Abstract:

To overcome the drawback of traditional alternative evaluation approaches, which only considered the fuzziness of evaluation information without considering the randomness of the information, an improved interactive multi-attribute decision making (TODIM) approach based on cloud model is proposed, in which the reference dependence and loss aversion behaviors of the decision maker are considered. Firstly, the cloud model is used to quantize the linguistic evaluation variable. Secondly, arithmetic mean cloud is defined. The dynamic weights of experts are determined by analyzing the similarity between the given expert evaluation cloud and the arithmetic mean cloud, based on the distance measure about the cloud model. Besides, the weighted decision matrix is constructed by integrating the cloud models of experts. Thirdly, the positive and negative ideal cloud models are defined. To maximize the overall closeness degree for all of the alternatives, a linear programming model is established to determine the criteria weights. Finally, the cloud-TODIM method based on the distance measure about the cloud model is used to calculate the overall dominance degree of each alternative relative to all others and obtain the ranking result of alternatives. A case study of cloud service providers evaluation and selection is presented to illustrate the effectiveness of the proposed approach.