Hybrid Quantum/Nonquantum QUBO Solving for Large-Scale NP-Hard Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current quantum and quantum/classical hybrid computing devices are limited in their ability to efficiently solve large-scale NP-hard combinatorial optimization problems.
Innovation Solution
A hybrid quantum/nonquantum computing system reformulates an optimization program as a Quadratic Unconstrained Binary Optimization (QUBO) model, using a quantum computing solver to generate multiple solutions, which are then input into a nonquantum computing solver to continue solving the optimization program, ultimately outputting an optimal solution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If current quantum or quantum/classical hybrid computing devices are used to solve large-scale NP-hard combinatorial optimization problems, then quantum computing capability is leveraged, but the devices are limited in their configuration and ability to solve such problems
Solution Approach 1:
The computing system is divided into two distinct solvers: a quantum computing solver for generating initial solutions and a non-quantum computing solver for refining and verifying solutions. This segmentation allows each component to be optimized for its specific function, overcoming the configuration limitations of single-device quantum systems.
Solution Approach 2:
The patent introduces an intermediary workflow that bridges quantum and non-quantum computing. The quantum solver generates solutions that are then passed to the non-quantum solver for further processing, creating a hybrid approach that leverages quantum capabilities while maintaining classical computational reliability and flexibility.
2Productivity
If a quantum computing solver is used to generate multiple solutions, then solution generation speed is improved, but the complexity of the computing system increases
Solution Approach 1:
The system segments the optimization process into two phases: solution generation (handled by the quantum solver) and solution refinement (handled by the non-quantum solver). This division of labor enables the quantum component to focus on generating multiple solutions efficiently while the classical component handles verification and optimization, managing overall system complexity.
Solution Approach 2:
The quantum computing solver generates a plurality of solutions in a single iteration, potentially exceeding the number of solutions needed. This partial action approach allows the system to leverage quantum parallelism for rapid solution generation, then use the non-quantum solver to process only the necessary solutions, optimizing the balance between speed and complexity.
3Measurement precision
If multiple solutions are generated and input into a nonquantum computing solver, then solution accuracy is improved, but processing time increases
Solution Approach 1:
The quantum computing solver performs preliminary action by generating multiple candidate solutions in advance. These solutions are then fed into the non-quantum solver which uses them as starting points for refinement. This preliminary solution generation leverages quantum computational speed to reduce the overall time required compared to using only classical methods for exhaustive search.
Data Source
AI summary
A system and method include reformulating an optimization program as a Quadratic Unconstrained Binary Optimization (QUBO) model and in a single iteration, solving the optimization program by inputting the QUBO model into a quantum computing solver, instructing the quantum computing solver to generate a plurality of solutions to the optimization program based on the QUBO model, receiving the plurality of solutions from the quantum computing solver, inputting each of the plurality of solutions into a nonquantum computing solver, wherein the nonquantum computing solver uses each of the plurality of solutions as a starting point to continue solving the optimization program, and outputting an optimal solution to the optimization program from the nonquantum computing solver.