A Novel Honey Badger Algorithm with Golden Sinusoidal Survival Rate Selection

Abstract

The original Honey Badger Algorithm as one of the newest meta¬heuristic techniques has a better convergence speed. However, HBA has the potential disadvantages of poor convergence accuracy, insufficient balancing among exploration and exploitation, and the propensity to slip into local optimization. In this paper, a novel golden sinusoidal survival honey badger algorithm is proposed. Firstly, an opposing learning and chaos mechanism are applied to the initial individual generation so that they can be distributed throughout the entire search area, which improves the precision of initial populations. Secondly, in the position update phase, we use a nonlinear convergence strategy to balance the weight of prey in the next walk and to increase the global search ability. After that, evaluating the quality of honey badger by golden sinusoidal survival rate and updating precocious individuals by Lévy flight, through which the premature convergence of the algorithm can be avoided. Finally, 23 benchmark function, CEC2019 tests are employed to assess the effectivity of improved algorithm. Test results indicate that the algorithm's capabilities to evolve, to extricate the local optimal and to detect the global optimal placements are enhanced.