Evolutionary Computation Based Methods as Enabling Technologies for Prescriptive Analytics

Abstract

Prescriptive Analytics is a very promising field, considered by renowned think tanks, trends and futurologists as the next level of intelligent systems.

Prescriptive Analytics is built on the areas of Descriptive, Diagnostic and Predictive Analytics. While Descriptive Analytics aims to provide a better understanding of the past through descriptive and descriptive concepts, Diagnostic Analytics aims to investigate the causes, effects and interactions of states. Predictive Analytics goes one step further by attempting to predict future events and anticipate the occurrence of future events.
The goal of Prescriptive Analytics is to derive actions from past analysis and present knowledge and to derive optimized solutions and results that will be returned to the real world.

Metaheuristics in general and evolutionary algorithms in particular have a high potential to provide valuable contributions to current best practices in the field of prescriptive analytics. In the analysis on structured numerical data, white-box approaches can make a significant contribution in terms of interpretability, transfer learning, further processing and integration. Approaches based on genetic programming for symbolic regression, classification and time series analysis represent one of the strongest methodological concepts in the predictive field. In the prescriptive phase, in which the analytics findings must be transferred into decision making, metaheuristics such as genetic algorithms and evolution strategies play a decisive role in optimization, which can also take place on a simulation-based basis, especially at the strategic level.

The presentation will cover theoretical aspects as well as real world examples demonstrating how the open source framework HeuristicLab https://dev.heuristiclab.com/ can be used for modeling, optimization and machine learning tasks for concrete challenges in the domain of production, logistics and systems research.

Michael Affenzeller (Bild: Michael Affenzeller)

CV

Michael Affenzeller has published several papers, journal articles and books dealing with theoretical and practical aspects of evolutionary computation, genetic algorithms, and meta-heuristics in general. In 2001 he received his PhD in engineering sciences and in 2004 he received his habilitation in applied systems engineering, both from the Johannes Kepler University of Linz, Austria. Michael Affenzeller is professor for heuristic optimization and machine learning at the University of Applied Sciences Upper Austria, Campus Hagenberg, and head of the research group HEAL http://heal.heuristiclab.com/ . Since October 2014 he serves as the head of studies for the Master degree program Software Engineering and as vice-dean for R&D at the faculty of informatics, communications and media.