The estimation of power curve is the central task for efficient operation and prediction of wind power generation. It is often the case, however, that the actual data exhibit a great deal of variations in power output with respect to wind speed, and thus the power curve estimation necessitates the detection and proper treatment of outliers. This study proposes a novel procedure for outlier detection and elimination for estimating power curves of wind farms by employing clustering algorithms of vector quantization and density-based spatial clustering of applications with noise. Testing different parametric models of power output curve, the proposed methodology is demonstrated for obtaining power curves of individual wind turbines in a Korean wind farm. It is asserted that the outlier elimination procedure for power curve modeling outlined in this study can be highly efficient at the presence of noises.

A novel approach to power scheduling is introduced, focusing on minimizing both economic and environmental impacts. This method utilizes deep contextual reinforcement learning (RL) within an agent-based simulation environment. Each generating unit is treated as an independent, heterogeneous agent, and the scheduling dynamics are formulated as Markov decision processes (MDPs). The MDPs are then used to train a deep RL model to determine optimal power schedules. The performance of this approach is evaluated across various power systems, including both small-scale and large-scale systems with up to 100 units. The results demonstrate that the proposed method exhibits superior performance and scalability in handling power systems with a larger number of units.

This paper studies the problem of finding an optimal subcarrier and power allocation for downlink communication in the orthogonal frequency division multiplexing (OFDM) wireless systems. With the constraints on the power consumption and transmission rates for users, the problem of maximizing total throughput is mathematically formulated. The developed mathematical model is so general as to include almost of existing ones in which the original problems have been usually divided into smaller and tractable ones to make a practical application of some simple greedy algorithms. In this study, developed is the subgradient dual optimization method for the problem. Extensive experiments with realistic data show that in a reasonable time the subgradient method converges to the stable point, almost feasible to the original problem. We proved that the subgradient method is applicable to the dynamic resource allocation, by showing that environmental changes at a stable state need much lesser additional iterations to converge.