论文部分内容阅读
在多属性决策问题中,不同的属性权重会产生不同的评价结果。由于实际问题的复杂性与不确定性,决策者对于属性权重的确定也存在不确定性。这些不确定既来自现实问题的复杂性和可变性,也来自决策者选择的模糊性与随机性。目前已有的研究主要是将不确定的权重信息转化为相对确定的信息(如转化为区间数等),硬性地消除了不确定,从而给决策结果带来较大风险。本文从方案排序的视角出发,研究在权重空间下,方案的占优关系和排序的稳健性。首先,定义了占优矩阵用于刻画不确定权重信息下方案两两比较的占优关系;其次,分析了方案的排序区间,即在所有可能存在的权重组合下,方案的最好排序和最差排序。然后,定义了方案的全排序排序概率,并且给出了排序概率的计算方法。进而,我们给出了方法的决策步骤和实施过程。最后,本文将该方法应用到某远洋集团的港口评估当中。
In the multi-attribute decision-making problem, different attribute weights will produce different evaluation results. Due to the complexity and uncertainty of practical problems, policymakers also have uncertainties about the determination of attribute weights. These uncertainties come not only from the complexity and variability of real problems, but also from the ambiguity and randomness of decision makers’ choices. At present, the existing research mainly transforms the uncertain weight information into relatively definite information (such as converting into interval number), and eliminates the uncertainty, which brings a greater risk to the decision-making result. From the perspective of program ordering, this paper studies the dominance relationship of programs and the robustness of ordering under the weight space. First of all, the dominance matrix is defined to describe the dominance relation of each pair of alternatives under uncertain weight information. Secondly, the ranking interval of the scheme is analyzed, that is, the best ranking and the best choice of the scheme under all possible combinations of weights Poor sorting. Then, the total sort ordering probability of the scheme is defined, and the method of calculating the ranking probability is given. Furthermore, we give the method of decision-making steps and implementation process. Finally, this paper applies this method to the assessment of a COSCO’s port.