ENBIS-8 in Athens

21 – 25 September 2008 Abstract submission: 14 March – 11 August 2008

My abstracts

 

The following abstracts have been accepted for this event:

  • A methods utilization model for the acquisition of information in logistics networks

    Authors: Sonja Kuhnt, Sigrid Wenzel
    Affiliation: Eindhoven University of Technology/ Institute of Production Engineering and Logistics, University of Kassel
    Primary area of focus / application:
    Submitted at 29-Apr-2008 09:45 by Sonja Kuhnt
    Accepted
    22-Sep-2008 11:35 A methods utilization model for the acquisition of information in logistics networks
    The design, organization und management of logistics networks usually
    involve model-based analyses of the network. The usefulness of such an
    analysis highly depends on the quality of the input data, which naturally
    should capture the real circumstances best possible. We provide a methods
    utilization model for a structured acquisition of information in this
    context, which guides the user through the steps of a procedure model. It
    further includes a methods utilization kid to allow for the integrative,
    goal-oriented use of methods from different disciplines such as data
    acquisition, visualisation and statistics.
  • Exploratory data analysis in quality improvement projects

    Authors: Jeroen de Mast
    Affiliation: IBIS UvA
    Primary area of focus / application:
    Keywords: Discovery; Entropy; Graphical Data Analysis; Hypothesis Generation; Pattern Discovery.
    Submitted at 29-Apr-2008 12:35 by Jeroen de Mast
    Accepted
    23-Sep-2008 14:00 Keynote by Jeroen de Mast: Exploratory data analysis in quality improvement projects
    Compared to the vast literature on confirmatory data analysis (hypothesis testing, estimation, modeling), the literature on exploratory data analysis (EDA) is far less elaborate, both in pure volume of texts devoted to the subject and in precision and depth of its theoretical development. Sometimes, EDA is even described as an art, rather than a science.
    In this presentation I will show a number of explicated principles for EDA that can be taught to practitioners and statisticians to help them master this art faster. The framework is developed on the basis of a large number of real-life applications. The purpose and process of EDA are defined, and contrasted to the purpose and process of confirmatory data analysis and descriptive data analysis.
    In the process of EDA, three steps are discerned: display the data, identify salient features, and interpret salient features. The details of each of these steps are elaborated, and I will present the underlying principles, such as Shewhart’s assignable causes, the maximum entropy principle, abduction, and explanatory coherence. Furthermore, the roles of probabilistic reasoning and automatic statistical procedures in EDA are discussed. Finally, I will place EDA in the wider context of hypothesis and idea generation, a discipline that is studied in philosophy of science (discovery), the cognitive sciences (problem solving), and the medical sciences (diagnosis). We will study what approaches for hypothesis generation there are besides EDA, and we will analyse how EDA compares to these other approaches.
    The resulting framework provides structure and practical advice which facilitates teaching of EDA to practitioners and statisticians alike. The precise definitions, delineations and references to relevant scientific disciplines helps the further theoretical understanding and development of EDA.
  • Desirability Analysis in Construction Design Quality Improvement

    Authors: E.S. Telis, G. Besseris and C. Stergiou
    Affiliation: Technological and Educational Institute of Piraeus
    Primary area of focus / application:
    Submitted at 29-Apr-2008 13:28 by ELEFTHERIOS - STAMATIOS TELIS
    Accepted (view paper)
    22-Sep-2008 14:20 Desirability Analysis in Construction Design Quality Improvement
    Construction project management is one of the fields that necessitates continuous support from modern quality improvement tools because of local government enhanced building specifications as well as customer increased awareness on construction issues. This work regards a multi-response optimization problem that includes well-known quality responses in construction such as safety factor, steel tension and total displacements. The problem possesses a high degree of interest because each of these responses may follow a different optimization direction. Desirability analysis is utilized to decipher the optimum levels of active factors and their corresponding two-way interactions. A preliminary screening of an 8-run fractional factorial design for seven nominated control factors showed that only one was a dominant influence on all three responses. However, the interactions engaged pairs of all factors considered in the screening phase. Therefore, during robust design, a sixteen-run design was employed to accommodate seven control factors and eight interactions. Data were collected by running professional design software package Plaxis. Goal setting is gauged against several weight adjustments. The importance of responses was decided through a brainstorming session that involved operations and quality managers as well as several experienced engineers. It is important to stress that the improvement effort is focused early in the product design phase during the blueprint drawing process of a real excavation project that involves a six-floor building with a three-level underground garage. We discuss the results that seem to promote a linear dependence between responses and active factors and interactions.
  • An Application of Design of Experiments in a Real Lift Test Rig

