AEDA Papers AEDA Papers

PDF copies of all AEDA papers are available for download by AEDA Members logged into this website.

September 2016

Analysis of Uncertainty in Engineering Design Optimization Problems

Authors: Philipp Fleck (HEAL), Michael Kommenda (HEAL), Michael Affenzeller (HEAL), Thorsten Prante (VRES)

Abstract: In this paper, we analyze popular benchmark instances in the field of engineering design optimization regarding the robustness of published solutions. First, we implement selected benchmark problems with HeuristicLab and show the advantages of having a framework that enables rapid prototyping for optimization and analysis. Then, we show that many solutions quickly become infeasible when considering uncertainty like production inaccuracies. Based on these findings, we motivate why robust solutions for engineering design are important and present methods for measuring, identifying and visualizing robustness. Finally, we present how solutions can be compared and selected using a novel robustness measure.

Keywords: Engineering Design Optimization, Constraint Handling, Uncertainty, Robustness

Conference / Journal: European Modeling and Simulation Symposium (Simulation in Industry) - EMSS 2016

September 2016

Evolution Under Strong Noise: A Self-Adaptive Evolution Strategy Can Reach the Lower Performance Bound - The pcCMSA-ES

Authors: Michael Hellwig (FhV) and Hans-Georg Beyer (FhV)

Abstract: According to a theorem by Astete-Morales, Cauwet, and Teytaud, “simple Evolution Strategies (ES)” that optimize quadratic functions disturbed by additive Gaussian noise of constant variance can only reach a simple regret log-log convergence slope ≥−1/2 (lower bound). In this paper a population size controlled ES is presented that is able to perform better than the −1/2 limit. It is shown experimentally that the pcCMSA-ES is able to reach a slope of −1 being the theoretical lower bound of all comparison-based direct search algorithms.

Keywords: -----

Conference / Journal: International Conference on Parallel Problem Solving from Nature - PPSN 2016

September 2016

A Path Planning Algorithm for a Materials Handling Gantry Robot and its Validation by Virtual Commissioning

Authors: Ruth Fleisch (VRES), Robert Schöch (VRES), Thorsten Prante (VRES), Reinhard Pfefferkorn (Schelling Anlagenbau)

Abstract: Adaptive and intelligent path planning is required to provide fully automated and optimized transportation systems directed at reducing the non-productive time related to materials handling. This paper presents a case study from the machine-tool industry sector: In order to optimize the efficiency of a cut-to-size plant with feedbacks in the material flow, a materials handling gantry robot is employed for transporting panels of largely varying sizes. For each of these panels, a path within spatial limitations, which are partially very tight, is planned so that collisions are prevented. This paper describes the developed path planning algorithm and its integration with the programmable logic controller (PLC) of the robot. Towards improving solution quality and decreasing lead time, first, the integration with the PLC is realized by software components automatically generated based on code developed in a high-level programming language. Second, virtual commissioning of the materials handling robot was performed.

Keywords: Robot Path Planning, Control System, Virtual Commissioning, Gantry Robot.

Conference / Journal: International Conference on Manufacturing Research - ICMR 2016

July 2016

Reuse of Engineering Data, Information and Knowledge in Collaborative and Automated Design Processes

Authors: Jullius Cho (TUW), Thomas Vosgien (VRES), Thorsten Prante (VRES), Detlef Gerhard (TUW)

Abstract: This paper addresses the challenges related to KBE-PLM systems integration in order to unlock engineering knowledge from proprietary representations and to manage the lifecycle of KBE models as well as their usage by different design automation applications. Essential constituents of product knowledge are identified and analyzed and the concepts of design intent and design rationale are re-introduced as key enablers to re-use this product knowledge in the appropriate KBE context. The paper introduces a KBE-PLM integration framework including a platform-independent Open KBE repository structured according to the KBE-PLM integration schema. This schema is a multi-layer neutral product and knowledge data model designed for integrating information from parameterized Computer-Aided Design (CAD) models, rule-based KBE systems and PLM systems.

Keywords: Knowledge-Based Engineering, Design Automation, Knowledge re-use, Design Intent, Design Rationale.

