Base Station Load Balancing via Dynamic Algorithm Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional rule-based algorithms for base station load balancing fail to adapt to sudden changes in communication environments, leading to imbalanced loads and reduced wireless network performance.
Innovation Solution
A system and method that dynamically switches between rule-based and machine learning algorithms based on network loading performance, using a performance evaluator and algorithm switching device to allocate connections between base stations and user devices, with machine learning algorithms being employed during sudden environmental changes to balance loads.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a rule-based algorithm is used for base station load balancing, then the loading performance is good under known or predictable communication environment, but the algorithm cannot adapt to sudden changes in communication environment
Solution Approach 1:
The patent implements dynamic algorithm selection by introducing a performance evaluator that continuously monitors loading performance metrics and switches between rule-based and machine learning algorithms based on current communication environment conditions. This dynamic adaptation allows the system to maintain reliability under predictable conditions while gaining adaptability when environmental changes occur suddenly.
Solution Approach 2:
The system changes the algorithmic parameter (from deterministic rule-based to probabilistic machine learning) based on evaluated performance metrics. When the communication environment becomes unpredictable or changes suddenly, the system transitions to a machine learning algorithm that can adapt to new patterns, thereby resolving the contradiction between reliability and adaptability.
2Power
If the number of user devices served by the base station exceeds a certain number, then the data throughput required exceeds the hardware capacity, but reducing the number of served devices reduces network coverage
Solution Approach 1:
The patent segments the user device population into different groups based on their connection relations with base stations. By using machine learning algorithms to optimize this segmentation and allocation, the system can efficiently distribute devices across multiple base stations, ensuring that no single base station becomes overloaded while maintaining comprehensive network coverage across the service area.
Solution Approach 2:
The performance evaluator continuously monitors loading performance and provides feedback to the algorithm switching device. This feedback mechanism enables real-time adjustment of connection allocations, allowing the system to dynamically balance load distribution across base stations while maintaining optimal network coverage, thus resolving the contradiction between throughput capacity and coverage area.
Data Source
AI summary
The present disclosure provides a base station load balancing system including a communication network, a performance evaluator and an algorithm switching device. The communication network includes a base station and a plurality of user devices. The performance evaluator is connected to the communication network and evaluates a loading performance of the communication network in a time period according to a first loading data of the communication network in the time period. The algorithm switching device is connected to the performance evaluator and selects the algorithm used by the communication network as a selected algorithm according to the loading performance. The communication network, with the selected algorithm, allocates the connection relation between the base station and plurality of user devices according to the first loading data. The selected algorithm is a machine learning algorithm or a rule-based algorithm.