AI-Empowered Intelligent Electricity Markets: Forecasting, Optimisation and Autonomous Bidding

Primary supervisor

Hao Wang

Electricity markets are becoming increasingly dynamic as variable renewable generation, utility-scale batteries, distributed energy resources and flexible demand play larger roles in the energy system. Market participants must make decisions under uncertain prices, renewable generation, electricity demand, network conditions and market requirements while coordinating assets across multiple services and timescales.

These challenges arise for a wide range of participants. Utility-scale wind, solar and battery operators must forecast market conditions and coordinate generation, storage and bidding. Virtual power plants (VPPs) and aggregators must manage diverse portfolios of rooftop solar, household batteries, electric vehicles and flexible loads while respecting customer preferences and operational constraints. Market and system operators also require better analytics to understand market outcomes, system risks and the behaviour of emerging participants.

This project investigates how artificial intelligence, optimisation and intelligent agents can improve electricity-market analytics and operational decision-making. Depending on the student’s interests, the project may focus on forecasting, market participation, asset scheduling, portfolio coordination, decision support or market analysis.

Potential research topics include:

  • forecasting electricity prices, demand, renewable generation, reserve requirements or market volatility;
  • quantifying forecast uncertainty and its implications for market decisions;
  • scheduling and bidding utility-scale wind, solar and battery energy storage systems;
  • coordinating renewable generation and storage across energy and ancillary-service markets;
  • optimising the participation of VPPs and aggregators representing batteries, electric vehicles, rooftop solar and flexible demand;
  • developing price-taker or price-maker bidding strategies under uncertainty;
  • coordinating portfolios across day-ahead, real-time, balancing, reserve or frequency-control services;
  • developing reinforcement-learning or multi-agent reinforcement-learning methods for sequential market decisions;
  • designing AI agents that combine forecasting, reasoning, optimisation and market information to support or automate operational decisions;
  • applying large language models to interpret market information, support trading analysis or coordinate specialised analytical tools;
  • explaining and auditing AI-generated forecasts, bids and operational decisions;
  • examining strategic interactions among generators, storage operators, aggregators and other market participants;
  • detecting anomalous events, unusual bidding behaviour or changing market regimes; and
  • evaluating profitability, reliability, emissions, risk and consumer outcomes under alternative market strategies.

Each student will undertake a focused project within this broader theme. The selected topic can be primarily analytical, predictive or decision-oriented and can address either utility-scale assets or aggregated distributed resources.

#sustainability

Aim/outline

This project aims to develop and evaluate AI- and optimisation-based methods for analysing electricity markets and supporting the operation of renewable energy, battery storage, VPP and aggregator portfolios.

The project will typically involve:

  • Selecting a market problem: Identify a forecasting, scheduling, bidding, coordination or market-analysis problem relevant to a particular asset or participant.
  • Analysing real-world data: Explore historical electricity prices, demand, renewable generation, ancillary-service, weather or asset-operation data.
  • Modelling assets and markets: Represent relevant market arrangements, operational constraints, uncertainties and value streams.
  • Developing a methodology: Apply forecasting, mathematical optimisation, reinforcement learning, multi-agent systems, generative or agentic AI, explainable AI, or a combination of these methods.
  • Evaluating performance: Compare the proposed method with suitable benchmarks through historical-data experiments, simulations or realistic case studies.
  • Assessing practical value: Evaluate outcomes such as forecast accuracy, revenue, operational cost, risk, renewable utilisation, battery degradation, reliability and interpretability.

URLs/references

Some recent papers from the team.

1) Hao Wang, B. Zhang, Energy storage arbitrage in real-time markets via reinforcement learning, IEEE Power & Energy Society General Meeting (PESGM), 2018.

2) Muhammad Anwar*, C., Wang, F. de Nijs, Hao Wang, Proximal Policy Optimization Based Reinforcement Learning for Joint Bidding in Energy and Frequency Regulation Markets, IEEE Power & Energy Society General Meeting (PESGM), 2022.

3) J. Li^, C., Wang, Hao Wang, Optimal Energy Storage Scheduling for Wind Curtailment Reduction and Energy Arbitrage: A Deep Reinforcement Learning Approach, IEEE Power & Energy Society General Meeting (PESGM), 2023.

4) J. Li^, C., Wang, Hao Wang, Deep Reinforcement Learning for Wind and Energy Storage Coordination in Wholesale Energy and Ancillary Service Markets, Energy and AI, 2023.

5) J. Li^, C., Wang, Hao Wang, Attentive Convolutional Deep Reinforcement Learning for Optimizing Solar-Storage Systems in Real-time Electricity Markets, IEEE Transactions on Industrial Informatics, 2024.

6) J. Li^, C., Wang, Y. Zhang, Hao Wang, Temporal-Aware Deep Reinforcement Learning for Energy Storage Bidding in Energy and Contingency Reserve Markets, IEEE Transactions on Energy Markets, Policy, and Regulation, 2024.

7) X. Li*, J. Li^, C., Wang, Hao Wang, Model Predictive Control for Real-Time Price-Maker Bidding by Grid-Scale Battery Energy Storage Systems, IEEE Power & Energy Society General Meeting (PESGM), 2025.

8) R Tian, M Zhang, Y Liu, Hao Wang, Y Zhang, LLM-Enhanced Trading Decision Framework with Multi-Scale Memory for Electricity Markets, IEEE SmartGridComm 2025 (IEEE International Conference on Communications, Control, and Computing Technologies for Smart Grids), 2025. (Best Paper Award)

9) H Zhou*, S Prasad*, C Huang^, J Feng, Hao Wang, Hybrid Kolmogorov-Arnold Network and XGBoost Framework for Week-Ahead Price Forecasting in Australia's National Electricity Market, The 24th IEEE International Conference on Industrial Informatics (INDIN), 2026.

10) J Li^, Hao Wang, Interpretable Kolmogorov-Arnold Network with Feature-Isolated Temporal Attention Mechanism for Electricity Load Forecasting, Applied Energy 422, 1-18, 2026.

(*Thesis students. ^HDR students.)

Required knowledge

Required: Python programming and an interest in artificial intelligence, optimisation or electricity markets.

Preferred: Experience in one or more of machine learning, deep learning, time-series forecasting, reinforcement learning, mathematical optimisation, operations research or data analytics. Previous knowledge of electricity markets or energy systems is helpful but not essential.