مدل های تجزیه و تحلیل ریسک و تعیین درجه ریسک در توسعه محصول جدید : یک مطالعه موردی

This paper proposes a risk analysis model to determine the risk degrees of the risk factors occurring in product development processes. The model uses both fuzzy theory and Markov processes on a concurrent engineering (CE) basis. Fuzzy models determine the impact values of the risk factors, and Markov processes determine the probability of risk occurrences. The analysis model is used to compute the risk degrees by multiplying the probability of risk occurrences by the impact value. This study can be utilized for analyzing the influences of risk factors on product development projects and will contribute toward the development of a risk management framework (RMF) to defend against various risk factors. Implications and directions for future research are discussed.

The product development process determines whether a company survives or fails in competitive markets because a product life is generally determined by its market share (Fig. 1). In order to ensure sustainable competitiveness, innovative efforts to develop new products must focus on substituting existing products or entering new markets. As new products must be introduced to the market periodically, the development process is a critical strategic issue at business level because the product life-cycle is becoming shorter (Kim, 2003a and Kim, 2003b). However, about 80% of new product development (NPD) efforts have failed before project completion and more than 50% of the efforts have made no returns on the investment of money and time (Cooper, 2003). In other words, the product development process for new products is a complex and difficult business decision-making process because of the high capital investment required and exposure to low success probability. The critical explanations for the difficulty of the product development process are unexpected risks and their impact, and the inability of the firm to defend against those risks effectively and efficiently. This paper suggests a new systematic risk management framework (RMF), as shown in Fig. 2. RMF determines risk degrees for risk factors and total risk degrees of the product development project, and shows effective and efficient responding activities. Especially, RMF suggests a risk analysis model under a concurrent engineering (CE) environment. CE is an approach to link all functional areas such as manufacturing, financing and marketing with the design process (Savic and Kayis, 2006). There is a multidirectional exchange of information among all functional areas for better, easier, and more economical product development. Therefore, either a high degree of collaboration or a high concurrency level (CL) is desirable to construct the CE environment. Furthermore, the fluent information exchange under high CL enables the functional areas to handle the related risk factors more effectively and efficiently. The increasing difficulties of a product development project require a higher CL. In addition, a full understanding of the pitfalls and risks is required for successful CE implementation. The famous pitfalls include unobtainable-schedule, change-product-ineffective-team, requirements, business-as-usual vendoring, automate-everything, supplier dependent lead-time, teams unsupported by reward systems, lack of information technology support project development instead of process improvement, and discontinued change (Willaert et al., 1998). In this case study, the risk analysis model is used to determine CL on a CE basis, and the model quantifies the influence of a risk factor to the development project as an impact value. When a risk factor occurs in a functional area, process, department or project, its impact value or the amount of influence differs. Therefore, the impact value has to be defined precisely. This paper proposes the use of fuzzy theory for this purpose. Because a risk factor occurs probabilistically or stochastically, a Markov process can be adopted to determine the probability of occurrences for a given risk factor. Finally, the developed risk analysis model is used to compute the risk degrees by multiplying the probability of occurrences with the impact value.

This paper has proposed a new risk analysis model using fuzzy theory and Markov processes on a CE basis. We determined the project difficulty by using IDM. This level is then used to select the appropriate membership function in modeling the impact value of a risk factor, after which the frequency adjustment process is used to model the probability of risk occurrences. The impact value of a risk factor is calculated by the fuzzy theory, and the probability of risk occurrences is obtained by the Markov processes. Finally, the risk degree of the risk factor that will occur in a future project is computed by multiplying the probability of risk occurrences by the impact value of the target risk factor. This risk degree can be used for selecting the appropriate responding activities to remove or minimize the possible effects of the risk factor in the product development project. The possible responding activities shown in Table 7 can be suggested for the target risk factor used in the example. Few studies have investigated risk analysis for product development. Therefore, it is hoped that this study will stimulate wide-ranging discussions about risk management in product development on the CE basis. Traditional risk analysis has been used to evaluate a certain risk factor by calculating the probability of risk occurrences according to an expert's subjective point of view. However, the proposed risk analysis model can be applied for various risk factors that are realized in product development projects. Further study is required to construct a robust RMF. New algorithms must be developed to compute the total risk degrees of the entire system. Especially, all possible functions and algorithms must be investigated for optimized application to and development of Eqs. (9), (10), (11) and (12) in the near future. The algorithm should also include a method to determine the minimum total risk degree in a development project by using the proper decision-making criteria. Furthermore, the performance of the risk analysis model must be verified by application to real industries. Finally, a huge volume of accurate data from real industries is needed to develop a more reliable, Markov-based model.