應用模糊多準則決策方法建構網站服務品質之評估模式

Abstract

本研究主要以兩個實證研究來說明如何運用模糊多準則決策方法解決資訊科技產業中的網站服務品質之評估議題，本研究透過整合因素分析法與權重技術法來建構評估模式。研究結果顯示，本研究不僅可幫助網站管理者從顧客的角度瞭解公司的競爭態勢，同時更能激勵管理者提供適當的網站服務品質滿足顧客的需求。 在旅遊網站的研究中，對許多旅遊相關公司而言，旅遊網站服務品質(travel website service quality , TWSQ)之評估是一個多準則決策問題，旅遊網站服務品質(TWSQ)有一個策略性質的重要性。但在實際環境中，由於評估人員主觀的認知所隱含的模糊性，使得決策者面對一個多變的旅遊網站服務品質(TWSQ)通常具有模糊性的。因此，此時利用傳統方法與精確的單一數值可能無法合理地描述評估者對網站服務品質的認知。為了克服此類真實的決策評估問題所具有的模糊性，本研究之主要目的在於建構一套有效的階層式模糊多準則決策模式來評估分析旅遊網站服務品質(TWSQ)。在本研究中有兩種多準則決策方法被使用，即模糊TOPSIS(FTOPSIS) 與簡單平均加權法(SAW)等，以推導出各評估方案之綜合評估值並據以進行優劣排序，同時比較這兩種方法之間的異同之處。因此，本研究結合因素分析法、模糊層級分析法(FAHP)、模糊TOPSIS(FTOPSIS)與簡單平均加權法(SAW)來建構一套有效且客觀的旅遊網站服務品質之比較分析模式。在本研究中，各準則的語意評估值及語意權重值， 皆利用正三角模糊數來表示其語意變數的隸屬函數之關係。最後以一真實的旅遊網站案例來驗證，階層式模糊多準則決策方法較傳統的簡單平均加權法(SAW)合適。本研究結果顯示這兩種方法所得到的排序順序是相同的，但模糊TOPSIS有比較好的辨別度。 在入口網站的研究中，近年來，由於網際網路的快速發展及全球資訊網的興盛，不僅促使相關的網際網路使用者與服務業者急速地增加，更使得入口網站在商業競爭中愈來愈顯現其成長能力與經濟效益。但在真實的環境中，當網際網路使用者面對多樣化的入口網站服務品質時，此時由於人類的主觀認知與判斷上，通常具有相當程度的不明確性與模糊性。然而，大多數的評估準則之間經常存在相依性之情況；這些問題是無法利用傳統的加法型與精確的單一數值來合理地表達網際網路使用者對入口網站服務品質的認知。為了克服這些問題，本研究之主要目的在於嘗試藉由網際網路使用者的角度，並同時考慮任何兩個評估準則之間經常存在相依性特性之情況，提出一套非加法型的評估模式來處理入口網站服務品質之評估問題，此模式是結合因素分析法、層級分析法、模糊積分與簡單平均加權法等方法。最後以一案例，實際說明本評估模式之可行性；研究結果顯示入口網站經營業者若能適當地改善其入口網站服務品質，以便滿足顧客需求是很重要的。

The research introduces the fuzzy multiple criteria decision making (FMCDM) method and explains the process of these approaches by two empirical real cases of website service quality evaluation issue for information technology industry. The study will combine MCDM methods with factor analysis and weighting techniques to construct the hybrid evaluation models. Moreover, the results not only help website providers understand the competition status from the view points of online customers, but also motivate website providers to provide appropriate levels of service quality in response to e-customers’ needs. In this research, the evaluation of TWSQ is a multiple criteria decision making (MCDM) problem and has a strategic importance for many travel-related commercial companies. In practical environments, the evaluators face a variety of travel websites service quality that is often vague from human beings’ subjective judgments. These problems may not be properly evaluated by the conventional models, and crisp data cannot represent precisely the e-customers’ perceptions of TWSQ. In order to overcome the situation of uncertainty among the evaluated criteria in real problem, the main purpose of this study is to present a hierarchical fuzzy MCDM model for assessing the service quality of travel websites. In this study, the two MCDM methods (i.e., Fuzzy Technique for Order Preference by Similarity to an Ideal Solution (FTOPSIS) and Simple Additive Weighted (SAW)) are utilized to derive the final appraisal values from which one can choose the best option and determine the preferred order according to these values. Meanwhile, this study also provides a comparison analysis of the above two methods. Therefore, this thesis combines factor analysis, fuzzy analytic hierarchy process (FAHP), FTOPSIS, and SAW methods to construct a proper and effective comparison analysis. In this study, the ratings and weights of these criteria are assessed by linguistic (natural language) variables expressed in triangular fuzzy numbers. Lastly, an empirical example of travel websites in Taiwan is illustrated to show that the FTOPSIS methods are preferred to the conventional additive methods (e.g., SAW method). The results of this study show that the two methods have identical rank overall, but the FTOPSIS methods have a better distinguishing capability. Due to the rapid growth of usage of Internet and WWW, there has been a fast growth in Internet users and related service industries in the past few years. Portal websites have shown its growing ability and economic benefit on business competition. In practical environments, Internet users face a variety of service quality of portal websites that are often inaccurate or vague from human beings’ subjective judgments. However, most criteria are interdependent or interactive, so these problems cannot be correctly evaluated by conventional additive measures and crisp data cannot represent precisely the perception of portal website service quality. To overcome these problems, this study aims to propose a hybrid non-additive model that integrates factor analysis, AHP, fuzzy integral and the simple additive weighted (SAW) method to deal with evaluating the service quality of portal websites with consideration of inter-dependent characteristics between any two criteria from Internet users’ viewpoint in the context of fuzzy environment. In order to show the practicality and usefulness of this model, an empirical case of the Taiwan portal website is demonstrated to highlight the procedure of proposed model at the end of this thesis. The results show that providers of portal website improve properly the service quality of portal website for satisfying of customer’s needs is the most important.