中文摘要：

网络分析法（ANP）是一种能够有效处理复杂决策问题的多准则决策方法。然而现有ANP文献在对收益、机会、成本、风险（统称为BOCR）评价值综合集成时会因评价值之间的不匹配而可能得出错误的方案排序结果和绩效表示。为克服上述缺陷,本文基于数据包络分析理论提出一种新的针对ANP/BOCR评价值的综合集成方法,使用摆幅置权区间估计方式反映出了决策者在判断BOCR相对权重时所面临的不精确性和模糊性。实例验证结果表明,所提方法对BOCR评价值的处理更符合实际情况。

英文摘要：

The analytic network process（ANP） is an efficient multiple criteria decision method to deal with complex decisions.However,the existing ANP with consideration of BOCR merit factors does not always yield correct ordering of alternatives and indication of profitability,due to the fact that different hierarchies or networks produce weights on different derived ratio scales which are usually not commensurate.To overcome the above drawbacks,this paper proposes a new ANP meatsynthesis method with respect to BOCR merits,based on data envelopment analysis.When estimating each weight for four BOCR factors,we adopt the means of swing weighting to reflect the inaccuracy and fuzziness of a decision-making.Applied to a specific example,the suggested approach is validated to be suitable to real world ranking problems for complex systems.