Deep Reinforcement Learning Task Scheduling for IoV Edge Nodes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In the context of Internet of Vehicles (IoV), efficiently scheduling data processing tasks across distributed computing resources is challenging due to high communication costs and the need for load balancing, while ensuring processing efficiency and reducing power consumption in dynamic network environments.
Innovation Solution
A method and apparatus for task scheduling using deep reinforcement learning, which involves dividing tasks into subtasks, building state data sets for each subtask, and using a pre-trained deep neural network to determine optimal node allocation based on CPU requirements, data storage, execution time, and CPU clock frequency, employing the Asynchronous Advantage Actor-Critic (A3C) algorithm to minimize average execution time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If centralized processing in cloud computing center is used, then computing capability is strong, but communication cost increases
Solution Approach 1:
The patent segments the centralized cloud computing system into a distributed edge computing network. Computing tasks are divided and distributed to multiple edge nodes (vehicles, roadside units, base stations) instead of concentrating all processing in a central cloud center. This segmentation reduces communication overhead by processing data closer to its source while maintaining strong aggregate computing capability across the distributed network.
2Speed
If deep reinforcement learning model is trained offline, then scheduling decision speed is fast, but model adaptability to dynamic environments decreases
Solution Approach 1:
The patent performs preliminary training of the deep reinforcement learning model offline using historical task data and environmental information. This preliminary action prepares the model with general scheduling strategies before deployment. The pre-trained model can then make fast scheduling decisions in real-time while retaining adaptability through its training on diverse scenarios, balancing speed and adaptability.
Data Source
AI summary
Disclosed are a method and apparatus for task scheduling based on deep reinforcement learning and a device. The method comprises: obtaining multiple target subtasks to be scheduled; building target state data corresponding to the multiple target subtasks, wherein the target state data comprises a first set, a second set, a third set, and a fourth set; inputting the target state data into a pre-trained task scheduling model, to obtain a scheduling result of each target subtask; wherein, the scheduling result of each target subtask comprises a probability that the target subtask is scheduled to each target node; for each target subtask, determining a target node to which the target subtask is to be scheduled based on the scheduling result of the target subtask, and scheduling the target subtask to the determined target node.