Development and applications of an intelligent crow search algorithm based on opposition based learning. (April 2020)

• Record Type: Journal Article
• Authors: Shekhawat, Shalini; Saxena, Akash
• Is Part Of: ISA transactions. Volume 99(2020)
• Page Start: 210
• Page End: 230
• Publication Date: 2020-04
• Subjects: CSA -- Metaheuristic algorithms -- Benchmark functions -- Bio-inspired algorithms; Engineering instruments -- Periodicals; Engineering instruments; Periodicals; Electronic journals; 629.805
• DOI: 10.1016/j.isatra.2019.09.004
• Languages: English
• ISSNs: 0019-0578

Abstract

Abstract: Metaheuristics are proven beneficial tools for solving complex, hard optimization problems. Recently, a plethora of work has been reported on bio inspired optimization algorithms. These algorithms are mimicry of behavior of animals, plants and processes into mathematical paradigms. With these developments, a new entrant in this group is Crow Search Algorithm (CSA). CSA is based on the strategic behavior of crows while searching food, thievery and chasing behavior. This algorithm sometimes suffers with local minima stagnation and unbalance exploration and exploitation phases. To overcome this problem, a cosine function is proposed first, to accelerate the exploration and retard the exploitation process with due course of the iterative process. Secondly the opposition based learning concept is incorporated for enhancing the exploration virtue of CSA. The evolved variant with the inculcation of these two concepts is named as Intelligent Crow Search Algorithm (ICSA). The algorithm is benchmarked on two benchmark function sets, one is the set of 23 standard test functions and another is set of latest benchmark function CEC-2017. Further, the applicability of this variant is tested over structural design problem, frequency wave synthesis problem and Model Order Reduction (MOR). Results reveal that ICSA exhibits competitive performance on benchmarks and real applications when compared with some contemporary optimizers. Highlights: Intelligent Crow Search Algorithm (ICSA) is proposed. Cosine function based acceleration factor and opposition based learning is applied. Modifications ensure effective bridging between exploration and exploitation. Performance of algorithm are validated on latest benchmark functions. Real applications of proposed variants are reported on challenging problems.