Optimal scheduling of battery energy storage system plays crucial part in distributedenergy system. As a data driven method, deep reinforcement learning does not requiresystem knowledge of dynamic system, present optimal solution for nonlinear optimizationproblem. In this research, financial cost of energy consumption reduced by schedulingbattery energy using deep reinforcement learning method (RL). Reinforcement learning canadapt to equipment parameter changes and noise in the data, while mixed-integer linearprogramming (MILP) requires high accuracy in forecasting power generation and demand,accurate equipment parameters to achieve good performance, and high computational costfor large-scale industrial applications. Based on this, it can be assumed that deep RL basedsolution is capable of outperform classic deterministic optimization model MILP. This studycompares four state-of-the-art RL algorithms for the battery power plant control problem:PPO, A2C, SAC, TD3. According to the simulation results, TD3 shows the best results,outperforming MILP by 5 % in cost savings, and the time to solve the problem is reducedby about a factor of three.
Original languageEnglish
Pages (from-to)391-402
Number of pages12
JournalVestnik Sankt-Peterburgskogo Universiteta, Prikladnaya Matematika, Informatika, Protsessy Upravleniya
Volume19
Issue number3
DOIs
StatePublished - 1 Nov 2023

    Research areas

  • reinforcement learning, energy management system, distributed energy system, numerical optimization

ID: 111486361