The final goal of quality of service （QoS） guarantees is to assure high quality of experience （QoE） for users. We cannot control QoE, but it is feasible to use QoS control schemes at lower levels to keep QoE high....The final goal of quality of service （QoS） guarantees is to assure high quality of experience （QoE） for users. We cannot control QoE, but it is feasible to use QoS control schemes at lower levels to keep QoE high. We need to know the necessary QoS parameters values which need the minimum cost but could satisfy required QoE. An enhanced method is proposed to get the necessary QoS parameters which cost minimum but could satisfy required QoE well. It abstracts the problem of obtain necessary QoS parameters to be how to get the solution of linear regression equations considering cost coefficient. The method establishes equations utilizing principal component analysis and multiple regression analysis based on normalized data. This paper defines the cost of necessary QoS and introduces the calculation of minimum cost necessary QoS parameter by inverse or generalized inverse matrix operations. The numerical example proves the normalization to be effective and the calculated values are feasible.展开更多
文摘The final goal of quality of service （QoS） guarantees is to assure high quality of experience （QoE） for users. We cannot control QoE, but it is feasible to use QoS control schemes at lower levels to keep QoE high. We need to know the necessary QoS parameters values which need the minimum cost but could satisfy required QoE. An enhanced method is proposed to get the necessary QoS parameters which cost minimum but could satisfy required QoE well. It abstracts the problem of obtain necessary QoS parameters to be how to get the solution of linear regression equations considering cost coefficient. The method establishes equations utilizing principal component analysis and multiple regression analysis based on normalized data. This paper defines the cost of necessary QoS and introduces the calculation of minimum cost necessary QoS parameter by inverse or generalized inverse matrix operations. The numerical example proves the normalization to be effective and the calculated values are feasible.
