复杂系统ANP-BOCR立体网络结构建构新方法

表达复杂决策问题的传统ANP-BOCR网络结构(即综合考虑收益(B)、机会(O)、成本(C)、风险(R)四类ANP(网络分析法)子网络结构所形成的新型网络结构)不仅忽视了隶属于不同B,O,C,R子网络中的准则集和准则之间所存在的依赖关系,而且没有科学反映O,R子网络内部因素随预期时刻不同所呈现出的动态变化性。此外,传统ANP-BOCR网络结构针对系统内部准则(集)的确定方法也缺乏科学严密性。为克服上述缺陷,基于DEMATEL(决策试行和评价实验室)和动态决策思维,提出一种全新的ANP-BOCR立体网络结构建构方法。相对于现有ANP-BOCR网络结构,新方法突破了传统ANP-BOCR固定不变的准则、方案结构体系,构建出立体维度框架下的动态决策网络结构,并且在确定系统内部准则(集)时运用DEMATEL体现出更强的科学性。实例对比验证结果表明新方法是科学合理的,能够为进一步完善ANP-BOCR方案优选方法提供前期理论基础。

A Construction Method of Spatial Network of ANP-BOCR for Complex Systems

For the complex decision-making issue, it often could be expressed by a new network structure of analytic network process(ANP) with benefits, opportunities, costs and risks(BOCR) merits, which is called conventional ANP-BOCR structure. However, the structure has the following disadvantages. First, the dependence relation among criteria(criteria cluster) lying different BOCR sub-network is ignored. Second, in opportunities and risks sub-network, the dynamic characteristics of the system elements(i.e., system elements and their priorities could with vary with different expected time) cannot be reflected scientifically. Third, the determination way for criteria(criteria cluster) in the network structure of conventional ANP-BOCR is preformed poor due to lacking of well scientificity. To overcome drawbacks abovementioned, in this paper, a novel constructing approach for spatial network structure of ANP-BOCR is given based on decision making trial and evaluation laboratory(DEMATEL) and dynamic decision idea. Compared with exiting ANP-BOCR structure, the new approach breaks the limit of strict structure system consisting of fixed alternatives and criteria, and gives a dynamic decision structure under spatial dimension framework. Furthermore, the DEMATEL is applied to determine the key criteria, which shows well stringency. Finally, based on real case data, the new method is validated to be reliable by contrasting to the conventional ANP-BOCR structure of current case, and it builds a foundation for further research on improving the alternative synthesis method for BOCR weights.