An Enhanced Dung Beetle Optimization Algorithm for Global Optimization

Abstract

This paper proposes an enhanced dung beetle optimization (EDBO) algorithm in order to address the issues of the dung beetle optimization (DBO) algorithm which include easy convergence to the local optimal, slow convergence speed, and poor global search capability. The improvements in the EDBO are implemented via the following four aspects. Firstly, the SPM chaotic mapping designed through combing Sine mapping and Piece-Wise Linear Chaotic Mapping is introduced to initialize the population for increasing diversity of population. Secondly, the position update formula in the Golden Sine Algorithm (Golden-SA) is used to replace the formula for the mathematical model of dung beetle ball-rolling behavior without obstacle with the purpose of improving the convergence accuracy and accelerating the convergence speed. Thirdly, the spiral foraging strategy in the tuna swam optimization (TSO) is hybridized with the mathematical model of dung beetle breeding and foraging behavior. The hybridization not only balances the global exploration and local exploitation but also keeps the diversity of the population. Finally, the EDBO can enhance the capability of escaping the local optima and extending the search space by means of bringing in the two different sets of adaptive weight coefficients. The performance of the EDBO is evaluated and compared with other swarm intelligence optimization algorithms via the benchmark functions of different characteristics. The results demonstrate that the EDBO outperforms the classical DBO and other compared algorithms in terms of convergence speed and accuracy.