Joint Research Conference
June 24-26, 2014
Complex quality of Products and Evaluation of Distance from Target Variant
For computer aided design purposes is essential to calculate the distance between target product (with required properties) and various variants evaluated experimentally, calculated virtually or selected from database system. Let the target product be characterized by a set of values for m utility properties and variant product is characterized by the set values for the same utility properties. The distance dj between target fabric and variant fabric is calculated. The main shortcoming of these distances is the assumption of symmetry. For the case of calculation distance for computer aided product design it is in fact distance characterizing loss of quality. It means that target product has prescribed quality and aim is to construct variant product with similar “or better” quality. Higher distance then should correspond to smaller quality only. The purpose of the paper is to describe the construction of utility value U based criterion suitable for expressing pseudo-distance in the area of computer aided products design. This criterion can be simply created based on the complex quality indices. The degree of quality (complex criterion) is often expressed as utility value U [1]. These are such properties that make it possible for the product to fulfill its function. Utility value aggregates then in some certain way partial quality properties. The purpose of the paper is to describe the construction of this type criterion suitable for computer aided products design. The application of U and corresponding pseudo-distance is demonstrated on the example of characterization of the effect of catalysts on the quality of crease resistant finished fabrics under laboratory conditions. The target fabric is selected according to the requirements to minimize the influence of finishing on the loss of mechanical properties and surface abrasion on the one hand and improve the recovery angles on the other hand. The application of so called Bootstrap simulation for calculation of basic statistical characteristics is discussed. The program COMPLEX written in MATLAB is briefly mentioned.