The Improved DEA Model for R&D Project Selection under Risk Condition
Xiao-yan GU,Feng HE
DOI: https://doi.org/10.16381/j.cnki.issn1003-207x.2018.07.006
2018-01-01
Abstract:For many enterprises, Research and Development (R&D) projects play a vital role in improving competitiveness, and the R&D project is the impetus of enterprise survival and development.The R&D project selection is a process of logical analysis and comprehensive judgment on an investment system with complex multi factors, and it is one of the most important and complex problems in the project management field.Meanwhile, R&D project selection is closely related to enterprise strategy development and future trends.Compared with general projects, high innovative R&D projects have greater uncertainty, which are not only reflected by the technical factors of projects themselves, but also by the factors of market changes and unexpected competitors.The high uncertainty of R&D projects brings high risks.It is also the case that R&D projects are affected by market, technology and emergency risks during the life cycle.Market risk is the uncertainty of potential R&D project cash flow, technology risk is the probable loss caused by technology uncertainty, and emergency risk is the impact of unexpected events on R&D project cash flow.The existed risks make R&D project selection more challenging.In general, R&D projects have large investment and long life cycles, which are useful for enterprise competitive advantage and future growth.It will affect enterprise development for several years whether to choose the R&D project suitable for enterprise strategy and whether to response to risks during R&D process.Therefore, it has great theoretical and practical significance to study R&D project selection under risk condition.In this paper, drawing on the risk characteristics associated with R&D projects and enterprise strategy, the R&D project selection model based on efficiency sorting is constructed and presented.To begin with, the effectiveness of R&D projects and the relative advantages compared to alternative projects are analyzed;Secondly, focus is put on the drawbacks of DEA model, combining with the concept of BSC.According to enterprise strategy, the relative importance of individual R&D project evaluation indicator is different.During the evaluation, it is necessary to introduce weight change information into the DEA method, thus a R&D project selection model is built based on DEA method with guarantee domain.Balanced Score Card (BSC) is a multiple performance evaluation system.It can transform enterprise tasks, decisions and internal departments into a variety of interrelated goals, and then decompose theses targets into multiple indicators.Based on the idea of BSC, the enterprise strategy is introduced into the DEA model, which can allocate the change range of input indicator weights under the same score card.It shows that DEA model is improved by introducing the real option theory and constraining the proportion of R&D project inputs, then the efficiency sorting of R&D projects is realized;Finally, the proposed model in this paper is verified through an actual example, with transverse contrast and sensitivity analysis.The R&D project selection differences between using the proposed DEA model and the standard DEA model in an aerospace enterprise are compared, and the corresponding operation code is developed by MatLab7.0.The results reveal that the model presented in this paper has higher efficiency division compared with the standard DEA model, and it could capture sensitively the evolution process that R&D project input and output changes impact upon efficiency.Above all, facing enterprise strategy, the selection model of multiple R&D projects in an enterprise is discussed under risk condition.Based on real option theory, constraining DEA model weight assignment with the BSC concept, R&D project investment factors and enterprise strategy are combined.Taking into account risks and uncertainties, the weight changes are limited within a special range, the standard C2 R model is improved, and the sort results validity is made.The proposed model has higher discrimination.At the same time, it can capture, with precision and accuracy the process of input and output parameters on efficiency influence.The model proposed in this paper can effectively sort the efficiency of R&D projects, and effectively solve the selection of multiple R&D projects.