Improved Dandelion Optimizer Based on Elite Pool and Mirror Optimization Strategy.

A new swarm intelligence optimization algorithm is proposed to simulate the long-distance flight process of dandelion seeds as a function of wind by the dandelion optimizer (DO). An improved version of DO, called IDO, is proposed to overcome the shortcomings of slow optimization speed and vulnerability to local optimum of the algorithm. In the landing stage of the dandelion seeds, IDO introduces the elite pool roulette wheel selection mechanism to maintain the diversity of the dandelion population, which enables the algorithm to maintain a better optimization ability. In the rising stage, a gamma distribution is introduced to make the algorithm population more diverse, thus improving its optimization capabilities. Finally, the mirror reversal strategy is introduced to search for the mirror solutions of inferior particles in the population surrounding the optimal individual, which helps the population to escape from local optima. The performance of the proposed improved IDO algorithm was tested against the CEC-2017, CEC-2020, and CEC-2022 benchmark functions. Finally, three engineering design problems were optimized, namely, three-bar truss design, cantilever design, and pressure vessel design. The experimental results show that the proposed IDO algorithm significantly improves the convergence rate and optimization accuracy. The experimental results show that the proposed IDO algorithm optimizes the exploitation and exploration process, improves the convergence rate and optimization accuracy. [ABSTRACT FROM AUTHOR]