A Visual Web Tool to Perform What-If Analysis of Optimization Approaches
Artificial Intelligence
2017-03-20 v1 Performance
Abstract
In Operation Research, practical evaluation is essential to validate the efficacy of optimization approaches. This paper promotes the usage of performance profiles as a standard practice to visualize and analyze experimental results. It introduces a Web tool to construct and export performance profiles as SVG or HTML files. In addition, the application relies on a methodology to estimate the benefit of hypothetical solver improvements. Therefore, the tool allows one to employ what-if analysis to screen possible research directions, and identify those having the best potential. The approach is showcased on two Operation Research technologies: Constraint Programming and Mixed Integer Linear Programming.
Cite
@article{arxiv.1703.06042,
title = {A Visual Web Tool to Perform What-If Analysis of Optimization Approaches},
author = {Sascha Van Cauwelaert and Michele Lombardi and Pierre Schaus},
journal= {arXiv preprint arXiv:1703.06042},
year = {2017}
}