一种高效的快速近似控制向量参数化方法

控制向量参数化(Control vector parameterization, CVP) 方法是目前求解流程工业中最优操作问题的主流数值方法,然而,该方法的主要缺点之一是 计算效率较低,这是因为在求解生成的非线性规划(Nonlinear programming, NLP) 问题时,需要随着控制参数的调整,反复不断地求解相关的微分方程组,这也是CVP 方法中最耗时的部分.为了提高CVP 方法的计算效率,本文提出一种新颖的快速近似方法,能够有效减少微分方程组、函数值以及 梯度的计算量.最后,两个经典的最优控制问题上的测试结果及与国外成熟的最优控制 软件的比较研究表明:本文提出的快速近似CVP 方法在精度和效率上兼有良好的表现.

An Efficient Fast Approximate Control Vector Parameterization Method

The control vector parameterization (CVP) method is currently popular for solving optimal control problems in process industries. However, one of its main disadvantages is the low computational efficiency, because the relevant differential equations in solving the generated nonlinear programming (NLP) problem should be calculated repeatedly with the adjustment of control, which is the most time-consuming part of the CVP method. A new fast approximate approach with less computational expenses on differential equations, function values and gradients is therefore proposed to improve the computational efficiency of the CVP method in this paper. The proposed approach is demonstrated to have marked advantages in terms of accuracy and efficiency, in contrast to mature optimal control softwares, on two classic optimal control problems.