    Authors: E. Viles M. Tanco I. Isasa U. Arteche X. Sagarzazu
    Affiliation: TECNUN-University of Navarra & Ikerlan – IK4
    Primary area of focus / application:
    Submitted at 29-Apr-2008 14:51 by Elisabeth Viles
    Accepted (view paper)
    23-Sep-2008 12:20 An Application of Design of Experiments in a Real Lift Test Rig
    A well known Basque lift manufacturer wishes to increase the comfort of their produced lifts, in order to gain a competitive advantage over the competence. Therefore, they wished to study the guiding system, which is the responsible of transmitting vibrations to the cabin avoiding having good comfort for passengers. Consequently, a lift test rig was constructed to simulate the real situation.

    The test rig was a device specially designed to carry out the experiments on it. It was built in order to have results extremely correlated to the real situation. The device is 5 metres height, but only 2.3 metres is the useful trip for the lift. And it was constructed in order to make possible the modification of many variables, such as loads, speed, off-centres, guides narrowing, etc.

    The experimental goal was to know about the influence of many factors (parameters) on the lift comfort. So the research team chose the Design of Experiments (DoE) techniques as the best techniques available to maximize the information that we would obtain from the experiments.

    The first option was to analyze the influence of many factors using to screen a factorial fractional design. However, when we were planning the experiments aroused many complications which were dealt with different literature solution. We finally chose a split-plot design with two dummy factors simulating a three-level factor.

    This article shows the steps carried out to plan and analyse the DoE problem, giving special importance to the planning steps along every phases of the DoE application.
  • Spc tools for short production runs - overview and case study from Polish industry

    Authors: Dr Eng. Agnieszka KUJAWIŃSKA, MSc. Eng. Michał ROGALEWICZ
    Affiliation: Poznan University of Technology, Faculty of Mechanical Engeneering and Management, Institute of Mechanical Technology, Div. of Production Management
    Primary area of focus / application:
    Submitted at 29-Apr-2008 15:48 by Michal Rogalewicz
    Accepted (view paper)
    22-Sep-2008 10:55 Spc tools for short production runs - overview and case study from Polish industry
    Many companies feel they cannot utilize statistical process control (SPC) charts because their average product run length is too short. This situation leads statisticians and quality managers to develop, encourage to use and apply statistical tools for short production runs. In Poland these SPC tools are not so popular, especially in small companies. It is a result of very low awareness of existing such methods or distrust to using statistical methods in case of short production runs. Because usage of SPC methods is sometimes a requirement for subcontractors from the side of big clients, Polish companies are not as competitive as foreign companies more frequently using SPC tools for short production runs.
    The purpose of this article is to describe the control charts (e.g. difference-from-nominal) and indices for providing statistical control of a short-run process.
    A case from Polish machining industry is presented.
  • Optimal strip-plot experimental plans for two-stage batch processes

    Authors: Heidi Arnouts (1), Bradley Jones (2), Peter Goos (1)
    Affiliation: (1) University of Antwerp, (2) SAS Institute Inc.
    Primary area of focus / application:
    Submitted at 29-Apr-2008 16:05 by Heidi Arnouts
    Accepted
    24-Sep-2008 09:00 Optimal strip-plot experimental plans for two-stage batch processes
    The cost of experimentation can often be reduced substantially by forgoing complete
    randomization. Examples of designs with a restricted randomization are split-plot
    and split-split-plot designs, which are commonly used in industry when some ex-
    perimental factors are harder to change than others. Another, lesser known type
    of experimental design plan is the strip-plot experimental design, also known as
    the strip-block experimental design (Miller, 1997). Strip-plot con gurations are an
    economically attractive design option in situations where the process under inves-
    tigation consists of two distinct stages, and it is possible to apply the second stage
    to groups of semi- nished products from the rst stage. They have a correlation
    structure similar to row-column designs and can be seen as special cases of split-lot
    designs (Mee and Bates, 1998; Butler, 2004). In this contribution, we present an
    algorithm for the construction of optimal strip-plot designs. Several examples will
    be discussed.