روش شناسی مبتنی بر گسترش عملکرد کیفی فازی برای دستیابی به ملزومات انتخاب نرم افزار شرکت

In many software acquisition methods, functional software requirements are resolved, but non-functional requirements are more or less deliberately put aside. A large body of research exists on the necessity of handling specific non-functional requirements as major drivers in the software development process. However, prior research does not provide adequate support for managing non-functional requirements in the software selection process, and suggests a unique technique and methodology for identifying the selection criteria. This paper presents a fuzzy quality function deployment approach for determining which of the non-functional requirements reported by earlier studies are important to a company’s software selection decision based on and integrated with its functional requirements. The solution provided in this study not only assists decision makers in acquiring software requirements and defining selection criteria, but also supports determining the relative importance of these criteria. An actual case in Audio Electronics of Turkey’s electronic industry demonstrates the feasibility of applying the proposed framework in practice.

During the acquisition of large-scale software systems, effective and efficient management of user requirements is one of the most crucial issues (Karlsson, 1997). Software that lacks the appropriate functionality (what the system does) and non-functionality (how the system behaves with respect to some observable attributes like performance, reliability, etc.) carries the risk of failing to meet user needs. In many software acquisition methods functional requirements are resolved, but non-functional requirements (NFRs) are more or less deliberately put aside (Karlsson, 1997). Representing functions in terms of inputs, processes and outputs may be relatively straightforward. However, representing and modelling NFRs such as usability or reliability may be less straightforward, since such requirements are often inherently problematic. NFRs play an important role for software architecture decisions because the selection of the main architecture structure is mainly driven by these requirements (Kruchten, 2003). There exist a lot of software architecture design methods discussing the necessity of handling specific NFRs as major drivers (Hofmeister et al., 2007). On the other hand, an increasing trend of purchasing enterprise software packages (e.g. enterprise resource planning (ERP) systems, customer relationship management (CRM) systems, and supply chain management (SCM) systems) has also highlighted the need to take NFRs seriously in the software selection process. In this situation, requirements need to be elicited from the end-user for commercially available software, and issues including the representation of the requirements, their prioritisation, level of detail, and the relation to functional requirements become more complex. Since the end-user has no influence on the functionality provided by enterprise software packages, there is little use in providing detailed functional requirements. It is more important to identify the constraints that software must meet and the overall quality of the product. Much research has addressed the different sets of NFRs (Erol and Ferrell, 2003, Illa et al., 2000, Kunda and L., 1999, Ochs and M.A., 2000, ISO/IEC Standards 9126, 1991, Wei et al., 2005 and Wei and Wang, 2004), what criteria are used in the software selection process, and also what criteria are the most important for firms (Baki and Çakar, 2005, Bernroider and Koch, 2001, Chau, 1995, Keil and Tiwana, 2006, Kontio and J., 1995, Maiden and Ncube, 1999 and Montazemi et al., 1996). However, these empirical investigations do not provide adequate support for managing NFRs, and suggest a unique technique or methodology for identifying the selection criteria in the software selection process. We have recently developed an integrated decision support system dealing with qualitative and quantitative criteria for enterprise software selection (Şen, Baraçlı, Şen, & Başlıgil, 2009). In this paper, we focus on the second step of this six steps methodology, identifying the criteria for enterprise software selection. The first goal of this paper is to identify possible NFRs reported by earlier studies. The second goal is to present a systematic procedure for determining which of these are important to a company’s software selection decision in accordance with its functional requirements. We develop a fuzzy quality function deployment (QFD) approach that focuses on translating functional requirements formed with linguistic variables into non-functional criteria. The solution proposed in this study not only assists decision makers in defining non-functional selection criteria on the basis of company’s own conditionality, but also supports determining the relative importance of these criteria. Consequently, the paper presents the end-user’s perspective rather than that of organizations that are developing enterprise software packages (suppliers). The rest of the paper is organized as follows: Section 2 describes briefly the QFD process and its use in software requirements analysis. In Section 3 we present, in detail, the proposed fuzzy QFD approach. An actual case in Audio Electronics of Turkey’s electronic industry is described to demonstrate the proposed approach in practice.

This study proposes a fuzzy QFD approach to help decision-makers in the software selection process by managing NFRs. This quantitatively-based method supports both a systematic definition of non-functional criteria based on the functional requirements and their prioritization. A hierarchical criteria structure, including non-functional criteria such as quality attributes, technology and socio-economic factors, is provided by organizing the factors addressed in prior studies. The identified criteria are used in this approach as the non-functional selection criteria template. The QFD serves as a formal method to determine the order in which these criteria should be satisfied in a company’s software selection decision according to its functional requirements. The proposed approach provides the flexibility to include any specific criteria that the team may wish to consider in any other situation. The use of fuzzy-logic enables the decision makers to eliminate, or at least contain, the problems stemming from the subjective and ambiguous nature of their information. By applying our proposed approach to Audio Electronics, the functional requirements and non-functional criteria for ERP software selection are clearly defined and prioritized. This will enable the decision-makers to examine the strengths and weaknesses of software alternatives by comparing them with respect to appropriate criteria and sub-criteria in the further steps of the software selection process. Identifying appropriate criteria instead of using all criteria not only eliminates or reduces the risk of making inappropriate selection decisions, but also decreases the number of comparisons and the related computational effort. Hence, the use of our approach can reduce time and effort in making software selection decisions, as a result of decreasing the number of criteria through taking into account the current business situation. During the application of this approach, software support needs to carry out all calculations. We develop VBA-based macros in Microsoft Excel and transfer the results to spreadsheets for easy computations. Using the detailed checklist for obtaining the functional requirements ensures that needed requirements are not overlooked. Although the main goals of the proposed method are achieved by applying it to Audio Electronics, more practical applications should be employed on the method to address its feasibility. It should be noted that further verifications should be carried out on the proposed approach to ensure the validation of the inputs. We can conduct a survey involving software experts to identify the relative occurrence of important/unimportant factors in preparing non-functional criteria template. As the number of requirements increases, the method becomes more complex to solve, even using Excel. For future study, we can generate guidelines to help clarify decision making processes and develop decision support or an expert system to make fuzzy QFD-based calculations more rapidly.