针对现有概率权重非参数确定方法存在的两方面缺陷,即因假设概率权重仅与概率有关而没有考虑概率权重受概率与结果共同影响的问题和没有针对决策者主观判断不准确性予以判断偏差有效控制的问题,通过引入两两比较思维构建了反映概率和结果对概率权重复杂影响的基元前景两两比较交叉判断新模式,并在此基础上运用误差控制优化技术提出了基元前景价值确定模型和前景优劣排序模型.数值模拟分析结果表明,应用前景优劣排序模型得到的前景排序与“假定能够事先知道、能够真实反映决策者在有限理性下实际选择行为”的真实排序具有高度一致性,证实了其克服现有非参数法固有缺陷的有效性.
There are two drawbacks in current nonparametric methods for eliciting probability weights (PW). One is no consideration of PWs being influenced by both probabilities and outcomes due to the viewpoint of PWs being influenced only by probability. The other is that there is no means of effectively controlling subjec- tive judgment errors due to less consideration of the inaccurateness of subjective judgments. To overcome these shortages, a new mode of basic prospect cross judgments (BPCJ) for reflecting complex relations of probabili- ties and outcomes on PWs is presented by the thought of pairwise comparison. Based on BPCJs, a value-deter- mination model for basic prospects and a ranking model for prospect choice, are developed through an error- controlling optimization technique for subjective judgments. The analysis results of numerical simulation show that the prospect ranking derived by the application model is highly consistent with the ranking of the decision- maker' s real preferences supposed to be known prior to the simulation, and that the models can well overcome the drawbacks as mentioned above.