Weighted interval type 2 fuzzy rule based system approach and applicaton in FMEA

During recent years, the globalization is increasing the competition in the markets. Customer requests in terms of product quality, cost and flexibility and complexity is also increasing. These are forcing companies. In this thesis, a comprehensive model that can respond to shorthcomes of traditional HTEA have been targeted for improvement. With this goal, we evaulated a model for HTEA which (i) is taken into account uncertainty arising from the use of expert opinion, ( ii) can be considered the variability in the data received from experts with different knowledge and experience, (iii) can be handled the traditional FMEA model is blocked error arising from the formulation , (v) RPN value get different results based on different possibilities with the weighting process is performed. The behavior of fuzzy rule based system techniques is expressed with a language that is easily interpretable by humans. The usage of linguistic terms in the fuzzy approach turns the FMEA much more applicable than traditonal form, because the experts can assign a more meaningful value for the factors considered. Fuzzy logic allows imprecise data usage, so it enables the treatment of many situations in decision making. Moreover, the studies about FMEA considering fuzzy approach use the experts who describe the quantitative data and qualitative information about risk factors O, S, and D by using the fuzzy linguistic terms. But FRBS, can not completely handle groups' variety uncertainties. In order to handle it we used Interval Type 2 Fuzzy Rule Base System. And then for getting more flexible model, we weighted the rules. So we developed a Weighted Interval Type 2 Fuzzy Rule Base System model. The proposed model is used in the process of a national marketing company that aims to create a new coffee chain in order to identify potantial failures may face. The company has been operating successfully for many years in different sectors and in the short time period, a rapid growth in this area is also planning. However, the coffee industry is a new area for the company and requires strategic steps towards branding. During this major investment process, the types of errors can be predicted and potential errors can be prevented before giving any material and moral damages. It is very important for the company. In order to consider in detail the errors, Phillip Kotler 's 7P marketing (product, price, place, people, process, physical facilities, promotion) criteria were used. Errors that may occur with regard to these criteria determined by experts in a systematic way, and this error types for severity (S), the probability (O) and the detectibility (D) risk factors were evaluated. We used five expert for this application. The experts obtained 41 errors associated with the process. And they determined affects which is failure modes could bring, evaluated these errors according to the probability severity risk factors. And then, experts deterimend the rules for all the combinatons of severity, occurence and detectibility risk faktors. Developed AAT2BKTS model is solved using Matlab IT2FLS interface. Application, also explored and solved, using with traditional FMEA model and type 1 fuzzy rule base system. Risk priority ranking compared to this type of 41 error for 3 different models. As a result, according to the output of AAT2BKTS model, some sort of failure modes found different from traditional and rule based system models solutions. AAT2BKTS models found H30 failure mode has the highest priority. This failure is "The place designing and decoration, is not to appeal to customers" Moreover, the failure mode, were also explored for their criteria they belong. Using the Guided Rules Reduction System (GRRS), the RPN value of the reduced rules (with 43, 55, 61, 84) and unreduced rules results deviations are calculated. The performance of GRRS models solutions for 43,55 61 and 84 rules were explored. Although the 84 rules deviation occurs slightly lower than the 61, 21 rules were defined less. This found that more advantageous by experts.