On July 22, the National Industrial Alliance for Artificial Intelligence Applications (Power Sector in the Energy Field, abbreviated NIAIA-PS) released the compilation of exemplary achievements in "AI + Energy and Power". "Key Technologies and Applications of 'Mingyue', a Multi-Agent Collaborative AI Dispatcher for Ultra-Large Provincial Power Grids", jointly developed by Prof. Jianwei Huang and Prof. Junhua Zhao of the Center for AI for Science and Engineering at SLAI and the Power Dispatch and Control Center of Guangdong Power Grid Co., Ltd., was selected as an exemplary achievement in the "AI + Dispatch" category. Only two achievements were included in this category.

The compilation was exhibited for global AI experts to exchange and discuss during the 2026 World Artificial Intelligence Conference (WAIC), held in Shanghai from July 17 to 20.

Cover and contents pages of the "AI + Dispatch" case compilation

01 An Intelligent Approach to Large-Scale Grid Dispatching

Power dispatch is one of the most challenging scenarios for applied artificial intelligence, requiring systems to respond within seconds while ensuring that every decision complies with the power grid operation constraints/grid physical operation rules.

"Mingyue" is China's first AI dispatcher for a large provincial power grid and the country's first intelligent dispatching product system integrating perception, decision-making, and control. Built on architectures supporting multi-agent coordination and hybrid large-small model synergy, "Mingyue" has now been deployed across the Guangdong Power Grid.

02 Shifting from Monitoring to Rapid Control with Quantifiable Gains

• Surge in decision efficiency:The system can make rapid parallel decisions in complex, unpredictable complex grid contingencies, delivering a perception performance improved by 200 times and a 30-fold increase in decision-making speed compared to five years ago. During major disasters such as typhoons, it has helped avert cumulative economic losses of RMB 4.8 billion.

• Streamlined operations: By relieves dispatchers of the manual work of drafting operational procedures for complex conditions, the system cuts the time required to prepare operation order tickets by 95%.

• 24/7 Full-Coverage monitoring: Real-time surveillance of the operation status of millions of power equipment assets ensures safety and stable frequency and voltage across the grid.

• Quantifiable Economic & Social Returns: Since full deployment across the Guangdong Power Grid, the system has generated over RMB 130 million in annual incremental economic benefits while delivering significant social benefits in increasing renewable energy integration, improving efficiency, and ensuring safety.

03 Embedding AI into Physical Systems via Three Core Innovations

Power grids pose ongoing practical challenges real-world challenges and operational requirements; achieving trustworthy AI performance on ultra-large-scale physical systems, however, relied on SLAI's long-standing interdisciplinary research expertise. The project's three core innovations represent a direct translation of academic methodologies into industrial systems:

• "Comprehensive perception": Multimodal large models, VLM-BLIP, and knowledge graphs enable power-grid-wide holographic perception and semantic fusion.

• "High-Reliability Physical Constrained AI": Technologies including PINN-GNN and Physics-DPO are integrated to build a physics-causality dual-guided large model for power dispatch, constraining the model's outputs with physical operating constraints.

• "Fast Coordinated Control": The team proposed an innovative hierarchical multi-agent knowledge-embedded learning framework to enable hierarchical, parallel, and coordinated control across source-grid-load systems.

Details page for the achievement case, selected into the "AI + Dispatch" case compilation

04 Driving AI + Energy Forward Through Interdisciplinary Teams and Real-World Scenarios

Professors Jianwei Huang and Junhua Zhao have long led an interdisciplinary research team focused on multi-agent systems, smart grids, electricity markets, and low-carbon energy networks. Their sustained efforts are supported by platforms including the Center for Crowd Intelligence at the Shenzhen Institute of Artificial Intelligence and Robotics for Society (AIRS) and the Shenzhen Key Laboratory of Crowd-Intelligence-Empowered Low-Carbon Energy Network. The Center for AI for Science and Engineering at SLAI brings together relevant research groups through joint appointment and adjunct faculty mechanisms, driving the deep integration of AI methodologies with large-scale engineering applications.

In April 2026, SLAI jointly established the "AI + Energy" Laboratory with Guangdong Power Grid and AIRS. Leveraging this platform, the partners will advance fundamental research, technical validation, and scaled applications in power dispatch operations, new-type power systems, and low-carbon energy networks. The lab will also cultivate a sound industry-university-research collaborative ecosystem, in which technological research and industrial practice mutually reinforce and iterate through the output of products, patents, academic papers and high-end talents.

05 Sustaining Industry Recognition While Extending Global Reach

In 2025, the project received the Grand Prize of the Power Innovation Award from the China Electricity Council. In January 2026, it was again awarded the Grand Prize of China International Competition for AI Application Scenarios (CICAS)‌ and was named one of the National Typical AI Application Scenarios. This latest inclusion in the NIAIA-PS case compilation further underscores industry-wide recognition of the solution's technological excellence, proven effectiveness, and replicability.

06 Research Team Profiles

Prof. Jianwei Huang

Joint Appointment Professor at the Center for AI for Science and Engineering, SLAI

Presidential Chair Professor, Associate Vice President, and Dean of SSE at The Chinese University of Hong Kong, Shenzhen

IEEE Fellow and AAIA Fellow

His long-standing research focuses on network economics and crowd intelligence.

Prof. Junhua Zhao

Visiting Professor at the Center for AI for Science and Engineering, SLAI; Secretary-General of the "AI + Energy" Laboratory

Professor at SSE, The Chinese University of Hong Kong, Shenzhen

IET Fellow and AAIA Fellow

Named one of China’s Top Innovators in Intelligent Computing by MIT Technology Review.

His research focuses on smart grids, electricity markets, and low-carbon energy transitions.

07 About Industrial Alliance

The National Industrial Alliance for Artificial Intelligence Applications (Power Sector in the Energy Field, abbreviated NIAIA-PS) is a non-profit industrial alliance chaired by China Southern Power Grid Company Limited. Since October 2025, it has operated under the National Pilot Testing Base for AI Applications (Power Sector in the Energy Field) and now has more than 120 member organizations.

Related Links:（Click to view）

• Alliance Unveils "AI + Energy and Power" Case Compilation

• National Finals of the 3rd China's Innovation Challenge on Artificial Intelligence Application Scene Held in Suzhou

• Winners Announced at the National Finals of the 3rd China's Innovation Challenge on Artificial Intelligence Application Scene