Research On Evaluation And Decision-Making For R&D Projects In Uncertain Environments

R&D (research and development) as an important part in the investment ofenterprises is the source of main competition of enterprises and the motivity ofdevelopment of enterprises. Recently, the investment for R&D activities has beenincreased by almost all countries in the world. This indicates that the research on theoption value of R&D project, R&D decision-making models and R&D performancemeasurement is of significance in theory and practice.The main contributions of this dissertation are organized as follows.(1) Random fuzzy theory is introduced to evaluate the option value of R&Dproject. In the classical researches on the option value of R&D project, the waitingtime until arrival of new information was assumed to be a random variable, whichbelongs to exponential distribution, and the jump amplitude is assumed to be arandom variable which submits to a weibull distribution and investment costs andfuture cash flows are considered as deterministic values, and the geometric Brownianmotion is applied to describe the properties of movement process of underlying assetof R&D project. The point estimation and interval estimation for the unknownparameters in the distributions are gained through Statistic method. However, inrealistic environments, it is very difficult or impossible to obtain enough data becauseof shortage of enough trials. The random fuzzy theory provides a powerful tool tohandle this kind of problems. In this dissertation, the waiting time until arrival of newinformation and jumpamplitude are depicted as random fuzzy variables, andinvestment costs and future cash flows are depicted as fuzzy valiables. What's more,jump process is applied to describe the properties of movement process of underlyingasset of R&D project.(2) The random R&D decision-making problems without rivalry are studied inthis dissertation. In the past R&D decision models, the interarrival times betweenjumps were exponentially distributed. In this dissertation, the interarrival times areassumed to be stochastic variables, which have general distributed functions.Furthermore, to get the stopping time of the R&D project, three kinds of models ——expected value models, chance-constrained programming models anddependent-chance programming models —— are proposed. The simultaneousperturbation stochastic approximation (SPSA) algorithm based on random simulationis designed to solve these models.(3) The R&D decision-making problems in random fuzzy environments arestudied. The randomness and fuzziness often appear at the same time during theprocess of actual R&D decision-making. Under this circumstance, the singlestochastic process theory cannot solve these problems. This dissertation introduces therandom fuzzy theory into the process of R&D decision-making. According to thegoals of the decision-makers, three kinds of models —— expected value models,chance-constrained programming models and dependent-chance programming models—— are proposed. Furthermore, the SPSA algorithm based on random fuzzysimulation is designed to solve the above models.(4) The R&D performance measurement is discussed, and an analytic hierarchyprocess (AHP) method based on fuzzy simulation is proposed to evaluate R&Dperformance. The meaning of R&D performance measurement is given and theperformance measurement system is put forward. Finally, the countermeasure forenhancing R&D performance of enterprises is provided.