Iterative Linear Model Partitioning for Feasible Sub-Model Solving
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Electronic trading systems face challenges in efficiently managing and reducing the volume of data related to positions held in financial instruments, which leads to increased computational resource needs, communication bandwidth limitations, and higher costs, particularly in systems handling large volumes of concurrent transactions.
Innovation Solution
A computer-implemented system that iteratively partitions large linear optimization models into smaller sub-models, allowing for independent solution of these sub-models using existing algorithms, thereby reducing the computational resources required and optimizing the solution process, while maintaining the integrity of the original model's constraints and objectives.
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 consumption increases and solution time increases
Solution Approach 1:
The patent divides a large linear optimization model into multiple smaller sub-models by partitioning the set of variables into different groups. Each sub-model contains a subset of the original constraints and variables, making them computationally more efficient to solve while collectively representing the full problem space.
Solution Approach 2:
The patent introduces an iterative dimension to the solution process, solving sub-models in multiple passes and progressively refining the solution. This transforms a single large-scale problem into a sequence of smaller problems solved across different iteration levels, reducing memory requirements and computational complexity.
2Reliability
If large linear optimization models are solved directly, then complete solution is obtained, but solution time increases
Solution Approach 1:
The patent partitions the optimization model into multiple sub-models that can be solved independently and in parallel. This segmentation allows the solution process to be distributed across multiple computational units, significantly reducing overall solution time while maintaining solution completeness through iterative refinement.
Solution Approach 2:
The patent performs preliminary partitioning of the model into sub-models before the actual optimization process. This pre-processing step organizes the problem structure in advance, enabling more efficient computation during the optimization phase by avoiding the need to handle the full model structure at each iteration.
3Quantity of substance
If data volume is reduced, then computational resource needs decrease, but data management complexity increases
Solution Approach 1:
The patent segments the large dataset into smaller partitions corresponding to different sub-models. This segmentation reduces the amount of data that needs to be held in memory at any given time and allows for more manageable data processing operations on each subset while maintaining the ability to reconstruct the full solution.
Data Source
Figure 1
Figure 2
Figure 3
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.