Multi-Objective Optimization Scheduling of Air Conditioning Load Based on the PPO Algorithm

As a typical type of adjustable flexible load, air conditioning load can participate in grid demand response through load aggregators, demonstrating significant potential in alleviating summer peak demand and optimizing the load profile. To validate its optimization capabilities, a source-storage-load coordinated optimization model is constructed by integrating energy storage systems, and a deep reinforcement learning (DRL)-based optimization scheduling strategy is proposed. This strategy aims to minimize distribution network operational costs while maximizing user comfort, achieved by designing the state space, action space, and reward function accordingly. The Proximal Policy Optimization (PPO) algorithm is employed to dynamically balance economic efficiency and thermal comfort. Numerical case studies demonstrate that the proposed strategy effectively reduces daily operational costs while ensuring user satisfaction, thereby validating the feasibility and effectiveness of PPO in coordinating air conditioning load aggregation for grid optimization.