Implementation of Trusted Traceability Query Using Blockchain and Deep Reinforcement Learning in Resource Management.

Area of Science:

• * Agricultural Science
• * Computer Science
• * Information Technology

Background:

• * Existing agricultural product traceability systems suffer from significant shortcomings in tracking goods and maintaining quality.
• * Current systems lack robust mechanisms for trusted queries and comprehensive information management.
• * The integration of advanced technologies is crucial to overcome these limitations.

Purpose of the Study:

• * To develop a trusted traceability system for agricultural products using blockchain technology.
• * To address the task scheduling problem in cloud computing resource management for traceability systems.
• * To propose and evaluate a novel resource management algorithm based on deep reinforcement learning.

Main Methods:

• * Application of blockchain technology to establish a new agricultural product traceability model.
• * Expansion of the task model for resource clusters in cloud computing.
• * Utilization of the deep Q-network algorithm from deep reinforcement learning for task scheduling optimization.
• * Performance analysis of the proposed algorithm against classic scheduling algorithms like Shortest Job First (SJF).

Main Results:

• * The developed blockchain-based traceability system achieved an accuracy of 99% in the first experimental group and 98% in the second.
• * Overall traceability accuracy across eight experimental groups consistently exceeded 98%.
• * The proposed deep Q-network-based resource management algorithm demonstrated superior performance compared to traditional algorithms.

Conclusions:

• * The blockchain-based system offers a reliable solution for agricultural product traceability, enhancing trust and quality management.
• * The novel resource management algorithm effectively optimizes task scheduling in cloud environments for traceability applications.
• * This research provides valuable insights into applying reinforcement learning for building advanced traceability systems.