A Dynamic Concrete Scheduling Method Based on Multi-Level Reinforcement Learning and Agent Simulation

By constructing ABM and DDQN networks through multi-level reinforcement learning and agent simulation, the deviation problem of traditional concrete scheduling in dynamic environments is solved, achieving efficient and accurate concrete scheduling decisions and improving transportation efficiency and demand response capabilities.

CN120409194BActive Publication Date: 2026-05-26SOUTHWEST JIAOTONG UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-03-28
Publication Date
2026-05-26

AI Technical Summary

Technical Problem

Traditional concrete scheduling methods struggle to respond in real time to sudden changes in demand and traffic congestion in dynamic environments, and lack micro-level traffic behavior modeling, resulting in large deviations in scheduling schemes, low computational efficiency, and difficulties in multi-objective optimization.

Method used

A concrete dynamic scheduling system is constructed using a method based on multi-level reinforcement learning and agent simulation, including ABM micro-traffic simulation, hierarchical reinforcement learning framework, DDQN policy network and dynamic scheduling mechanism, and scheduling decisions are optimized through a two-layer fully connected feature extractor.

Benefits of technology

It achieves efficient scheduling in dynamic environments, reduces transportation time prediction bias, improves the accuracy of micro-traffic modeling, and enhances fuel consumption optimization rate and demand satisfaction rate.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120409194B_ABST
    Figure CN120409194B_ABST
Patent Text Reader

Abstract

This invention relates to the field of dynamic concrete scheduling, specifically to a dynamic concrete scheduling method based on multi-level reinforcement learning and agent simulation. The technical solution includes: constructing an ABM (Automatic Traffic Management) micro-traffic simulation environment for dynamic concrete scheduling; designing a hierarchical reinforcement learning framework for dynamic concrete scheduling; training a DDQN (Directed Decision-Making) policy network; and triggering the dynamic concrete scheduling mechanism. This invention utilizes a dynamic decision-making mechanism based on hierarchical reinforcement learning to respond to environmental changes in real time, improving the accuracy of micro-traffic behavior modeling. Through refined simulation of two-way single-lane scenarios using ABM, including random deceleration and overtaking decisions, this invention reduces the deviation rate in transportation time prediction. Micro-behavioral modeling reduces transportation cost estimation errors and improves fuel consumption optimization efficiency. This invention introduces a two-layer fully connected feature extractor to effectively capture the nonlinear features in the production-transportation coupling relationship. This invention is applicable to dynamic concrete scheduling.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Automatic driving lane changing decision and control method based on DQN series reinforcement learning algorithm

    CN115158349A

  • Online car-hailing scheduling method based on hybrid hierarchical reinforcement learning

    CN115222251A