constraint-satisfying solutions.
2Reliability
If continuous optimization methods are used to satisfy complicated constraints, then constraint satisfaction is improved, but the solution takes too much time and is delayed
Solution Approach 1:
The patent divides the optimization process into two separate search operations: a first search using a first solver for objective function optimization, and a second search using a second solver for constraint satisfaction. This segmentation enables the use of different solving strategies optimized for their specific purposes, improving overall efficiency.
Solution Approach 2:
The patent performs a second search in the vicinity of the annealing solution obtained by the first solver, rather than searching the entire solution space. This partial action focuses computational resources on the most promising regions, significantly reducing solution time while maintaining constraint satisfaction.
Data Source
AI summary
An annealing unit performs a first search for an annealing solution to lower an objective function value in the vicinity of a constraint satisfaction solution, by using a first solver. A constraint logic unit performs a second search for a constraint satisfaction solution in the vicinity of an annealing solution obtained by the annealing unit, by using a second solver, and performs iterative processing of the first search and the second search to obtain an optimum solution.