A Novel Approach Based on Average Swarm Intelligence to Improve the Whale Optimization Algorithm

In this study, a new technique has been introduced by changing the convergence of the whale optimization algorithm, which has the principle of approaching its prey by following the pack leader strictly. For this, first of all, average position values of the swarm were obtained in each iteration. Later, when the "p" parameter, which is used to add randomness to the progress of the swarm members, is below a certain value, the swarm average was used for each individual to move to the new position. Thus, slow convergence and frequent falling to the local optimum which is considered to be the biggest shortcoming of the algorithm, has been eliminated. The distance of whales from each other and from prey was modeled as a fitness function and the Euclidean distance formula was used for this. A complex engineering problem was chosen to reveal the power of both the classical whale optimization algorithm and the algorithm that includes the proposed new technique. As a result, this new technique introduced has provided a 10 million times improvement in solving this complex engineering problem used in the control of serial robot manipulators.