Aiming at the deficiencies of ant colony clustering algorithms in distance measure, convergence rate, similarity clustering, and other aspects, WAlf algorithm is presented as an improvement to the lf algorithm. Using ...
详细信息
ISBN:
(纸本)9781467380027
Aiming at the deficiencies of ant colony clustering algorithms in distance measure, convergence rate, similarity clustering, and other aspects, WAlf algorithm is presented as an improvement to the lf algorithm. Using weighted hybrid distance as distance measure, and introducing the adaptive mechanism during the process of clustering, ant colonies can adjust the radius of neighborhoods dynamically and merge similar clusters in the process, meeting the clustering requirements while improving the convergence speed. Finally WAlf algorithm has been shown through experiments to be better than lf algorithm both in clustering results and efficiency.
Due to deregulation of distribution network investment businesses, geographically divided microgrids may be owned and operated by different entities. In this study, distributed multi-area load flow (lf) calculation me...
详细信息
Due to deregulation of distribution network investment businesses, geographically divided microgrids may be owned and operated by different entities. In this study, distributed multi-area load flow (lf) calculation method that can preserve the operational independence of each entity is necessary. However, conventional fixed-point iteration-based distributed lf (Dlf) algorithm may encounter convergence problems in active distribution networks integrated with high penetration of distributed generators (DGs). In this work, the authors propose a more general and novel Dlf algorithm to better accommodate the integration of DGs. In this algorithm, the original non-convex lf model for multi-microgrid systems is convexified into a second-order conic lf problem and then decomposed into a regional lf subproblem using a modified alternating direction method of multipliers (ADMM). The conventional ADMM is enhanced by asymmetric decomposition and partial boundary penalty relaxation with regard to the specific lf problem structure to accelerate the convergence procedure. Numerical tests on an IEEE 69-bus distribution system and a 345-bus system show that the proposed method outperforms the fixed-point iteration-based Dlf method.
This paper provides a new method for arranging data sets into clusters. The proposed model, called FlyAntClass, starts from the ants collective sorting behavior and overwrites it with additional behaviors inspired fro...
详细信息
ISBN:
(纸本)9781509043200
This paper provides a new method for arranging data sets into clusters. The proposed model, called FlyAntClass, starts from the ants collective sorting behavior and overwrites it with additional behaviors inspired from birds and spiders: in this context, birds' moving behavior is used to control next relative positions for a moving ant;and spiders' homing behavior is provided to manage movements of ants with conflict situation.
暂无评论