摘要
随着计算机技术的发展与非线性规划求解技术的日趋成熟,动态优化方法在过程工业、机械控制等研究领域受到越来越多的关注。联立法由于采用了将连续动态优化问题离散为非线性规划(NLP)问题进行求解的思路,使得该方法简单易行并具有很高的求解效率,逐渐成为当前最具发展前景的动态优化求解策略之一。但是,在一些对计算的实时性有较高要求的应用环境中,仍然需要对现阶段的基于联立法的动态优化工具不断改进来提高优化的效率。本文围绕动态优化问题的快速求解策略展开研究,共分析讨论了4个方面的内容:
1.对利用有限元正交配置法(OCFE)离散微分代数方程(DAE)优化问题的一般形式进行了深入分析,通过严格证明得到了直接离散格式与间接离散格式的等价条件。在此基础上,对多种常用的配置方式进行了模拟与分析,并发现了间接Lobatto配置在诸多方面均可获得更好的效果。此外,进一步深入讨论了配置法离散的本质,提出基函数离散的思想,并以径向基函数(RBF)离散为例验证了非多项式离散的可行性。
2.对记忆增强优化方法(MEO)的本质进行了讨论,提出根据历史经验数据生成高等初值点(ASP)的实质为多元散乱数据拟合的观点,并成功应用RBF插值方法改进了MEO方法,使应用效果大幅提升。在此基础上,提出两层MEO的思想,并设计出效果更好的具有误差修正的RBF-MEO方法。
3.为克服RBF插值方法存在的先天不足,提出采用支持向量机(SVM)替代RBF来实现MEO的功能。同时,还针对SVM中QP求解效率不高的现状,提出了Hesse阵稀疏化方法及多元输出问题的联立求解策略,成功地提升了SVM方法的执行效率。进一步将SVM-MEO应用于解决动态参数化优化问题,提出了高效的“部分MEO"策略。
4.针对动态优化研究平台缺少的现状,构建了DAE优化框架,并在Matlab与AMPL环境中成功实施。该平台采用模块化设计思路,将离散后的NLP问题划分为模型部分与方法部分。在此平台上,可进行离散方法、模型开发与求解工具等方面的研究。该部分内容还包括如何将不同的动态优化模型进行标准化描述,以及根据离散结构的特点而设计的快速求解方法。
With the development of computer technology and the increasingly perfecting of non-linear programming (NLP) solution technology, dynamic optimization methods have at-tracted more and more attention in the process industries, machine control and other re-search areas. Since the simultaneous strategies adopt the idea that the continuous dynamic optimization problems are discretized into NLP problems and solved successively, the strategies can be executed easily and have extremely solution efficiency. The simultane-ous strategies gradually become more prospective. However, current simultaneous meth-ods still can not meet much application requirements, so need to be improved further to enhance the efficiency and practicality. This work mainly focuses on fast solving strategies for dynamic optimization with differential-algebraic equations (DAEs), including the next four aspects:
1. It is analyzed deeply that the orthogonal collocation on finite element (OCFE) meth-ods are used to discretize the DAE optimization problems, and the equivalent con--ditions between direct and indirect discretization structures are proved theoretically. On this basis, combined with the specific collocation formulae for analysis and verifi-cation, it is concluded that the Lobatto formula has the best or better effects on many solution aspects. In addition, the nature of collocation discretization methods is dis-cussed further, and then the idea of basis function discretization are proposed. The feasibility of non-polynomial discretization is exemplified by radial basis function (RBF) methods.
2. The nature of mnemonic enhancement optimization (MEO) is discussed, and it is pro-posed that the problems of advanced starting point (ASP) generated from empirical data are multivariate scatter data fitting problems in fact. Thus, the RBF interpolation principle is used to improve the MEO method, and the application results are signifi- cantly improved. Furthermore, an improved RBF-MEO method with error correction is proposed.
