不确定环境下整合风险与属性绩效交互影响的多属性决策方法及应用研究

Multiattribute decision making (MADM) methods evaluate decision alternatives based on their values with respect to each attribute. The risks that may affect the achievability of the attribute values are seldom taken into consideration. Risks may come from various sources. Their different characteristics and interrelationships will have different impact on the attribute values, which however, have not been addressed by existing MADM-based methods. To solve decision problems under uncertainty with enhanced risk awareness and improved accuracy of decision-making, this project proposes a novel approach to integrating interrelated risks and attribute values under the MADM framework. The approach will address unsolved issues in MADM under uncertainty by (a) identifying the risk factors that affect the achievability of the attribute values, analyzing their interrelationships by constructing fuzzy cognitive maps, and evaluating their magnitude of impact by a modified failure mode and effects analysis under decision context, (b) developing a method to integrate the impact of risk factors and the attribute values based on their interaction mechanism for obtaining the achievability of attribute values under risk, and (c) identifying the interrelationships between attributes, and developing bi-capacities based attribute weighting and aggregation methods for the evaluation and selection of decision alternatives based on the principle of satisfaction. Empirical studies on leading e-waste recycling companies are conducted to illustrate how the approach works. The approach represents an original contribution to the methodological development of MADM-based methods. In practice, it provides e-waste recycling companies with a proactive mechanism for resolving decision problems under risks and uncertainty.

传统的多属性决策方法通常根据备选方案的属性绩效来评价方案，较少考虑影响属性绩效达成的风险，也未考虑不确定环境下风险的不同来源、特征，以及风险间的交互作用对属性绩效的影响程度。本项目创新的提出了整合风险与属性绩效交互影响的多属性决策方法，旨在提高决策者对风险的认知能力和决策的准确性。本项目首先通过识别影响属性绩效达成的风险因子，分析其特征要素和交互关系，构建基于模糊认知图、优化理论和失效模式影响分析理论的风险影响力评价模型；其次，通过分析风险因子和属性绩效间交互作用的机理，建立风险对绩效的影响指数，取得整合多种风险影响力的属性绩效达成度；在此基础上，通过识别属性间交互作用，设计基于双极容度的属性赋权法与集结算法，建立基于满意原则的方案评价与选择模型并求解。最后，以我国处于转型期不确定环境下的电子废弃品回收处理企业为对象，对所提理论与方法进行验证。本项目可丰富多属性决策领域的理论、方法及应用。

多属性决策是现代决策科学的一个重要组成部分。由于决策环境的不确定性，客观事物本身的复杂性和人类认知的局限性往往会对决策结果产生影响，对于决策中风险信息、模糊信息、以及决策要素交互影响关系的处理是该领域的研究热点。目前的多属性决策方法通常根据备选方案的属性绩效来评价方案，较少考虑影响属性绩效达成的风险，也未考虑不确定环境下风险的不同来源、特征，以及风险间的交互作用对属性绩效的影响程度。更少有研究探讨决策方案间的交互影响对决策结果的影响。为了提高决策者对风险的认知能力和决策的准确性，本课题以构建整合风险与属性绩效交互影响的多属性决策方法为主，探讨了风险间的交互影响，属性间的交互影响，风险对属性绩效间的交互影响，以及决策方案间的交互影响，创新的开发出一套考虑决策要素交互影响关系的多属性决策方法。为验证所提方法理论的有效性与实用性，本课题研究了我国电子废弃物回收处理行业在转型期的不确定环境中所面临的实际决策问题，通过所建立方法分别探讨了回收企业策略规划、运营决策制定、提升措施管理、发展阻碍因素应对、消费者对基于互联网的创新回收模式的行为意向分析等急需解决的实际问题。研究结果为回收企业提出决策支持方案与战略规划路径，为政府相关管理部门提出政策建议，对促进电子废弃物回收行业的可持续发展具有应用价值和指导意义。本课题为决策者解决在风险和不确定环境下的决策问题提供了一套主动决策机制，研究成果丰富了多属性决策领域的理论和方法。