Abstract:

Aiming at the problems of sine cosine algorithm, such as easy to fall into local optimum, low accuracy and slow convergence speed, an improved sine cosine algorithm based on dynamic classification strategy is proposed. Firstly, the Latin hypercube sampling method is introduced to divide the search space evenly so that the initial population covers the whole search space to maintain the diversity of the initial population. Secondly, the dynamic classification strategy is used, according to the order of fitness value, the population dynamics is divided into three grades: good, medium and poor. The destruction strategy and elite guidance method are used to disturb the population dynamics, so as to improve the convergence accuracy of the algorithm and enhance the ability to jump out of the local optimum. Finally, the global search strategy of dynamic reverse learning is designed by introducing the reverse learning method to improve the convergence speed of the algorithm. At the same time, the performance of the improved algorithm in the aspects of complexity, convergence and stability is tested, and 15 standard test functions are selected to carry out the simulation experiment analysis in the low and high dimensional state, and compared with particle swarm optimization algorithm, backtracking search algorithm and other improved sine cosine algorithms. The simulation results show that the proposed algorithm can improve the convergence and stability of the algorithm.

Key words: improved sines cosine algorithm, Latin hypercube population initialization strategy, dynamic classification strategy, global search strategy, numerical simulation analysis