Optimal power flow (OPF) plays a crucial role in addressing asymmetric operational problems of bipolar direct current distribution networks (DCDNs). However, the effectiveness of existing OPF models applied to bipolar...
详细信息
Optimal power flow (OPF) plays a crucial role in addressing asymmetric operational problems of bipolar direct current distribution networks (DCDNs). However, the effectiveness of existing OPF models applied to bipolar DCDNs exhibits sensitivity to the problem type, underlying assumptions, and choice of objective functions. Hence, this paper proposes a dynamic decoupling-based convex-concave programming (CCP) framework to formulate an OPF model, aiming for a more rational impact across diverse applications in asymmetric operational bipolar DCDNs. More specifically, the coupled power between poles and ports of bipolar DCDNs is analyzed in detail, leading to the derivation of decoupled pole-to-ground equivalent circuits without relying on any preliminary assumption. Subsequently, the OPF problem of the equivalent circuits is formulated via CCP, involving the introduction of a cutting plane through difference-of-convex inequalities to limit the feasible region of the traditional OPF based on second-order cone programming, thereby diminishing dependence on exact relaxation conditions. Finally, a solution method is proposed to dynamically correct both the decoupled equivalent circuits and the cutting plane. The numerical results indicate that this method effectively addresses power optimization problems with a non-strictly monotonic objective function. Furthermore, it exhibits scalability, showcasing its applicability to large-scale radial bipolar DCDNs.
In this letter, we target the problem of model selection for support vector classifiers through in-sample methods, which are particularly appealing in the small-sample regime. In particular, we describe the applicatio...
详细信息
In this letter, we target the problem of model selection for support vector classifiers through in-sample methods, which are particularly appealing in the small-sample regime. In particular, we describe the application of a trimmed hinge loss function to the Rademacher complexity and maximal discrepancy-based in-sample approaches and show that the selected classifiers outperform the ones obtained with other in-sample model selection techniques, which exploit a soft loss function, in classifying microarray data.
暂无评论