Stochastic gradient-push for economic dispatch on time-varying directed networks with delays

Economic dispatch problem (EDP) is an elementary optimization problem of power systems whose purpose is to minimize the total generation cost while meeting total demands and complying with individual generator output constraints. This paper proposes a distributed optimization algorithm with noisy gr...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:International journal of electrical power & energy systems Ročník 113; s. 564 - 572
Hlavní autoři: Yuan, Ye, Li, Huaqing, Hu, Jinhui, Wang, Zheng
Médium: Journal Article
Jazyk:angličtina
Vydáno: Elsevier Ltd 01.12.2019
Témata:
Distributed optimization algorithm
Economic dispatch
Stochastic gradient-push approach
Time delays
Time-varying directed communication networks
Distributed optimization algorithm
Time delays
Economic dispatch
Stochastic gradient-push approach
Time-varying directed communication networks
ISSN:0142-0615, 1879-3517
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Economic dispatch problem (EDP) is an elementary optimization problem of power systems whose purpose is to minimize the total generation cost while meeting total demands and complying with individual generator output constraints. This paper proposes a distributed optimization algorithm with noisy gradient based on stochastic gradient-push approach to solve the EDP on time-varying directed communication networks potentially with time delays. It shows that the proposed algorithm can be ensured to solve the EDP when the time-varying directed communication networks are uniformly jointly strongly connected. The algorithm with asynchronous step-sizes can deal with the finite time-varying delays on communication links as well. The generation cost functions are considered as strictly convex and strongly convex functions respectively with convergence rate of Olnk/k and Olnk/k. Simulation results are provided to validate the availability of the algorithm.
ISSN:0142-0615
1879-3517
DOI:10.1016/j.ijepes.2019.05.051