Huang (2025) Actor–Critic Deep Reinforcement Learning for Multi-Objective Intelligent Irrigation Scheduling: Algorithm and Edge-Cloud Management System

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Informatica
• Year: 2025
• Date: 2025-11-23
• Authors: Peng Huang
• DOI: 10.31449/inf.v49i14.11138

Research Groups

Information not provided in the paper text.

Short Summary

This study developed an intelligent agricultural irrigation scheduling algorithm and management system based on a deep reinforcement learning model, demonstrating a 12.7% improvement in water resource utilization and an 8.3% gain in crop yield through field trials.

Objective

• To construct a technical solution for agricultural irrigation that combines decision-making adaptability and resource utilization efficiency using a deep reinforcement learning algorithm and an integrated management system.

Study Configuration

• Spatial Scale: 35-hectare wheat–corn site, comprising 12 plots.
• Temporal Scale: 4 months (for field trials).

Methodology and Data

• Models used: Deep reinforcement learning model using an improved Deep Q-Network (DQN) combined with policy gradient fusion (DQN–Policy Gradient hybrid). The strategy function was optimized using the Time Difference (TD) method.
• Data sources: Multimodal data including soil moisture, evapotranspiration, and meteorological predictions collected by field sensing networks.

Main Results

• The system achieved 88.1% ± 1.7% water use, representing a 12.7% improvement in water resource utilization (n=30, p<0.01).
• It resulted in an 8.3% ± 1.2% yield gain (n=30, p<0.05), outperforming predefined thresholds.
• The multi-objective reward function successfully integrated crop yield (0.3 weight), water resource utilization rate (0.5 weight), and energy consumption (0.2 weight).

Contributions

• Provides a scalable technical path for intelligent management of agricultural water conservancy.
• Offers practical verification for the application of deep reinforcement learning in complex resource scheduling scenarios, specifically for agricultural irrigation.

Funding

Information not provided in the paper text.

Citation

@article{Huang2025ActorCritic,
  author = {Huang, Peng},
  title = {Actor–Critic Deep Reinforcement Learning for Multi-Objective Intelligent Irrigation Scheduling: Algorithm and Edge-Cloud Management System},
  journal = {Informatica},
  year = {2025},
  doi = {10.31449/inf.v49i14.11138},
  url = {https://doi.org/10.31449/inf.v49i14.11138}
}