Optimally scheduling power consumption of appliances is the essential feature of smart grid, which enables Demand Response Management (DRM) and helps to shape the power usage profile. This problem is often required to be solved in face of a large number of appliances and many time slots; thus the computational efficiency of solving a large scale optimal power scheduling problem with limited computational resources becomes the major concern of algorithm design. To this end, a novel algorithm is proposed based on Karush-Kuhn-Tucker (KKT) conditions to solve the optimal power scheduling problem with temporally spatially coupled constraints in a distributed manner. The proposed algorithm converts the original problem into equivalently solving an optimal KKT operator in a much lower dimension, thus the computation speed is greatly enhanced. In addition, the proposed method dose not require a step size in the iteration process, thus avoids the oscillation of numerical solution caused by problem parameter changes. Compared with the widely used conventional algorithms, e.g., interior point method and dual decomposition, the higher computational speed and less sensitivity to the problem parameter setting are observed in numerical simulations.

中文翻译：

---

具有时空耦合约束的大型设备最优功率调度的快速算法

最佳调度设备的功耗是智能电网的基本功能，它可以启用需求响应管理（DRM）并有助于调整功耗使用情况。面对大量的设备和许多时隙，通常需要解决这个问题。因此，以有限的计算资源来解决大规模最优功率调度问题的计算效率成为算法设计的主要关注点。为此，提出了一种基于Karush-Kuhn-Tucker（KKT）条件的新颖算法，以分布式方式解决具有时间空间耦合约束的最优功率调度问题。所提出的算法将原始问题转化为在更低的维度上等效求解最优KKT算子，从而大大提高了计算速度。此外，所提出的方法在迭代过程中不需要步长，从而避免了由于问题参数变化而引起的数值解的振荡。与广泛使用的传统算法（例如，内部点方法和对偶分解）相比，在数值模拟中观察到更高的计算速度和对问题参数设置的敏感性更低。