Graph Neural Network Radio Resource Scheduling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Next-generation wireless networks face challenges in efficiently scheduling communication resources for multiple devices due to increased demand and complexity, especially with the introduction of technologies like multi-user MIMO systems, which require cost-effective and flexible management of wireless access networks.
Innovation Solution
The use of artificial intelligence/machine learning models, specifically graph neural networks (GNNs), to allocate and schedule communication resources based on channel information and device attributes, including fairness and interference considerations, to determine the optimal allocation of communication resources among multiple communication devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If traditional scheduling techniques are used to schedule communication resources for multiple devices, then the scheduling process can be implemented with existing processors, but the complexity and latency increase due to increased demand and multi-user MIMO systems
Solution Approach 1:
The patent replaces traditional mechanical/computational scheduling algorithms with a neural network-based system. The neural network model processes channel information and device attributes to predict optimal resource allocations, substituting complex iterative optimization algorithms with a trained neural network that can make decisions more efficiently. This substitution reduces scheduling complexity while maintaining or improving allocation efficiency.
Solution Approach 2:
The patent transforms the scheduling problem by changing the input parameters and representation. Instead of using traditional scheduling algorithms that process raw channel state information directly, the system transforms this information into a format suitable for neural network processing, including normalized channel information, device attributes, and fairness/interference parameters. This parameter transformation enables more efficient processing while capturing the essential characteristics needed for optimal scheduling.
2Adaptability or versatility
If multi-user MIMO systems are introduced to increase data rate and connectivity, then the number of communication devices that can be served increases, but the scheduling algorithms become more complex and harder to scale
Solution Approach 1:
The neural network scheduling system provides universality by handling varying numbers of communication devices and different channel conditions through a single trained model. The model can accommodate multi-user MIMO systems with different configurations (number of antennas, users, data rates) without requiring retraining or algorithm redesign. This universal approach scales the system to serve increasing numbers of devices while maintaining consistent performance.
Solution Approach 2:
The system performs preliminary training of the neural network model offline using extensive simulations and historical data that capture various multi-user MIMO scenarios. During online operation, the pre-trained model simply needs to process new channel information and device attributes, avoiding the need for real-time complex optimization calculations. This preliminary action enables the system to handle large-scale multi-device scenarios efficiently during actual operation.
3Ease of operation
If conventional scheduling methods are used, then implementation is straightforward, but the latency in user scheduling increases due to computational requirements
Solution Approach 1:
The neural network model is trained in advance using offline data and simulations, capturing the optimal scheduling decisions for various scenarios. During real-time operation, the system only needs to process current channel information and device attributes through the pre-trained network, which is computationally lightweight compared to running complex optimization algorithms. This preliminary training action eliminates the need for time-consuming real-time optimization while maintaining implementation simplicity.
Solution Approach 2:
The system creates a computational model (neural network) that copies and learns from optimal scheduling patterns found in training data and simulations. Instead of solving the scheduling optimization problem from scratch in real-time, the model retrieves and adapts pre-learned optimal patterns, significantly reducing computation time. This copying approach maintains implementation ease while dramatically reducing scheduling latency.
Data Source
AI summary
A device may include a processor configured to provide input data that is based on channel information representative of attributes associated with one or more established communication channels of a plurality of communication devices to a trained machine learning model configured to determine a score for each of the plurality of communication devices based on the input data, wherein the score represents a likelihood of the respective communication device to be scheduled for a communication resource to perform a communication, determine, for the communication resource, one or more communication devices from the plurality of communication devices based on the determined score for each of the plurality of communication devices, and provide an output to schedule the communication resource for the one or more communication devices