Abstract: With the improvement of systematicness and integrality of the object model description, the complexity of process optimization problem is increased and then higher requirements are put forward for the performance of optimization algorithm. The existing nonlinear programming algorithms have both merits and demerits in their solving performance. A convergence depth control based polynary hybrid nonlinear programming algorithm was proposed in this paper. Each nonlinear programming algorithm was regarded as a meta-algorithm. In order to take full advantage of each meta-algorithm, convergence depth was used to control interactive cooperation among them and then the ability for solving large-scale complex optimization problem could be enhanced. Data reconciliation problem of air separation system and optimization problem of depropanizer and debutanizer distillation column systems were taken to test the proposed algorithm. The numerical results showed that the solving ability of the polynary hybrid nonlinear programming algorithm was better than the single nonlinear programming algorithm.

Key words: optimization, convergence depth, polynary, nonlinear programming, process systems, numerical simulation

摘要： 随着对象模型描述的系统性和完整性的提高，过程优化问题的复杂程度逐步增加，对优化算法的性能提出了更高的要求。现有的非线性规划算法在求解性能上各有优劣，本文提出了一种基于收敛深度控制的多元混合非线性规划算法，将各个非线性规划算法视为元算法，利用收敛深度来控制这些元算法之间的相互协作，更好地发挥元算法各自的优势，从而提高求解大规模复杂优化问题的能力。采用空分系统的数据校正问题以及脱丙烷塔和脱丁烷塔联塔系统的优化问题对多元混合算法进行了测试，数值结果表明相比各个单独的非线性规划算法而言，多元混合算法具有更好的求解性能。

关键词: 优化, 收敛深度, 多元, 非线性规划, 过程系统, 数值模拟