Article Preview
Top1 Introduction
A Web Service is a software component which can be located by its identifier on the Web and its input and output parameters, which represent its functional information; however, there are other non-functional pieces of information which can characterize the service as response time, availability, cost, reliability, etc. This latter kind of information is called quality of service (QoS).
The problem of identifying the best candidate web services from a set of functionally-equivalent services is a Multi-Criteria Decision Making (MCDM) problem (Alrifai et al. 2010). Non-functional quality of service aspects are crucial criteria for enhancing Web services selection.
Unfortunately, the QoS got from service descriptors (WSDL) or services providers (UDDI) do not reflect the real quality of these services, they differ from a user to another, from a context to another and change dynamically in time according to several parameters (Yu et al. 2010), for example, the response time can vary according to the traffic of the network. It is important to propose an approach that not only returns a pertinent results but also to deals with the presence of uncertainty in QoS data (Hadjila et al. 2020).
To select a service among several candidate services, with similar functionality but with different qualities of services, the user cannot try all candidate services by himself, but, he can leverage historical invocation experiences performed by other users (Shao et al. 2007). Exploiting the experiences of other users allows to predict the QoS before using them, then, to help the user to choose the relevant services to his needs.
In the last years, the notion of Skyline is appeared as a new and popular method to select the best or the most relevant Web services (Benouaret et al. 2012) (Yu et al. 2010). It is considered as a promising direction which allows reducing the space of decisions of the user by preselecting the best one and prunes others. However, dominance relationship used by Skyline presents some drawbacks: i) the number of Skyline Web services cannot be controlled, either a huge or a small number which can be more or less informative for the user, ii) all attributes have the same importance, for example in financial services, security attribute is very important than other attributes, iii) Skyline returns generally a conflicting and incomparable results.
To deal with the large number of Skyline Web services candidates, the concept of K Representative Skyline (Lin et al. 2007) (Tao et al. 2009) has been proposed, which can return the k services that represent the full Skyline result (Chen et al., 2011). However, K-Representative Skyline allows generally returning incomparable and conflicting results, or the user often encounters some difficulties to select the best services which answer most to the user needs.
The idea to use MCDM methods to resolve some selection problems is not new. Analytical Hierarchy Process (AHP) and Fuzzy AHP is the most significant method being implemented as a Decision Making tool in many domains for selection of service, material, equipment, contractor, customer or supplier (Vesyropoulos et al. 2015; Godse et al. 2008; Chang et al. 1996; Kwong et al. 2002; Ayağ et al. 2006); Wu et al. 2007; Kumar et al., 2018; Shaygan et al. 2016). The main advantages of the wide usage of AHP is its ability to evaluate qualitative and quantitative measures, simplicity, flexibility, decomposing a problem into a hierarchy and its ability to check the consistencies of Decision-makers judgments (Huang et al. 2008; Li et al. 2013; Wan et al. 2008; Nefeslioglu et al. 2013; Kubler et al. 2016).