Abstract: In the knowledge network of industrial cluster, the exploratory knowledge and exploitative knowledge will be sent from the knowledge supply source, through the knowledge intermediary to the users, and then form the feedback information to the knowledge source. In this paper, the knowledge closed-loop system based on the network power is studied, the differences about knowledge pricing and profits in different decision models are compared. The method of game theory and numerical simulation is used, the different decision-making behaviors under the different power networks of knowledge entity are compared. The result shows that (1) in each model, knowledge profits decreases with the increase of the preference coefficient,(2) in the decentralized decision mode, the income of knowledge source and the intermediary is related to the power, the more power, the greater the corresponding profits, and(3) the centralized decision model is the optimal decision mode, it can make the price lowest in both of exploratory knowledge and exploitative knowledge, and the system can reach the highest profits. However, this model can only be achieved through coordination mechanism. Thus the effective coordination mechanism is designed to achieve the centralized decision model and to provide beneficial enlightenment to the industrial cluster knowledge management practice. The conclusions would be helpful to enrich the knowledge closed-loop theory and popularize the application of cluster knowledge closed-loop system in reality.

Key words: industrial clusters, network power, ambidextrous innovation, knowledge closed-loop, knowledge pricing