A Heterogeneous Evolving Cuckoo Search Algorithm for Solving Large-Scale Combined Heat and Power Economic Dispatch Problems

• Published in: IEEE access Vol. 7; pp. 111287 - 111301
• Main Authors: Huang, Zhenyu, Gao, Zhengzhong, Qi, Liang, Duan, Hua
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2019; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Cogeneration; Combined heat and power economic dispatch; constraint-handling mechanism; cuckoo search; Evolution; Global optimization; large-scale global optimization; Machine learning; Optimization; Power dispatch; Robustness; Robustness (mathematics); Schedules; Search algorithms; Search problems; STEM; System effectiveness; valve-point effect
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2019.2933980

Combined heat and power economic dispatch (CHPED) problem aims to optimally schedule the output of generating units with minimum fuel cost, which is a highly non-linear, non-convex and large-scale global optimization problem with many practical constraints. The complexity of the problem demands solution methods with powerful search ability, robustness, and computational efficiency. This paper proposes a heterogeneous evolving cuckoo search (HECS) algorithm with a novel constraint-handling mechanism to solve the large-scale CHPED problem considering valve-point effect. Based on the cuckoo search algorithm, we apply a comprehensive learning strategy to enhance the search ability in the high-dimensional environment, and a heterogeneous evolving strategy to improve the robustness of the algorithm. Moreover, we develop a novel constraint-handling mechanism that uses strict mathematical methods to repair unfeasible solutions and avoid redundant calculation. 5 tests are conducted on 24-unit, 48-unit, 84-unit, 96-unit, and 192-unit systems and the results are compared with the state-of-the-art algorithms published in the year 2015-2019. The comparisons show that the HECS could annually save millions of dollars in some large-scale systems, which verify its effectiveness.