3. In order to avoid the inherent deficiencies of RBF method, the support vector machine (SVM) is employed as an alternative to realize the functions of MEO. According to the fact that QP problems can not be solved quickly enough in the SVM, the Hessian matrix sparseness methods and the simultaneous strategy of multi-output problems are put forward, successfully raising the solution efficiency of the SVM. The SVM-MEO method is applied to solve the parametric dynamic optimization problems, and an efficient "partial MEO" is proposed.
4. Since scarcity of research platform for dynamic optimization, a DAE optimization framework is built and successfully implemented in MATLAB and AMPL environ-ment. The platform adopts the modular design concept and divides the discrete for-mula into the model part and the method part. On this platform, the further research of discretization methods, model development and solver tools can be distributed. This section also includes standardization of dynamic optimization models and fast solution strategy based on discretization structure.
引文
丁同仁,李承治.1991.常微分方程教程[M].北京:高等教育出版社.
童季贤,张显明.1994.最优控制的数学方法及应用[M].成都:西南交通大学出版社.
胡健伟,汤怀民.2007.微分方程数值方法[M].2.北京:科学出版社.
蔡艳宁,胡昌华.2008.一种基于Cholesky分解的动态无偏LS-SVM学习算法[J].控制与决策,23(12):1363-1367.
刘培玉.1990.应用最优控制[M].大连:大连理工大学出版社.
邵之江.1997.连续工业过程的在线优化[D].杭州:浙江大学.
吴宗敏.1998.函数的径向基表示[J].数学进展,27(3):202--208.
于寅.2001.高等工程数学[M].武汉:华中科技大学出版社.
郑大钟.2002.线性系统理论[M].2.北京:清华大学出版社.
吴宗敏.2002.径向基函数、散乱数据拟合与无网格偏微分方程数值解[J].工程数学学报,19(2):1--12.
张贤达.2004.矩阵分析与应用[M].北京:清华大学出版社.
江爱朋.2005.大规模简约空间SQP算法及其在过程系统优化中的应用[D].杭州:浙江大学.
吴宗敏.2007.散乱数据拟合的模型、方法和理论[M].北京:科学出版社.
王可心.2008.大规模过程系统非线性优化的简约空间理论与算法研究[D].杭州:浙江大学.
方学毅.2009.过程系统记忆增强型实时优化方法[D].杭州:浙江大学.
马利敏.2009.径向基函数逼近中的若干理论、方法及其应用[D].上海:复旦大学.
丛爽.2009.面向MATLAB工具箱的神经网络理论与应用[M].3.合肥:中国科学技术大学出版社.
胡寿松,王执铨,胡维礼.2005.最优控制理论与系统[M].2.北京:科学出版社.
马振华,刘坤林.等.2005.现代应用数学手册-计算与数值分析卷[M].北京:清华大学出版社.
顾小丰,孙世新,卢光辉.2005.计算复杂性[M].北京:机械工业出版社.
ANDERSEN E D, YE Y.1998. A computational study of the homogeneous algorithm for large-scale convex optimization[J]. Computational Optimization and Applications, 10:243-269.
ASCHER U M, PETZOLD L R.1998. Computer Methods for Ordinary Differential Equa-tions and Differential-Algebraic Equations[M]. Philadelphia, PA:SI AM:231--299.
ASCHER U M, MATTHEIJ R M, RUSSELL R D.1995. Numerical Solution of Boundary Value Problems for Ordinary Differential Equations[M]. Philadelphia, PA:SIAM:218-225.
BAXTER B J C.1992. The Interpolation Theory of Radial Basis Functions[D]. Cambridge, UK:Trinity College of Cambridge Univeristy.
BELLMAN R.1957. Dynamic Programming[M]. Princeton, NJ:Princeton University Press:245--282.
BENSON D.2005. A Gauss pseudospectral transcription for optimal control[D]. Cam-bridge, MA:Massachusetts Institute of Technology. http://hdl. handle.net/1721. 1/28919.
