4 - Improved crow search algorithm based on arithmetic crossover—a novel metaheuristic technique for solving engineering optimization problems

Abstract

The metaheuristic optimization techniques’ role is inevitable in solving real-world problems. Following the crow's intelligent trait, the classical crow search algorithm (CSA) was demonstrated. From the optimization point of view, it can be viewed that the crows are analogous to searchers and the environment is considered as the solution space. Every location in the solution space indicates a feasible solution. Also, the grade of food source is synonymous to the objective or fitness function and the optimum food source is the attainment of the global solution of the problem. The crows have good memory power and classical crow search optimization was proved to be efficient, generates promising outcomes, when compared with the classical optimization techniques. The performance of the genetic algorithm (GA) was improved by arithmetic crossover and the improved crow search optimization utilizes the arithmetic crossover operation of GA. The arithmetic crossover operation was incorporated in the position update of crows and formulated the hybrid crow optimization (CO) algorithm. The performance of the hybrid CSA was verified on benchmark functions and satisfactory results are produced while comparing with the classical crow search optimization. The hybrid CO technique was also coupled with support vector machine (SVM) for the classification of a brain tumor on MR brain images; effective results were achieved while comparing with SVM coupled with a GA, simulated annealing, and crow search optimization techniques. The hybrid crow search optimization was also validated on two design problems (three-bar truss and cantilever beam) and compared with the classical methods.