关键词: 混合多属性决策, VIKOR, 改进离差最大化法, 概率语义集

Abstract: Hybrid multiple attribute decision making(MADM) problems have broad applications in the fields of economy, management and social science, etc. The existing methods to support hybrid MADM are more and more common which can process many different types of information such as crisp, interval, fuzzy, hesitant fuzzy. However, hybrid information is often converted into the same form which leads to the loss of information in most of these methods. In addition, only few research concerned the uncertainty caused by random variable. Aiming to avoid any transformation step and take random variable into account, Modular Random VlseKriterijumska Opti-mizacija I Kompromisno Resenje (Mo-RVIKOR) is proposed which can break heterogeneous information into modules and process information in a straightforward way without unifying. Firstly, real numbers and random variable are used by experts to process quantitative evaluation information, probabilistic linguistic term set to process qualitative evaluation information. Secondly, the weights of attribute are determined by the improved deviation maximization method. Finally, Mo-RVIKOR is adopted to rank the alternatives. This method can effectively handle certain or uncertain mixed evaluation information, and a case study of C2B customized service quality assessment of one company is presented to illustrate the effectiveness of the proposed approach.

Key words: hybrid multi-attribute decision-making, VIKOR, improved deviation maximum method, probabilistic linguistic term set