【摘要】 从应用的角度提出一种基于带约束非线性规划的多项式偏最小二乘算法(QPLS),利用差分进化算法(DE)计算最优输入权值和内部关系式的最优参数。在对常压塔常三线柴油凝点的软测量模型应用结果表明,该算法与基于传统优化算法进行参数优化的多项式QPLS相比,模型拟合精度高,且对初值的依赖性小,并具有良好的预报能力。更多还原

【Abstract】 A quadratic partial least squares(QPLS) algorithm based on nonlinear constrained programming was presented.And differential evolution(DE) algorithm was used to calculate the optimal input weights and the parameters of inner relationship.An application of the technology to soft measurement of diesel oil solidifying point on a crude distillation unit was demonstrated.In comparison with linear partial least squares(PLS) and QPLS algorithms using other traditional optimization techniques,the proposed algorithm can improve the fitting accuracy of the model and decrease the computation burden.更多还原