Modified Cuckoo Search Algorithm: A Novel Method to Minimize the Fuel Cost

• Published in: Energies (Basel) Vol. 11; no. 6; p. 1328
• Main Authors: Nguyen, Thang, Vo, Dieu, Vu Quynh, Nguyen, Van Dai, Le
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.06.2018
• Subjects: Adaptive algorithms; Algorithms; Convergence; cuckoo search algorithm; Data buses; Evolutionary algorithms; Genetic algorithms; IEEE networks; Methods; Neural networks; Operating zones; Optimization; prohibited operating zone; Prohibited operating zones; Search algorithms; State of the art; System effectiveness; transmission network constraints; valve point loading effects
• ISSN: 1996-1073; 1996-1073
• DOI: 10.3390/en11061328

Economic load dispatch (ELD) is an important optimization problem for operating and controlling modern power systems, and if ELD is effectively executed, power systems work stably and economically. The main objective of this paper is to develop a novel method to solve the ELD with the purpose of minimizing the total fuel cost of all available generating units while requirements are to satisfy all constraints regarding thermal units, generators, and transmission power networks. The proposed high performance cuckoo search algorithm (HPCSA) is developed from the efficient technique for the second new solution generation of conventional cuckoo search algorithm (CCSA), called adaptive mutation technique. This proposed technique diversifies the local search ability based on a new comparison criterion. The HPCSA is verified on difference systems under special conditions, namely the 10-unit system with multi fuels, 15-unit system considering prohibited operating zones, and three IEEE systems with 30, 57, and 118 buses considering transmission power network constraints. The specific evaluation of the HPCSA is compared to that of Lagrange optimization-based methods (LMS), neural network-based methods (NNMS), CCSA, and other popular methods such as Particle swarm optimization (PSO) variants, Differential evolution (DE) variants, Genetic Algorithm (GA) variants, and state-of-the-art methods. In comparison with CCSA, the proposed method is always more effective and more robust since the proposed method can find most solutions with better quality and faster convergence speed. In comparison with LMS and NNMS, the proposed method can also find solutions with approximate or equal quality. In comparison with popular methods and state-of-the-art methods, the proposed method has more potential since it can reach faster convergence to valid solutions with approximate or better quality. Consequently, it can be concluded that the proposed HPCSA is an effective optimization tool for dealing with ELD problems.