Annealing Execution Unit for Combinatorial Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing solutions for combinatorial optimization problems do not allow for efficient modification of already-obtained solutions in response to changing circumstances, such as altering worker schedules to accommodate urgent business needs.
Innovation Solution
A system that uses annealing execution and solution determination methods to derive multiple new solutions based on an energy function representing spin states, allowing users to modify specific aspects of the solution while minimizing changes to other variables, using either the Ising or QUBO models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If simulated annealing is used to obtain a solution to a combinatorial optimization problem, then a solution can be obtained using the energy function, but the solution cannot be efficiently modified when circumstances change
Solution Approach 1:
The patent stores the solution obtained from simulated annealing as initial values for subsequent optimization processes. This preliminary action allows the system to start from a known good solution rather than random initialization, enabling efficient modifications when constraints change while maintaining solution quality.
Solution Approach 2:
The patent implements a dynamic optimization process where the system can re-run simulated annealing or use gradient descent with the stored solution as initial values when constraints change. This dynamic approach allows the solution to adapt to new circumstances efficiently without complete re-optimization.
2Reliability
If the entire solution is re-optimized when constraints change, then the new constraints are satisfied, but unnecessary changes occur in parts of the solution that should remain unchanged
Solution Approach 1:
The patent applies local quality by using gradient descent optimization that can target specific variables or regions of the solution. When constraints change, the system can focus optimization efforts on the affected portions of the solution while maintaining stability in unrelated areas, thus satisfying new constraints without unnecessary global changes.
Solution Approach 2:
The patent segments the optimization process by allowing different optimization methods (simulated annealing for global optimization, gradient descent for local refinement) to be applied selectively. This segmentation enables targeted re-optimization of specific solution components rather than forcing complete re-optimization of the entire solution.
3Adaptability or versatility
If simulated annealing is re-executed to modify the solution, then new solutions can be found, but the computational cost and time increase significantly
Solution Approach 1:
The patent applies partial action by using gradient descent optimization as a more efficient alternative to full simulated annealing re-execution. When constraints change, the system can perform partial re-optimization using gradient descent with the stored solution as initial values, achieving adequate solution modification without the excessive computational cost of complete simulated annealing re-execution.
Solution Approach 2:
The patent changes the optimization parameters by switching between simulated annealing (for initial solution acquisition) and gradient descent (for efficient local refinement and modification). This parameter change in the optimization approach allows rapid solution modification when constraints change, significantly improving computational efficiency compared to always re-executing simulated annealing.
Data Source
AI summary
The annealing execution means 73 executes annealing using an energy function, when a solution to a combinatorial optimization problem, and the energy function of a model representing states of individual spins by a first value or a second value which is used to obtain the solution are given, and derives multiple new solutions to the combinatorial optimization problem. The solution determination means 74 determines a solution from among the multiple solutions derived by the annealing execution means 73, based on a solution determination rule for determining a solution among the multiple new solutions and the given solution to the combinatorial optimization problem.