经济模型预测控制基础

Fundamentals of Economic Model Predictive Control

当前大多数复杂控制系统的目标是快速、高效、可靠地将系统控制至目标设定点。模型预测控制由于可以处理大型的受系统状态和输入严格约束的多变量系统而成为当前比较流行的控制算法。在工业生产过程中,信息和目的的分层方法不再是最佳或最理想的处理方式。最近提出的替代分层的方法是直接将经济目标作为控制系统的目标函数。在这种称为经济模型预测控制的方法中,控制器直接实时地优化过程的经济性能,而不是跟踪到设定值。其中设定值为由其他信息管理系统提供的最佳稳态设定点,该最佳稳态设定点通常是系统所有稳态中代价最小最有利可图的状态量。目的是解释如何设计这类控制系统以及用其可以实现什么样的闭环特性。涵盖以下问题:渐近平均性能、闭环的稳定性和收敛性、强耗散性。

The goal of most current complex control systems is to quickly and reliably guide the process to the target set point.Model predictive control has become a popular control method in many systems because it can handle large scale,multivariable systems subject to strictly constraints on system states and inputs.In the industrial production process,the hierarchical method of information and purpose is no longer the best or most ideal way of processing.A recently proposed alter native to layering is to directly use economic goals as the objective function of the control system.In this method called economic model predictive control,the controller directly optimizes the economic performance of the process in real time,instead of tracking the set value.The set value is the best steady state set point provided by other information management systems,and the best steady state set point is usually the least costly and most profitable state quantity among all the steady states of the system.The purpose of this article is to explain how to design such control systems and what kind of closed-loop characteristics can be achieved with them.Covers the following issues:asymptotic average performance;closedloop stability and convergence,strict dissipativity.