No abstract

No abstract

No abstract

Summary We study the optimal operation of air‐conditioning loads for commercial consumers based on the time‐of‐use (TOU) ($/kWh) and demand ($/kW) pricing plan. The problem is formulated as a dynamic program that aims to adjust consumer needs for electricity charges and comfort flexibly. Moreover, the Monte Carlo method is used to imitate the environmental uncertainties due to predicted outdoor temperatures. To determine the day‐ahead temperature schedules for air‐conditioning loads, a quantum‐behaved particle swarm optimization based on space compression strategy (QPSO‐SC) is presented in this paper. Space compression and particle re‐initialization by chaotic initialization enhance the performance of the QPSO‐SC. Finally, we test the proposed method for the TOU and demand pricing plan from Duke Energy. Thermostatical, daily energy charge saving, and monthly charge saving strategies are first compared, and the QPSO‐SC is further compared with the particle swarm optimization, genetic algorithm, differential evolution, and other optimization algorithms. From the extensive simulations, the QPSO‐SC is observed to be capable of yielding higher‐quality solutions stably and efficiently than the other optimization algorithms; the proposed optimal strategy can also achieve a better balance between electricity charge and customer comfort, considering environmental uncertainties.

A data-driven optimal control method for an air supply system in proton exchange membrane fuel cells (PEMFCs) is proposed with the aim of improving the PEMFC net output power and operational efficiency. Moreover, a marginal utility-based double-delay deep deterministic policy gradient (MU-4DPG) algorithm is proposed as a an offline tuner for the PID controller. The coefficients of the PID controller are rectified and optimized during training in order to enhance the controller’s performance. The design of the algorithm draws on the concept of marginal effects in Economics, in that the algorithm continuously switches between different forms of exploration noise during training so as to increase the diversity of samples, improve exploration efficiency and avoid Q-value overfitting, and ultimately improve the robustness of the algorithm. As detailed below, the effectiveness of the control method has been experimentally demonstrated.