BETTS J T.2010. Practical Methods for Optimal Control and Estimation Using Nonlinear Programming[M].2. Philadelphia, PA:SIAM.
BIEGLER L.2009. Technology Advances for Dynamic Real-Time Optimization[M]//. DE BRITO ALVES R M, DO NASCIMENTO O, BISCAIA-JR E C.10th International Symposium on Process Systems Engineering:Part A.[S.I.]:Elsevier. Computer Aided Chemical Engineering, vol.27.
BIEGLER L T.2007. An overview of simultaneous strategies for dynamic optimization[J]. Chemical Engineering and Processing,46(11):1043-1053.
BIEGLER L T.2010. Nonlinear Programming:Concepts, Algorithms, and Applications to Chemical Processes[M]. Philadelphia, PA:SIAM.
BIEGLER L T, CERVANTES A M, WACHTER A.2002. Advances in simultaneous strategies for dynamic process optimization[J]. Chemical Engineering Science,57(4): 575-593.
BRYSON A E.1999. Dynamic Optimization[M]. Menlo Park, CA:Addison Wesley Longman, Inc.
BUSKENS C, MAURER H.2001. Sensitivity Analysis and Real-Time Optimization of Parametric Nonlinear Programming Problems[M]//. GROTSCHEL M, KRUMKE S O, RAMBAU J. Online Optimization of Large Scale Systems. Berlin, GER:Springer Verlag: 3-16.
BYRD R H, NOCEDAL J, WALTZ R A.2006. Knitro:An Integrated Package for Nonlin-ear Optimization[M]//. DI PILLO G, ROMA M. Large-Scale Nonlinear Optimization. New York, USA:Springer-Verlag:35--59.
CAREY G F, FINLAYSON B A.1975. Orthogonal collocation on finite elements[J]. Chemical Engineering Science,30(5-6):587--596.
CHEN W, SHAO Z, WANG K, et al.2011. Random sampling-based automatic param-eter tuning for nonlinear programming solvers[J]. Industrial & Engineering Chemistry Research,50(7):3907--3918.
CUTHRELL J E, BIEGLER L T.1989. Simultaneous optimization and solution methods for batch reactor control profiles[J]. Computers & Chemical Engineering,13(1-2):49--62.
DAVIS M E.1984. Numerical Methods and Modeling for Chemical Engineers[M]. New York, NY:John Wiley & Sons:97--126. http://resolver.caltech.edu/ CaltechBook:1984.001.
DE ARAUJO A C B. GOVATSMARK M, SKOGESTAD S.2007a. Application of plantwide control to the HDA process. Ⅰ--steady-state optimization and self-optimizing control[J]. Control Engineering Practice,15(10):1222-1237.
DE ARAUJO A C B, HORI E S. SKOGESTAD S.2007b. Application of plantwide control to the HDA process. Ⅱ--regulatory control[J]. Industrial & Engineering Chemistry Research,46(15):5159--5174.
DOLAN E D, MORE J J, MUNSON T S.2004. Benchmarking Optimization Software with COPS 3.0[R].9700 South Cass Avenue. Argonne, Illinois 60439:Mathematics and Computer Science Division, Argonne National Laboratory.
DOUGLAS J M.1988. Conceptual Design of Chemical Processes[M]. New York. NY: McGraw-Hill.
DOWNS J J, VOGEL E F.1993. A plant-wide industrial process control problem[J]. Computers & Chemical Engineering,17(3):245-255.
DUA V, PERKINS J D, PISTIKOPOULOS E N.2007. Real Time Optimization[M]//. PUIGJANER L, HEYEN G. Computer Aided Process and Product Engineering (CAPE). Weinheim, GER:WILEY-VCH:577--590.
FANG X, SHAO Z, WANG Z, et al.2009. Mnemonic enhancement optimization (MEO) for real-time optimization of industrial processes[J]. Industrial & Engineering Chemistry Research,48(1):499--509.
