Method and apparatus for generating alternative representation of optimization models

Abstract

A method for determining an alternative representation of an optimization model reduces model input through compact representation of model parameters. Model generation is performed at varying levels of complexity (approximation) depending on pre-defined, business approved thresholds.

Description

GOVERNMENT LICENSE RIGHTS [0001] This invention was made with Government support under Contract No. 43-82X9-3-5073 awarded by USDA, Forest Service. The U.S. Government has certain rights in this invention.BACKGROUND OF THE INVENTION [0002] 1. Field of the Invention [0003] The present invention generally relates to computer-implemented optimization models for solution of problems and, more particularly, to an approach for generating alternative representations for an optimization model while keeping model parameters at an acceptable, pre-determined accuracy threshold. [0004] 2. Background Description [0005] Real-world modeling problems often result in an implementation that makes the model instances intractable due to computationally prohibitive data size and structural complexity. The prior art solutions to such large problems involved simplification of the model by either aggregating the data or simplifying the model assumptions. Hence, the resulting solution is sub-optimal due to ...

AI Technical Summary

Benefits of technology

[0007] The present invention addresses the shortcomings of the prior art solutions by employing a flexible modeling approach which enables generation of model instances at varying levels of complexity, thereby enabling scalability. The present invention is thus able to solve model instances ranging from small size and / or low complexity to large size and / or high complexity, with graceful degradation of optimality.

[0015] The second coordinate (y-axis value) of the anchor point. This results in a compaction of input data by eliminating redundant data points while maintaining the precision of the input function.

[0023] It is another object of the invention to enable model generation at varying levels of complexity (approximation) depending on pre-defined, business approved thresholds. According to the invention, there is provided an approach for generating alternative representations for an optimization model which has the following advantages:

[0024] Instances can be generated at varying levels of complexity depending on a pre-determined threshold. This in turn enables scalability.

[0026] Solve an efficient model with reduced memory requirements, improved solution time, and improved solution quality. More specifically, the invention provides a method for determining an alternative representation of an optimization model based on establishing an acceptable threshold for modifying parameters. According to the invention,

[0028] Domain Reduction facilitates an efficient reformulation of the model, and

Problems solved by technology

Real-world modeling problems often result in an implementation that makes the model instances intractable due to computationally prohibitive data size and structural complexity.

As a result, the prior art solutions are not fully satisfactory.

Images