Linear Model Partitioning for Large-Scale Portfolio Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing linear optimization systems face challenges in efficiently reducing the computational resource needs for processing and storing large volumes of data in electronic trading systems, particularly in managing multilateral portfolios of bilateral financial instruments.
Innovation Solution
The system employs a computer-implemented method that iteratively partitions large linear optimization models into smaller sub-models, which are solved independently to approximate the solution of the full model, thereby reducing computational resource requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If large linear optimization models are solved directly, then solution accuracy is maintained, but computational resource requirements increase significantly
Solution Approach 1:
The patent divides a large linear optimization model into multiple smaller sub-models that can be solved independently. Each sub-model represents a partition of the original model's variables and constraints, allowing parallel processing and reduced memory requirements while maintaining solution accuracy through coordinated solving of all sub-models.
Solution Approach 2:
The patent introduces a new dimension of problem-solving by solving multiple smaller sub-models in parallel rather than one large model sequentially. This dimensional transformation from a single large computation to multiple smaller concurrent computations reduces the computational burden on any single system resource while maintaining overall solution integrity.
2Reliability
If large linear optimization models are solved directly, then complete solution is obtained, but processing time increases
Solution Approach 1:
The patent segments the large optimization model into smaller sub-models that can be solved simultaneously. This segmentation enables parallel processing where multiple sub-models are solved at the same time rather than sequentially, dramatically reducing total processing time while ensuring a complete solution is obtained by combining results from all sub-models.
Solution Approach 2:
The patent maintains continuous useful action by solving multiple sub-models in parallel without idle time. While one sub-model is being solved, other sub-models are simultaneously being processed, eliminating the sequential waiting time that would occur if a single large model were solved step-by-step, thus reducing total processing time while maintaining solution completeness.
3Productivity
If computational resources are increased to handle large data volumes, then processing capability improves, but costs increase
Solution Approach 1:
The patent segments large optimization problems into smaller sub-models that can be solved with fewer computational resources each. This allows existing hardware and software resources to handle multiple smaller problems in parallel rather than requiring a single massive computational system, thereby maintaining processing capability while reducing the quantity of computational resources needed.
Solution Approach 2:
The patent applies partial action by solving multiple smaller sub-models separately rather than attempting to solve one complete large model at once. Each sub-model requires only partial computational resources, and the cumulative effect of solving all sub-models achieves the same productivity as solving the full model, but with reduced peak resource requirements and lower overall costs.
Data Source
AI summary
The disclosed embodiments related to multilateral portfolio compression using general large-scale linear optimization which pre-processes a model to decrease model size using domain knowledge to remove variables to reduce dimensionality, thereby making the model faster to solve and improving numerical characteristics. but it would not remove, for example, as much as half of the model, but rather a smaller fraction. The disclosed pre-processing enables an approximate solution for large, linear optimization models by automatically iteratively and selectively partitioning them into independently easily solvable sub-models. The sub-models are themselves linear optimization models, which can be solved with any preferred algorithm or library. The solutions for each sub-model are aggregated to obtain an acceptable, e.g., approximate, solution for a large model without solving the full model. At each iteration the disclosed embodiments will have a valid, feasible solution, if the user is satisfied before full convergence.