FLETCHER R, MORTON W.2000. Initialising distillation column models[J]. Computers & Chemical Engineering,23(11-12):1811-1824.
FORNBERG B.1996. A Practical Guide to Pseudospectral Methods[M]. Cambridge CB2 1RP. UK:Cambridge University Press.
FORSGREN A, GILL P E, WRIGHT M H.2002. Interior methods for nonlinear optimiza-tion[J]. STAM Review,44(4):525--597.
FRANKE R.1982. Scattered data interpolation:tests of some method[J]. Mathematics of Computation,38(157):181--200.
GONDZIO J.1998. Warm start of the primal-dual method applied in the cutting-plane scheme[J]. Mathematical Programming,83(1):125-- 143.
GONDZIO J, GROTHEY A.2008. A new unblocking technique to warmstart interior point methods based on sensitivity analysis[J]. SIAM Journal on Optimization,19(3): 1184--1210.
HAIRER E, WANNER G.1996. Solving Ordinary Differential Equations II:Sitff and Differential-Algebraic Problems[M].2. Berlin Heidelberg, GER:Springer-Verlag.
HAIRER E, N(?)RSETT S P, WANNER G.1993. Solving Ordinary Differential Equations I:Nonsitff Problems[M].2. Berlin Heidelberg, GER:Springer-Verlag.
HARGRAVES C R, PARIS S W.1987. Direct trajectory optimization using nonlinear programming and collocation[J]. Journal of Guidance, Control, and Dynamics,10(4): 338-342.
HUANG R.2010. Nonlinear Model Predictive Control and Dynamic Real Time Optimiza-tion for Large-scale Processes[D]. Pittsburgh, PA:Carnegie Mellon University.
ISERLES A.1996. A First Course in the Numerical Analysis of Differential Equations[M]. Cambridge CB2 2RU, UK:Cambridge Univerisity Press:33-52.
JOCKENHOVEL T, BIEGLER L T, WACHTER A.2003. Dynamic optimization of the Tennessee Eastman-process using the OptControlCentre[J]. Computers & Chemical En-gineering,27(11):1513-1531.
KAMESWARAN S, BIEGLER L T.2006. Simultaneous dynamic optimization strategies: Recent advances and challenges[J]. Computers & Chemical Engineering,30(10-12): 1560-1575.
KAMIEN M I, SCHWARTZ N L.1991. Dynamic Optimization:The Calculus of Variations and Optimal Control in Economics and Management[M].2. Amsterdam, NL:Elsevier Science B.V.
KAWAJIRI Y, LAIRD C, WACHTER A.2010. Introduction to IPOPT:A tutorial for downloading, installing, and using IPOPT[Z/OL]. http://www.coin-or.org/Ipopt/ documentation/.
KUNKEL P. MEHRMANN V.2006. Differential-algebraic equations analysis and nu-merical solution[M]. Switzerland:European Mathematical Society.
LANG Y D, BIEGLER L T.2007. A software environment for simultaneous dynamic optimization[J]. Computers & Chemical Engineering,31(8):931-942.
LANG Y D, CERVANTES A M, BIEGLER L T.1999. Dynamic optimization of a batch cooling crystallization process[J]. Industrial & Engineering Chemistry Research,38(4): 1469-1477.
LUYBEN M L. TYREUS B D.1998. An industrial design/control study for the vinyl acetate monomer process[J]. Computers & Chemical Engineering,22(7-8):867-877.
LUYBEN W L, TYREUS B D, LUYBEN M L.1998. Plantwide Process Control[M]. New York, NY:McGraw-Hill.
MAI-DUY N.2005. Solving high order ordinary differential equations with radial basis function networks[J]. International Journal for Numerical Methods in Engineering,62(6): 824-852.
MAI-DUY N, TRAN-CONG T.2001. Numerical solution of differential equations using multiquadric radial basis function networks[J]. Neural Networks,14(2):185-199.
