Solving the Short Video Assignment Problem via Federated Learning and Group Multi-Role Assignment

The Short Video Assignment Problem (SVAP) is one of the main problems short video platforms face. This article suggests converting SVAP into a Many-to-Many Assignment Problem, which the Group Multirole Assignment can specify, to obtain the best video assignment plan as much as possible. GMRA is a sub-model of the E-CARGO, transforming the relationship between video classes and servers into the relationship between agents and roles. To improve the privacy and security of data as much as possible, this paper uses the method of Federated Learning (FL) to process data. The Entropy Weight Method (EWM) and Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) methods evaluate video classes in multiple dimensions and indicators to assess their matching degree with servers. Finally, the linear programming method was used to solve the problem, and the best group performance is 985.86639