Agility Evaluation Of Virtual Enterprise Based On Neural Network

At the beginning of the 21st Century, the world faces significant changes in almost all aspects, especially marketing competition, technological innovations and customer demands. Mass markets are continuing to fragment as customers become increasing demanding and their expectations rise. These developments have caused a major revision of business priorities and strategic vision. Companies have realized that agility is essential for their survival and competitiveness. This study further recognizes that no company possesses all of the resources required to meet every opportunity. Therefore, to achieve a competitive edge in the global market, the virtual enterprise (VE), which is in general the collaborative partnership between business partners in value chains, has become a key factor for survival in the competitive business environment. A VE is a temporary organization which is created according to a business opportunity and is dissolved when the business opportunity no longer exists. The core requirements for efficient collaboration among business partners are agility of the VE models to cope with the changing business environment. In addition, the VE is designed to increase competitiveness, to optimize resource utilization, to increase scale of the business, and to take advantages of the complementary capabilities of the business partners.Agility is the key characteristic of VE.This paper is aim at how to measure the agility.First of all, in order to figure out how to measure agility of VE member, principal of agility evaluation is presented. And then their criterion is addressed according to the principal. Base on it, evaluation modal based on Modular Neural Network (MNN) is put forward. It is including three parts. The Decomposition Decision Unit (DDU) using analytic hierarchy process (AHP) is used to provide training sample. The sample divided by the result of DDU is sent to each Sub-task Processing Unit (STU).By means of Bayesian regulari- zation BP neural network, STU calculated the final result. Conclusion Proce- ssing Unit (CPU) is used to keep the result fit for the reality. Simulation experiments show that the modal proposed in this paper is efficient and effective and is superior to non-modular methods due to its parallel network system structure. Furthermore compared with other optimized methods, Bayesian regularization BP neural network using in the modal kept the BP neural network advantages and overcome the over-fitting problems.The second problem be solved in this paper is how to measure Agility of VE. The relationship between VE and its member is analyzed. Base on it, using Bayesian Regularized Neural Network (BRNN) for Agility Evaluation of Virtual Enterprise is developed. A simulative example illustrates the usefulness of the proposed method. In contrast to the standard BP neural network, the result shows that BRNN overcome the over-fitting problems, it can be used to measure any size of VE.