Data and Knowledge Driven Optimization Model Generation for Flow Based Optimization Problems

Abstract

Optimization can provide exceptional value, but requires significant time and skills to implement, thereby significantly limiting its widespread use. In this talk, we will show how optimization models for flow-based problems can be automatically generated from a combination of data and easily specifiable knowledge. We will describe the underlying technology and present a brief demonstration. In addition, we will discuss several approaches for overcoming the inherent inaccuracies existing in automatically generated models, so as to ensure the quality of the provided optimization solutions.