多变量系统中多参数多目标满意优化研究

基金项目: 国家自然科学基金资助项目（69574026）；教育部高等学校骨干教师计划资助项目（教技司[2000]65号）

详细信息

SATISFACTORY OPTIMIZATION OF MULTI-PARAMETERS AND MULTI-OBJECTIVES IN MULTIVARIABLE SYSTEM

1.
School of Electrical Engineering, Southwest Jiaotong University, Chengdu 610031, China
2.
Southwest China Research Institute of Electronic Equipment, Chengdu 610036, China

摘要: 本文提出了多变量系统的多参数多目标满意优化方法，将系统性能指标要求的满意设计与控制器参数优化融为一体统一考虑，通过设计性能指标满意度函数和系统综合满意度函数，构造出多变量系统满意优化模型，并用改进遗传算法实现其满意优化．仿真结果显示该方法可获得比传统优化方法更满意的综合性能指标，表明了该方法的有效性和实用性．
- 优化 /
- 满意优化 /
- 控制系统 /
- 遗传算法
Abstract: This paper proposes a satisfactory optimization method for the multivariable control system with multi-parameters and multi-objectives. By designing satisfactory rate functions of performance criterions and comprehensive satisfactory rate functions of the system, satisfactory optimization model of multivariable control system is constructed, in which satisfactory design of performance criterions and optimization of controller parameters are considered together. Then, the satisfactory optimization design procedure is given in detail by using improved genetic algorithm. Simulation results show that better performances can be obtained with satisfactory optimization method than those with traditional methods, indicating that the method introduced is effective and practical.
- optimization /
- satisfactory optimization /
- control system /
- genetic algorithm

[1]	Jin W D, Chen L, Li G, et al. An approach for the optimization of control system parameters via composite satisfactory extent dunction[A]. The 5th Inter. Conf. on Automation Technology and 1998 Inter. Conf. of Production Research (Asia Meeting) [C]. Teipei:1998.325～330.
[2]	席裕庚. 复杂工业过程的满意控制[J]. 信息与控制, 1995, 24(1): 14～20.
[3]	欧阳黎明. MATLAB控制系统设计[M]. 北京:国防工业出版社,2001.168～203.
[4]	杨智民,王旭,庄显义. 遗传算法在自动控制领域中的应用综述[J]. 信息与控制,2000,29(4):329～337.
[5]	Youssef H, Sait S M, Adiche A. Evolutionary algorithms, simulated annealing and tabu search: a comparative study[J]. Engineering Application of Artificial Intelligence, 2001,14(2):167～181.
[6]	Takahashi H, Nakajima M. Evolutional design and training algorithm for feedforward neural networks [J]. IEICE Trans. Information & Systems, 1999,E82-D(10):1384～1392.