Multi-attribute dynamic two-sided matching method of talent sharing market in incomplete preference ordinal environment

Abstract With the development of the mobile Internet, big data and cloud computing, talent sharing has emerged and developed in the labor market. As a new economy mode of the Internet, the talent sharing can match surplus labor with relevant needs companies and achieve the maximum interests for two sides with the help of the network platform. In this paper, we deeply investigate the matching problem of talent sharing in incomplete preference ordinal environment. Unlike the traditional two-sided matching method, which directly ignores the incomplete preference ordinal, we firstly fill the incomplete preference ordinal by using the collaborative filtering algorithm. In the two-sided matching of talent sharing, we pay full attention to the individual differences of the seekers and the solvers. Because of individual differences, the seekers and the solvers have different preferences for different decision attributes, i.e., the attribute priority matrices. At the same time, considering the psychological expectations of the seekers and the solvers, we construct the satisfaction degree matrices based on the prospect theory. Given the constraints on the attribute priority and the satisfaction degree, the bi-objective optimization model for multi-stage dynamic decision-making is established. Moreover, with the aid of dynamic decision-making process, we can obtain more matches when meeting the psychological expectations of the matching subjects. Finally, a case study of the talent sharing platform Upwork is given to illustrate the validity of our proposed method.