Tensor Network Configuration Search for Unconstrained Optimization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods struggle to efficiently solve unconstrained optimization problems, particularly those with multiple variables and complex relationships, due to high computational complexity, making it difficult to find superior configurations for processes and systems in a timely manner.

Innovation Solution

A computer-implemented method using a Tensor Network to convert discrete variable cost function equations into unconstrained optimization problems, iteratively modifying tensor coefficients with different sets of parameters to reduce the cost function value, and checking for convergence criteria to achieve optimization.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If quantum computing methods are used to solve optimization problems, then solution speed is improved, but device complexity increases

Engineering Contradiction:
Improvesolution speedVSAvoiddevice complexity
Core Design Contradiction:
SpeedVSDevice complexity

Solution Approach 1:

The patent introduces a quantum processor as an intermediary component that specifically handles the quantum annealing computation, while classical processors manage the overall optimization workflow, data preprocessing, and result interpretation. This mediator approach allows the system to leverage quantum speedup for the core optimization task without requiring the entire system to be quantum, thus improving solution speed for optimization problems while keeping overall device complexity manageable through specialized architecture.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Manufacturing precision

If tensor network methods are used to solve unconstrained optimization problems, then manufacturing precision is improved, but device complexity increases

Engineering Contradiction:
Improveoptimization precisionVSAvoidcomputational complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent segments the optimization problem into discrete variables that can be represented and manipulated within the tensor network framework. By breaking down complex optimization problems into manageable discrete components, the system achieves high precision in finding optimal configurations while keeping the computational device complexity tractable through structured tensor representations and efficient contraction algorithms.

Inventive Principle:
Principle #1Segmentation

3Productivity

If multiple variables with complex relationships are optimized, then productivity is improved, but loss of time increases

Engineering Contradiction:
Improveprocess optimization efficiencyVSAvoidcomputation time
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent replaces traditional classical computational mechanics with quantum annealing mechanics to solve optimization problems involving multiple variables with complex relationships. The quantum system naturally explores the solution space through quantum tunneling and thermal annealing processes, finding optimal configurations faster than classical algorithms, thus improving productivity in optimizing multi-variable systems while reducing the time loss associated with complex computations.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP4047532A1Methods and devices for optimizing processes and configurations of apparatuses and systems
Publication Date: 2022.08.24 MULTIVERSE COMPUTING SL
  • EP4047532A1 patent drawingFigure 1~2
  • EP4047532A1 patent drawingFigure 3
  • EP4047532A1 patent drawingFigure 4~6

AI summary

A computer-implemented method whereby an equation with a cost function for minimization is solved by means of a tensor network. Coefficients of tensors of the tensor network are modified so as to reduce a value of the cost function in an iterative process until convergence is reached, at which point the concerned Unconstrained Optimization problem is solved and the values of the variables of the cost function are provided.