Abstract:

Concurrent execution and knowledge transfer among projects are fundamental characteristics of multi-project management， especially for the new product development （NPD） project. The accumulated knowledge and experience can be transferred from one project to other projects in the overlapping and concurrent process， thus improving the performance of the program. However， existing research on multi-project management don’t take the influence of knowledge transfer and learning among projects into account， as well as the clustering analysis for multi-project. Using the dependency structure matrix （DSM）， knowledge transfer theory and two-stage clustering algorithm， a new multi-project management method is proposed. It attempts to solve three key problems： 1） how to measure the knowledge maturity and acquirement model of projects； 2） how to analyze the interaction relationship among projects through knowledge transfer and learning； 3） how to cluster projects into programs via the improved two-stage DSM clustering algorithm.Hence， the knowledge transfer and learning process among projects is analyzed from the perspective of overlapping which is the basic characteristic of NPD projects. First， the knowledge transfer and learning among projects is analyzed in the overlapping and concurrent process. Then， the knowledge maturity model is built for transferring knowledge projects and knowledge acquirement model for receiving knowledge projects in the concurrent process using the DSM method. Next， the interdependency strength models resulted from knowledge transfer among projects is established. Additionally， by taking account the “knowledge transfer-learning” process， quantitatively measure model for the shorten time is built using the improved learning curve model. Further， a two-stage program DSM clustering algorithm with maximizing “Added Internal and External Average Crashed Time Ration” criterion is defined as an objective function. Finally， an industrial example is provided to illustrate the proposed model and method. The results indicate that the clustered program can enhance the dependency strength and shorten the time of the program， reducing the total coordination costs， and significantly improve the performance of program management.

Key words: project management, program, overlapping concurrency, knowledge transfer, dependency structure matrix （DSM）, clustering