MAURER H, AUGUSTIN D.2001. Sensitivity Analysis and Real-Time Control of Para-metric Optimal Control Problems Using Boundary Value Methods[M]//. GROTSCHEL M, KRUMKE S O. RAMBAU J. Online Optimization of Large Scale Systems. Berlin, GER:Springer Verlag:17--55.
MORRISON D D, RILEY J D, ZANCANARO J F.1962. Multiple shooting method for two-point boundary value problems[J]. Communications of the ACM,5:613--614.
MOSEK-APS.2011. The MOSEK optimization toolbox for MATLAB manual (Version 6.0)[Z/OL]. Copenhagen, Denmark:[s.n.]. http://www.mosek.com.
NOCEDAL J, WRIGHT S J.2006. Numerical Optimization (Springer Series in Operations Research)[M].2. New York, NY:Springer Science & Business Media, LLC:304-354.
POLYAK R.1992. Modified barrier functions (theory and methods)[J]. Mathematical Programming,54(1-3):177--222.
RAO A V, BENSON D A, DARBY C, et al.2010. Algorithm 902:GPOPS, A MAT-LAB software for solving multiple-phase optimal control problems using the gauss pseu-dospectral method[J]. ACM Transactions on Mathematical Software,37(2):1--39.
RUTQUIST P E, EDVALL M M.2010. PROPT-MATLAB Optimal Control Software[Z/ OL]. Pullman, WA:[s.n.]. http://tomopt.com/does/TOMLAB_PROPT.pdf.
SCHWARTZ A L.1996. Theory and Implementation of Numerical Methods Based on Runge-Kutta Integration for Solving Optimal Control Problems[D]. Berkeley, CA:Uni-versity of California at Berkeley.
SEBORG D E, EDGAR T F, MELLICHAMP D A.2004. Process Dynamics and Con-trol[M].2. Hoboken, NJ:John Wiley & Sons:510-533.
VANDERBEI R J, YURTTAN H.1998. Using LOQO To Solve Second-Order Cone Programming Problems[R].[S.1.]:Princeton University.
VAPNIK V N.1998. Statistical Learning Theory[M]. New York, NY:John Wiley & Sons.
VAPNIK V N.2000. The Nature of Statistical Learning Theory[M].2. New York, NY: Springer-Verlag.
VASILE M.2006. DITAN:A tool for optimal space trajectory design[Z/OL]. http: //trajectory.estec.esa.int/Astro/3rd-astro-workshop-presentations/.
VON STRYK O.1993. Numerical solution of optimal control problems by direct col-location[J]//. BULIRSCH R, MIELE A, STOER J, et al. Optimal Control-Calculus of Variations, Optimal Control Theory and Numerical Methods, International Series of Numerical Mathematics 111. Birkhauser, Basel:[s.n.]:129--143.
VON STRYK O, BULIRSCH R.1992. Direct and indirect methods for trajectory opti-mization[J]. Annals of Operations Research,37(1-4):357--373.
WACHTER A.2002. An Interior Point Algorithm for Large-Scale Nonlinear Optimization with Applications in Process Engineering[D]. Pittsburgh, PA:Carnegie Mellon Univer-sity.
WACHTER A, BIEGLER L T.2006. On the implementation of an interior-point filter line-search algorithm for large-scale nonlinear programming[J]. Mathematical Programming, 106(1):25-57.
WRIGHT M H.2010. Interior Methods for Nonlinear Continuous Optimization[Z/OL]. Selva di Fasano, IT:[s.n.]. http://www.ba.cnr.it/ISSNLA2010/suml0.nonlinint. pdf.
WRIGHT S J.1997. Primal-Dual Interior-Point Methods[M]. I. Philadelphia, US:Society for Industrial and Applied Mathematics:224--225.
WU Y C, DEBS A S.2001. Initialisation, decoupling, hot start, and warm start in direct nonlinear interior point algorithm for optimal power flows[J]. IEE Proceedings:Gener-ation, Transmission & Distribution,148(1):67--75.