Improving Sailfish Optimizer with Population Switching Strategy and Random Mutation Strategy

We propose two novel search strategies to further boost the performance of the standard sailfish optimizer (SFO) and present an enhanced SFO with advanced capabilities. Specifically, the first strategy, named the population switching strategy, takes into account the fitness consumption cost of the sardine population. It enables individuals from two different populations to switch, resulting in improved search efficiency. The second strategy, referred to as the random mutation strategy, assists the stagnant individuals in escaping from localized areas where they are trapped and facilitates their search for other potential regions. To evaluate the performance of these strategies, we conducted experiments using three SFO variants: SFO with the population switching strategy, SFO with the random mutation strategy, and SFO with both strategies. Additionally, we compared these variants with the standard SFO on 29 benchmark functions from the CEC2017 test suite. Each benchmark function consists of various dimensions, and each dimension was independently run 30 times. Moreover, we applied the improved algorithm to two classical engineering design problems and compared its performance with three other evolutionary computation (EC) algorithms, including the conventional SFO. The experimental results confirmed that the improved SFO exhibited superior convergence accuracy and provided improved solutions for the two engineering design problems.