Strategic Decision Model And The Key Factors Of Uncertainty Reasoning Research

A strategic decision of an enterprise is to predict and look to the future based on the present conditions, wherein not only current interests are taken into account, but also long-term developments and opportunities are further paid attention to. Success or failure of the strategic decision of the enterprise can directly affect a short-term development even a long-term development of the enterprise. The study of the paper aims at establishing a critical factor index model for the strategic decision of the enterprise by utilizing the management science and an artificial intelligence science.Firstly, the paper preliminarily establishes the strategic decision index of the enterprise by starting with the BSC (Balanced Score Card), four dimensions and a general special principle, thereby realizing balances of many aspects, such as a finance and a non-finance, a long-term objective and a short-term objective, an external and an internal, a result and a process, a management performance and an operating performance, and so on. Then the BSC index is converted into a strategic decision map of the enterprise.Secondly, the paper converts the strategic decision map based on the BSC into a Bayesian network model by utilizing the characteristics that a causal relationship exists in both the strategic decision map of the enterprise and a Bayesian network.After that, the paper extracts hidden variable nodes by utilizing a hidden variable theory according to a priori knowledge, thus the problems of complex network structure and chaotic levels can be improved with the join in of the hidden variable nodes.Afterwards, by studying a learning algorithm of the Bayesian network structure, the paper selects a K2algorithm with the fastest computation speed and the highest precision to perform structure learning to enterprise data, and rejects some non-critical nodes, thereby achieving the purpose of optimizing the strategic model of the enterprise. Finally, the paper proves a precision of the strategic decision of the enterprise by comparing the BSC index systems of a plurality of specialists in a plurality of fields at home and abroad.At the end of the paper, a Bayesian network visual management system is established by aiming at the characteristics that the Bayesian network is abstract and difficult to manage, wherein in a front end of the system, dragging and visual operations for the Bayesian network are realized by adopting the Javascript technology, and at the same time, in a back end of the system, across-platform extension of the network is realized by adopting an XML file format. The system in the paper has the advantages of convenient use and friendly operation.