Learning-based implementation method for adaptive dynamic heterogeneous network selection strategy

Abstract

The invention discloses a learning-based implementation method for an adaptive dynamic heterogeneous network selection strategy. The method comprises the following steps: step 1, performing multi-agent-based network selection scene modeling; step 2, executing a dynamic heterogeneous network selection strategy algorithm for each terminal user, with the algorithm comprising a network selection process and a network evaluation process, respectively constructing a candidate base station set and a standby base station set, and if the candidate base station set is not null, taking the base station with the maximum available bandwidth as a current optimal access base station; after the selection process is completed, a network evaluation process is started to be executed, namely, the quality of the base station is evaluated by the user i; and for each available base station k, calculating and detecting whether the base station can meet own specific requirements or not. According to the invention, better network performance such as load balance of a network end, user revenue and total network bandwidth utilization efficiency can be achieved; in addition, under the condition that the user does not obey the strategy, the system still has good robustness.

Description

technical field [0001] The invention relates to the technical field of wireless network resource management, in particular to a dynamic heterogeneous network selection method. Background technique [0002] With the rapid development of network technology, the network structure is developing in the direction of heterogeneity and intelligence. It is imperative to integrate multiple wireless access technologies together. The significant increase in the number of end users of mobile terminal smart devices and their bandwidth requirements have become two major characteristics of the new network environment. [0003] However, while bringing benefits, two features of the new network environment also introduce new challenges. One of the challenges is that in the case of dynamic changes in the network environment and wireless resources, end users are in a heterogeneous network environment with multiple end users, multiple network technologies, multiple application devices, and mult...

AI Technical Summary

Problems solved by technology

Although there are already many algorithms for network selection, they are implemented from various network performance perspectives and using various technologies, but there are still two defects in the following aspects: First, the demand for network information is too large

The acquisition of this information requires a high cost, and some information cannot be obtained at all in the actual network environment

Second, poor adaptability to dynamic network environments

These tasks may have good system performance when there is sufficient network status information or the network environment remains stable. However, when the environment information is missing and network resources change dynamically, it cannot cope well with such situations.

Images