A Multi-strategy Improved Dung Beetle Optimisation Algorithm and its Application

Abstract In order to solve the shortcomings of Dung Beetle Optimizer such as low convergence accuracy and easy to fall into local optimum, a multi-strategy improved Dung Beetle Optimizer (IDBO) is proposed. Firstly, the Cubic chaotic mapping strategy is introduced to improve the diversity of the initial population. Secondly, the global exploration strategy in the Fishhawk algorithm is introduced to give the dung beetle algorithm the exploration ability of identifying the optimal region and escaping from the local optimum, which initially improves the convergence speed and optimality-seeking accuracy of the algorithm. Finally, the dung beetle foraging behaviour is perturbed using the adaptive t-distribution perturbation strategy, making the dung beetle algorithm improve the global exploitation ability and local exploration ability while further accelerate its speed of convergence. The effectiveness of the three improved strategies is verified by testing and analysing the CEC2021 and CEC2017 test functions. The optimization results of the improved algorithms and the comparison algorithms are subjected to convergence analysis and Wilcoxon rank sum test, which proves that the IDBO algorithm has good convergence speed and optimization accuracy. In addition, the IDBO algorithm is adopted to optimise parameters of the HKELM prediction model which is applied to the short-term PV power prediction simulation and comparison experiments. Experimental results show that the IDBO-HKELM prediction model can effectively improve the prediction accuracy of short-term PV power, which further verifies the feasibility and validity of the IDBO algorithm in solving the problems of practical applications.