近红外光谱(NIRS)是一种间接分析技术,其应用需建立相应的校正模型。为了提高模型的解释能力、预测准确度和建模效率,需要对NIRS 进行波长选择,优选最小化冗余信息。选用蚁群优化(ACO)、遗传优化(GA)、粒子群优化(PSO)、随机青蛙(RF)和模拟退火(SA)5 种智能优化算法对烟叶总氮和烟碱近红外光谱数据进行特征波长选择,结合偏最小二乘(PLS)算法,构建了多个烟叶总氮和烟碱的校正模型,(Near infrared spectroscopy (NIRS) is an indirect an

改进的差分粒子群算法，并用于求函数极值，程序中分别于DE PSO 蛙跳进行对比，结果显示，改进的差分粒子群更好(The improved Differential Particle Swarm Optimization (DPSO) algorithm is used to find the function extremum. The program is compared with DE PSO frog jump, and the results show that the improve

蛙跳算法，基于蛙跳算法实现优化变分模态分解的K和a.(Frog jump algorithm, based on the frog jump algorithm to achieve optimal variational mode decomposition K and A.)