Network attribute matrix standardizing and weight calculating method based on service sensitive intervals

Abstract

The present invention provides a network attribute matrix standardizing and weight calculating method based on service sensitive intervals, comprising: step S1, constructing a decision matrix X; step S2, setting a service sensitive interval for each decision attribute; step S3, standardizing the decision matrix X to generate a standard decision matrix; step S4, calculating a mean value and a standard deviation of respective decision attributes; and step S5, calculating objective weights of respective decision attributes. The network attribute matrix standardizing and weight calculating method based on service sensitive interval in the present invention can accurately determine the objective weights of respective network attributes and lay the foundation for precisely selecting a wireless network through using a multi-attribute decision algorithm, so that a terminal can match an optimal or suboptimal network among numerous wireless networks, thereby improving the service experiences for users of the terminal and the utilization of network resources.

Description

technical field [0001] The invention relates to the technical field of wireless network selection and multi-attribute decision-making, in particular to a network attribute matrix normalization and weight calculation method based on service sensitive intervals. Background technique [0002] Multi-attribute decision-making, also known as limited-scheme multi-objective decision-making, refers to the decision-making problem of selecting the optimal alternative or ranking the solutions under the consideration of multiple attributes, which is an important part of modern decision science. The theory and method of multi-attribute decision-making are widely used in many fields such as engineering, technology, economy, management and military affairs. [0003] In recent years, with the development of mobile communications and the popularization of smart terminals, applications and services have become more abundant. People's requirements for service experience are getting higher and ...

AI Technical Summary

Problems solved by technology

[0009] (2) Normalization does not consider the difference in attribute value space

In this way, when the attribute value distribution is relatively scattered, it can avoid the difference in the value range of different attributes. However, when the attribute value distribution is relatively concentrated, the normalized attribute value still cannot reflect the specific physical meaning of the attribute value. That is, if a certain attribute of different candidate networks takes a large value or a small value in a concentrated manner, the values ​​after normalization are basically the same, which will inevitably lead to deviations in the final decision result.

Images