Conference / Journal: IFIP International Conference on Product Lifecycle Management - PLM16

May 2016

Estimating the Potential of State of the Art Design Automation - Tasks, Methods, and Benefits

Authors: Eugen Rigger (VRES-ETH), Clemens Münzer (ETH), Kristina Shea (ETH).

Abstract: In this study, multiple methods for categorization and characterization of design tasks are analyzed and a mapping of design task characteristics to methods is presented. Further, approaches from literature regarding estimation of potential and success validation are investigated. Due to the lack of an unambiguous categorization of design tasks as well as a general distinct mapping of tasks to methods, based on analysis of multiple recently published design automation methods, this paper introduces an approach that allows mapping of design task characteristics to design automation methods. Additionally, an implicit mapping of automation methods to potential benefits is performed. Therefore, it is expected to improve the understanding of opportunities and potential benefits of design automation in industry.

Keywords: Design Automation, Task Specification, Task Categorization, Computational Design Synthesis (CDS), Evaluation of benefits

Conference / Journal: International Design Conference - DESIGN 2016

April 2016

Mutation strength control via meta evolution strategies on the ellipsoid model

Authors: Michael Hellwig (FhV), Hans-Georg Beyer (FhV)

Abstract: In this paper, The ability of a hierarchical organized evolution strategy (Meta-ES) to optimally control its mutation strength is investigated on convex-quadratic functions (referred to as ellipsoid model). Applying the dynamical systems analysis approach a first step towards the analysis of the Meta-ES behavior is conducted considering isolation periods of length = 1. A nonlinear system of difference equations is derived to describe the mean-value evolution of the Meta-ES. In the asymptotic limit case of large search space dimensions this system is suitable to derive closed-form solutions which describe the longterm behavior of the Meta-ES. The steady state mutation strength is bracketed within an interval depending on the mutation strength control parameter . Compared to standard settings in cumulative steplength adaptation (CSA) evolution strategies the Meta-ES realizes almost similar normalized mutation strengths. The performance of the Meta-ES turns out to be more robust to the choice of its control parameters. The results allow for the derivation of the expected running time of the algorithm.

Keywords: Evolution strategies, Meta-ES, ellipsoid model, mutation strength adaptation

Conference / Journal: Theoretical Computer Science, Volume 623 Pages 160-179.

March 2015

Effizienzoptimierte Konstruktion von Kranen und deren Komponenten: Lösungswege und Ergebnisse

Authors: Gerald Frank (VRES), Thorsten Prante (VRES), Vaheh Khachatouri (VRES), Martin Schwarz (LWN), Klara Fink (LWN)

Abstract: Dieser  Beitrag zeigt, dass die Möglichkeiten des automatisierten Erstellens von CAD-Modellen und ihrer Derivate nicht vom Komplexitätsgrad der jeweiligen Baugruppe abhängen. Sowohl bei Detail-, Haupt- als auch bei Gesamtbaugruppen ist es möglich, die Effizienz des Konstruktionsprozesses mittels Konstruktionsautomatisierung erheblich zu steigern. Abhängig von den Charakteristiken und dem Einsatzziel (Vertrieb oder Konstruktion bzw. Produktion) der zu automatisierenden Baugruppe stehen, wie erläutert, unterschiedliche Herangehensweisen zur Verfügung, die in einen allgemeinen Lösungsansatz eingebettet sind und durch Software-Frameworks unterstützt werden. Auf Basis der diskutierten Faktoren kann die Anwendbarkeit der beschriebenen Ansätze aufgrund des jeweils resultierenden Kosten-Nutzen-Verhältnisses bzgl. Umsetzungsaufwand und Kostenersparnis bewertet und gewählt werden. Die Anwendung des vorgestellten Ansatzes ist nicht auf einzelne CAD-Systeme begrenzt. Er wurde so konzipiert, dass jedes CAD-System, das eine entsprechende Schnittstelle zur Verfügung stellt, eingebunden werden kann.

Event: 23. Kranfachtagung: Krane - Herausforderungen im Zeichen von Industrie 4.0. Technische Universität Dresden