Analysis of scalability and sensitivity for chaotic sine cosine algorithms

Abstract EN

Chaotic Sine-Cosine Algorithms (CSCAs) are new metaheuristic optimization algorithms.

However, Chaotic Sine-Cosine Algorithm (CSCAs) are able to manipulate the problems in the standard Sine-Cosine Algorithm (SCA) like, slow convergence rate and falling into local solutions.

This manipulation is done by changing the random parameters in the standard Sine-Cosine Algorithm (SCA) with the chaotic sequences.

To verify the ability of the Chaotic Sine-Cosine Algorithms (CSCAs) for solving problems with large scale problems.

The behaviors of the Chaotic Sine-Cosine Algorithms (CSCAs) were studied under different dimensions 10, 30, 100, and 200.

The results show the high quality solutions and the superiority of all Chaotic Sine-Cosine Algorithms (CSCAs) on the standard SCA algorithm for all selecting dimensions.

Additionally, different initial values of the chaotic maps are used to study the sensitivity of Chaotic Sine-Cosine Algorithms (CSCAs).

The sensitivity test reveals that the initial value 0.7 is the best option for all Chaotic Sine-Cosine Algorithms (CSCAs).