大数据环境下的管理决策创新研究

From decision support system (DSS) to business intelligence (BI), management decision making technology has been developing for more than thirty years. Its basic concept is to support decision activities by analyzing structural data from enterprise or organization. Currently, Big Data has changed human’s decision making behavior and method; that is, the “model-based” hypothesis test and reasoning decision making becomes the “data-driven” related analytic decision making. Its heterogeneous data also bring a new challenge for data-driven management decision. This proposal calls for investigation of research frontiers between Big Data complexity, Big Data mining and intelligent knowledge by systematically applying Management Science, especially Optimization theory and methods. First, it will explore, under cloud computing environment, how to use optimization techniques to transform semi-structured and unstructured data into structured data for Big Data mining. The project will investigate the mathematical structure of complexity and uncertainty of different types of semi-structured and unstructured data to build a basic theoretical framework of Big Data. Second, it will examine how Big Data mining influences the relationships of “data heterogeneity” and “decision making heterogeneity” in terms of intelligent knowledge. This intends to find out the reasons why Big Data mining can use intelligent knowledge to change the decision making structure and how people understand the impacts of the change on consequence of management decision making. Third, the project will study, in management decision practice, how Big Data analytical tools like data mining impact the decision-making quality and satisfaction of executive decision team of organizations, and identify how organizations can really make good use of the proposed new set of big data analysis methods in their daily management decisions by exploring how CDO of an organization should do to ensure the application success. Finally, the project will use three Big Data test banks: Xinhua 08, Shanghai Transportation Information System and Tencent (QQ) social networks to conduct experimental studies of the proposed theory and methods for their efficiency, extension, approximation and effectiveness. It may demonstrate the future direction of how to use Management Science to handle Big Data problems. The purpose of this project is to establish a new interdisciplinary research field between Management Science and Big Data.

管理决策从决策支持系统到商业智能经历了三十多年。当今，大数据不仅改变了人类决策行为，而且给管理决策带来了新的挑战。本项目将运用管理科学手段，特别是最优化理论和方法研究大数据复杂性、大数据挖掘与智能知识交叉领域的科研课题。首先，在云计算环境下，探讨如何用最优化技术把非结构化和半结构化数据转化成结构化的数据。从异构数据类型入手，考查其复杂性、不确定性特征描述的数学结构，建立统一大数据基础理论模型。然后，利用“智能知识”探讨大数据挖掘对“数据异构性”与“决策异构性”的影响，以揭示大数据挖掘和智能知识导致的决策结构变异对管理决策的影响。其次，研究在大数据管理实践中的决策质量与满意度问题。最后，以我国大数据行业重要示范:“新华08”金融信息平台、上海交通信息平台和腾讯（QQ）社交网络平台作为实证平台，检验新方法的有效性、可扩展性、解的近似性和实用性。本项目力争开辟新的管理科学与大数据交叉学术领域。

本“大数据环境下的管理决策创新研究”（71331005）重点项目旨在运用管理科学手段，特别是最优化理论和方法研究大数据复杂性、大数据挖掘与智能知识交叉领域的科研课题。首先在云计算环境下，探讨如何用最优化技术把非结构化和半结构化数据转化成结构化的数据；其次从异构数据类型入手，考查其复杂性、不确定性特征描述的数学结构，建立统一大数据基础理论模型；然后利用“智能知识”探讨大数据挖掘对“数据异构性”与“决策异构性”的影响，以揭示大数据挖掘和智能知识导致的决策结构变异对管理决策的影响；然后研究在大数据管理实践中的决策质量与满意度问题；最后，以我国大数据行业重要示范:“新华08”金融信息平台、上海交通信息平台和腾讯（QQ）社交网络平台等作为实证平台，检验新方法的有效性、可扩展性、解的近似性和实用性。在本项目进行期间，不仅在大数据理论研究方面取得重要进展，已经发表论文84篇，其中期刊论文53篇，SCI检索论文34篇，会议论文31篇，完成专著5部，获得3项发明专利和4项软件著作权，主办了11次国际学术会议，相关研究成果已成功应用到新华社、腾讯、上海交通信息平台、西安市公安局、春雨医生等机构。另外，本项目还资助培养了研究生92名，2名博士后。