Thresholded Extremal Optimization for Fast Large-Scale QUBO Solving
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current solutions for combinatorial optimization problems, particularly Quadratic Unconstrained Binary Optimization (QUBO) problems, face long run times when applied to commercial-sized problems with thousands or millions of variables.
Innovation Solution
A method and apparatus that simultaneously assess a plurality of fitness values and update configuration values using probabilistic selection to avoid deterministic loops, enabling efficient determination of an optimized configuration vector.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If current solutions are applied to commercial-sized combinatorial optimization problems, then the problems can be solved, but the run time becomes excessively long
Solution Approach 1:
The patent segments the combinatorial optimization problem into multiple smaller sub-problems or components that can be processed in parallel. By dividing the large-scale problem into manageable chunks, the computational burden is reduced and processing time is decreased while maintaining solution quality.
Solution Approach 2:
The patent performs preliminary actions by pre-processing the optimization problem data, pre-calculating certain parameters, or initializing data structures in advance. This preliminary work reduces the computational complexity during the actual optimization process, leading to faster execution for commercial-sized problems.
2Ease of operation
If deterministic algorithms are used to solve optimization problems, then the solution process is straightforward, but deterministic loops occur reducing efficiency
Solution Approach 1:
The patent introduces dynamic elements into the optimization algorithm by incorporating adaptive parameters that change during execution. This allows the algorithm to dynamically adjust its behavior based on the problem state, avoiding rigid deterministic loops while maintaining operational simplicity through structured control mechanisms.
Solution Approach 2:
The patent changes key parameters during the optimization process, such as adjusting convergence thresholds, modifying search step sizes, or altering probability distributions. These parameter changes prevent the algorithm from getting stuck in deterministic loops while keeping the overall process manageable through systematic parameter management.
Data Source
AI summary
In some embodiments, the present disclosure relates to a method. The method includes accessing an instance matrix having a plurality of instance values and a configuration vector having a configuration values. Iterations are performed on a computing apparatus to determine an optimized configuration vector. The iterations respectively include simultaneously determining a plurality of reduction values by multiplying the configuration values by instance values within a row of the instance matrix. A plurality of fitness values are respectively determined using the plurality of reduction values and configuration value associated with a row of the instance matrix. A current cost is determined by summing the plurality of fitness values. Unstable fitness values are simultaneously identified based upon a comparison of the plurality of fitness values with a threshold. At least one of the configuration values associated with the unstable fitness values are simultaneously updated based upon a probabilistic selection.