ساخت خانه‌ي كيفيت در فرايند QFD با استفاده از داده‌هاي نامتقارن چندزباني

Building the house of quality in QFD process by unbalanced multi-granular data

در فرايند توسعه‌ي كاركرد كيفيت (QFD)، مرحله‌ي ساخت خانه‌ي كيفيت به‌طور قابل توجهي با ارزيابي‌هاي انساني روبه‌رو است. با توجه به تاثيرپذيري اين ارزيابي‌ها از عوامل روحي و رواني، بي‌توجهي به اين عوامل منجر به كاهش دقت و صحت تصميم‌گيري مي‌شود. در اين پژوهش ضمن معرفي و بررسي عوامل انساني مرتبط با ساخت خانه‌ي كيفيت، تاثير دخالت اين عوامل با سطح دانش و تجربه‌ي مختلف و مجموعه‌هاي كلامي چندزباني، ريسك‌پذيري با عملگر OWA و نگرش فرد با مجموعه‌هاي كلامي نامتقارن ارزيابي مي‌شود. در اين بررسي همچنين داده‌هاي كلامي نامتقارن چندزباني توسط مدل سلسله‌مراتب كلامي يك‌دست و متقارن شده و براي جلوگيري از هدررفت اطلاعات در حين تبديل داده‌ها، از نمايش 2-tuple براي ساخت خانه‌ي كيفيت استفاده مي‌شود. سپس ارزيابي‌هاي فني (HOWs) اولويت‌بندي شده و مهم‌ترين آن‌ها مشخص مي‌شود. در پايان يك مثال عددي طرح و با مدل پيشنهادي حل شده است. نتايج حاكي از عملكرد مناسب مدل پيشنهادي و ارتقا عملكرد در ساخت خانه‌ي كيفيت با تمركز بر رتبه‌بندي ارزيابي‌هاي فني (HOWs) است.

Quality Function Deployment (QFD) is a customer-oriented tool that aims to meet customer requirements in a better way. Indeed, this function can be described as a methodology that makes every process of product development transparent, starting from understanding the qualities demanded by customers all the way down to establishing quality planning and determining design quality. For achieving the first step of QFD, namely understanding the qualities demanded by customers, this process used House of Quality (HOQ). To that end, major elements of the process are vagueness of the customer’s assessments and their different level of experience. Since every customer has different knowledge level and experience about the subject under consideration, they use different fuzzy sets with different cardinalities fitting their levels. With regard to this situation, there would appear multi-granularity in information for which we need a procedure to make data homogeneous for solving this problem. In the real world, there exist problems that need to assess their variables with linguistic term sets that are not uniformly and symmetrically distributed; we call this type of linguistic term sets “unbalanced linguistic term sets”. In some cases, the unbalanced linguistic information appears as a consequence of the nature of the linguistic variables that participate in the problem as it happens usually for showing the attitude of their users. So, for dealing with this situation, we need a process to make information uniformly distributed. Between choosing different Brands, customers usually have some risks. So, for building an effective HOQ, we need a weighting model to consider this issue. Considering the whole of these limitations, we use Hierarchical Ordinal Model which makes the data homogeneous and balanced without any loss of information. Moreover, for avoiding any loss of information during conversion process, this model uses 2-tuple representation model. For considering the risk factor, we also propose the OWA weighting model that has numerous benefits. In the next step, the balanced homogenous information is used in 9-step HOQ building procedure, and the paper is concluded with the ranking of HOWs based on customer’s assessments.