Decision support with wavelet neural networks in R&D project termination decision

Research and development (R&D) project termination decision is an important and challenging task for organizations with R&D project management Current research on R&D project management mainly focuses on project selection decisions. Very little research has been done on the termination decision of R&D projects In this paper a wavelet neural network system for assisting managers in deciding whether to abandon an ongoing R&D project at various stages of R&D is presented. The system is basically a multi-layered wavelet-based neural network which combines the function of time-frequency location of wavelet transform and self-studying of neural networks. A supervised gradient descent-based back-propagation algorithm is employed to adjust the parameters of the wavelet neural network by using the training patterns. It has also shown by the modeling and pattern recognizing results in terms of termination decisions of fifty R&D projects that the method possesses reinforcement learning properties and universalized capabilities. With respect to modeling and termination decision of R&D project, which has the fact that the evaluation criteria are hardly ever determined by conventional approaches such as statistical analysis, the method is available.