Dynamic distributed multi-objective mantis search algorithm based on Transformer hybrid strategy for novel power system dispatch

With massive renewable energy connected to the grid and increasing the complexity of power systems (PSs), traditional centralized PS models can not satisfy the needs of modern PSs. Consequently, a distributed novel power systems model (DNPSs) is proposed in this study aiming at efficient complementa...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Energy (Oxford) Jg. 324; S. 136075
Hauptverfasser: Lu, Quan, Zeng, Haozheng, Yin, Linfei
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Elsevier Ltd 01.06.2025
Schlagworte:
algorithms
biomass
carbon dioxide
Distributed multi-objective mantis search algorithm
Distributed novel power systems
energy
Levy flight and Gaussian mutation
Markov chain
mutation
prediction
Transformer prediction strategy
water power
wind
Levy flight and Gaussian mutation
Distributed multi-objective mantis search algorithm
Transformer prediction strategy
Distributed novel power systems
ISSN:0360-5442
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:With massive renewable energy connected to the grid and increasing the complexity of power systems (PSs), traditional centralized PS models can not satisfy the needs of modern PSs. Consequently, a distributed novel power systems model (DNPSs) is proposed in this study aiming at efficient complementarity and optimal dispatch of multiple energy sources by coordinating the utilization of thermal, wind, photovoltaic, hydro, nuclear, biomass, and tidal energy. To address the dispatch problem in DNPSs, this study presents a distributed multi-objective mantis search algorithm (DMOMSA) which combines a distributed computing framework and a multi-objective mantis search algorithm. Although DMOMSA has significant optimal capabilities, the problem of premature convergence still exists. To overcome the problem of premature convergence of heuristic algorithms, this study presents a dynamic distributed multi-objective mantis search algorithm based on Transformer hybrid strategy (Transformer-DDMOMSA). The Transformer-DDDMOMSA strengthens the exploratory capacity of Transformer-DDMOMSA to avoid premature convergence by introducing a hybrid strategy consisting of the Transformer prediction strategy, Levy Flight, and Gaussian mutation. Simulation results indicate that Transformer-DDMOMSA is compared with non-dominated sorting moth flame optimization, multi-objective gray wolf optimizer, multi-objective mantis search algorithm, and multi-objective ant-lion optimizer: (1) in case I, generator costs are reduced by at least 2.36 %, 2.54 %, 2.30 %, and 4.20 %, respectively, while carbon dioxide emissions are reduced by at least 6.11 %, 9.92 %, 4.64 %, and 13.45 %, respectively; (2) in case II, generator costs are reduced by at least 2.38 %, 2.37 %, 1.92 %, and 2.70 %, respectively, while carbon dioxide emissions are reduced by at least 8.80 %, 12.01 %, 8.76 %, and 13.14 %, respectively; (3) the Transformer-DDMOMSA performs excellent in three evaluation metrics values of diversity, Euclidean distance and hypervolume. •Distributed multi-objective economic dispatch problems are established.•Dynamic distributed mantis algorithm with Transformer hybrid strategy is proposed.•The proposed method avoids algorithm from falling into local optimal solutions.•Transformer hybrid strategies can help the proposed method explore expansive spaces.•Generator costs and carbon emissions are optimized simultaneously by proposed method.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:0360-5442
DOI:10.1016/j.energy.2025.136075