Improved grasshopper optimization algorithm using opposition-based learning

• Published in: Expert systems with applications Vol. 112; pp. 156 - 172
• Main Authors: Ewees, Ahmed A., Abd Elaziz, Mohamed, Houssein, Essam H.
• Format: Journal Article
• Language: English
• Published: New York Elsevier Ltd 01.12.2018; Elsevier BV
• Subjects: Algorithms; Benchmark functions; Benchmarks; Engineering; Engineering problems optimization; Expert systems; Grasshopper optimization algorithm; Machine learning; Mathematical models; Opposition-based learning; Optimization; Optimization algorithms; Benchmark functions; Engineering problems optimization; Grasshopper optimization algorithm; Opposition-based learning
• ISSN: 0957-4174; 1873-6793
• DOI: 10.1016/j.eswa.2018.06.023

•A modified GOA is proposed based on opposite-based learning strategy (OBLGOA).•OBLGOA is evaluated using 23 benchmark problems and four engineering problems.•The results were superior to those of well-known algorithms in optimization domain. This paper proposes an improved version of the grasshopper optimization algorithm (GOA) based on the opposition-based learning (OBL) strategy called OBLGOA for solving benchmark optimization functions and engineering problems. The proposed OBLGOA algorithm consists of two stages: the first stage generates an initial population and its opposite using the OBL strategy; and the second stage uses the OBL as an additional phase to update the GOA population in each iteration. However, the OBL is applied to only half of the solutions to reduce the time complexity. To investigate the performance of the proposed OBLGOA, six sets of experiment series are performed, and they include twenty-three benchmark functions and four engineering problems. The experiments revealed that the results of the proposed algorithm were superior to those of ten well-known algorithms in this domain. Eventually, the obtained results proved that the OBLGOA algorithm can provide competitive results for optimization engineering problems compared with state-of-the-art algorithms.