Title: Novel robustness measures for engineering design optimisation

Authors: Philipp Fleck; Michael Kommenda; Thorsten Prante; Michael Affenzeller

Addresses: Heuristic and Evolutionary Algorithms Laboratory, University of Applied Sciences Upper Austria, Hagenberg, Austria ' Heuristic and Evolutionary Algorithms Laboratory, University of Applied Sciences Upper Austria, Hagenberg, Austria; Institute for Formal Models and Verification, Johannes Kepler University, Linz, Austria ' Heuristic and Evolutionary Algorithms Laboratory, University of Applied Sciences Upper Austria, Softwarepark 11, 4232 Hagenberg im Mühlkreis, Austria; Institute of Formal Models and Verification, Johannes Kepler University Linz, Altenbergerstraße 69, 4040 Linz, Austria; Design Automation, V-Research GmbH, Stadtstraße 33, 6850 Dorinbirn, Austria ' Heuristic and Evolutionary Algorithms Laboratory, University of Applied Sciences Upper Austria, Hagenberg, Austria; Institute for Formal Models and Verification, Johannes Kepler University, Linz, Austria

Abstract: This paper presents novel robustness measures to analyse and compare the robustness of solutions for constrained optimisation problems in the field of engineering design optimisation. First, we define uncertainty in production processes and present a method to quantify uncertainty. Based on the variations of a solution that are introduced by uncertainty, we want to assess the robustness of those solutions towards those variations. We show how a solution's quality and feasibility (with regards to constraint violation) change with increasing uncertainty and discuss how those changes determine the robustness of that solution. Furthermore, we present a method of aggregating that information into a single, real-valued robustness measure. This novel robustness measure can be used to select solutions that have a high robustness along with a high quality. To test the presented measures extensively, we apply them to various solutions for benchmark problems from published literature in the field of engineering design optimisation.

Keywords: engineering design optimisation; uncertainty; robustness; optimisation; benchmark; constrained optimisation; multi-objective.

DOI: 10.1504/IJSPM.2018.093757

International Journal of Simulation and Process Modelling, 2018 Vol.13 No.4, pp.387 - 401

Received: 01 May 2017
Accepted: 26 Oct 2017

Published online: 03 Aug 2018 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article