A Research On Bidding Decision Model Of IT Project Based On Fuzzy Set And Rough Set Theory

With the development of the strategy to pull the industrialization by the informationization in China, more and more enterprise focus on IT project and invest so many resources and energy on it. Informationization has become a mainly approach to enhance core competence for enterprises. However, there are so many risks in implement of IT project and statistic data shows that less than 30% of IT projects succeed in achieve the expected goals. Introduced in bidding and competition in IT project, we can fairly choose the best contractor, and it helps to control implement risk from the beginning and guarantee project success.IT project bidding evaluation is more difficult in that it has a wide rang of unique characters: often referring a great investment and more complicated evaluation indexes, the close bidders'ability, the grate correlative degree and low diaphaneity of information, the high dynamic of communication, and there are many indexes that can't be quantitatively measured. Therefore it seems a nonstructural question, for which the experience, knowledge and preference of the decision-maker play an important role in decision-making. As the complexity of decision process, it is believed that hardly this problem could be resolved perfectly by only one theory. Most traditional research focus on qualitative analysis, they seldom pay attention to applying quantitative analysis and evaluation model in bidding practice, and therefore are subjective and casual to some extend.Aiming at the problems mentioned above, firstly this paper constructs a general bidding evaluation index system of IT project, according to analysis of characters in IT project bidding, and thus give a knowledge denotation system and data analysis through Rough set theory. After an elementary cut and filter of bidding by rules mining of history similar project, we introduce a multilayer model of IT project evaluation based on fuzzy set and rough set, which integrates the advantage of the two theories. The subjectivity of weight choice is avoided by analysis of the significance of attribute based on rough set; the complexity of computation is reduced by attributes reduction of inconclusive indexes and elementary filter of bidding based on RBR (Reasoning by Rule). Then the combination of a method of fuzzy comprehensive evaluation can integrate all experts' opinions effectively and eliminate inconsistent factors during the marking process of experts. Finally, an empirical study is conducted to validate the model mentioned above.