启发式优化算法中寻优代理过早收敛易陷入局部最优.本文对此进行机理分析并发现,虚拟碰撞作为一种隐性过早收敛现象将直接影响群体智能优化算法的准确性与快速性,而采样过程的无约束性和样本分布信息的缺失是导致虚拟碰撞的根本原因.为解决上述问题,本文提出雨林优化算法.该算法仿照植物生长模式,利用规模可变种群代替规模限定种群进行分区分级寻优采样,并结合均匀与非均匀采样原则来权衡优化算法的探索与挖掘,可以有效减少虚拟碰撞的发生,在提高寻优效率的同时,获取精准性和稳定性较高的全局最优解.与遗传算法、粒子群算法对标称函数的寻优对比实验表明,雨林算法在快速性、准确性以及泛化能力等方面均具有优势. Heuristic optimization algorithm premature convergence of preemptive agents tends to fall into local optimum.This paper analyzes the mechanism and finds that the virtual collision as a recessive premature convergence will directly affect the accuracy and rapidity of the group intelligent optimization algorithm , And the lack of sample information and the lack of sampling information are the root causes of the virtual collision.In order to solve the above problems, this paper proposes a rainforest optimization algorithm, which simulates the plant growth model and uses the variable-size population instead of the limited-size population Partitioning and graded optimal sampling, combined with the principle of uniform and non-uniform sampling to weigh optimization algorithm exploration and mining, which can effectively reduce the occurrence of virtual collision, while improving the efficiency of the search for the highest accuracy and stability of the global Optimal solution.Compared with genetic algorithm and particle swarm optimization algorithm, the experiments show that rainforest algorithm has advantages in fastness, accuracy and generalization ability.