Application of neural network in QFD matrix

Abstract

The major challenge for contemporary enterprises is to make products fulfill customers’ expectations. Conceptual design is the foundation of the development process, as it starts from the customer needs identification and decides about the whole product life cycle. Quality function deployment (QFD) helps to extract product characteristics from customer demands. Optimization of the product development process requires different product variant information at the early stage of product development. At the early designing stage, designers lack sufficient product information and have difficulty in determining it. The idea of the paper is to provide measurable engineering information for the quality function deployment method. For this purpose, a chosen artificial intelligence method was used. In the experiment, artificial neural network (NN) was applied. The results of analyses show that the intelligent estimation methods are useful and effective. The methods of estimation consist of four stages: goal setting, data acquisition, configuration of NN architecture, fulfilling of the QFD matrix. Finally, to illustrate the procedure of the chosen engineering characteristic estimation, a toothed gear box example was used.