GEATbx: Main page  Tutorial  Algorithms  M-functions  Parameter/Options  Example functions 
Previous: Contents:
2.3 Harvest system Contents

2 Optimization of dynamic systems


2.4 Push-cart system

The push-cart system [Mic92] is a two-dimensional system described by the following equations:

x1(k+1)=x2(k), k=1,2,...,N,
x2(k+1)=2·x2(k)-x1(k)+(1/N^2)·u(k).
The objective function for minimization is therefore defined as:

f(x,u)=-x1(N+1)+1/(2·N)·sum(u(k)^2), k=1:N.

The exact solution can be analytically found by:

Minimum=-(1/3)+((3·N-1)/(6·N^2))+1/(2·N^3)·sum(k^2), k=1:N-1.

Figure 1 shows the control vector for the push-cart system with N=20.

Fig. 1: Optimal control vector for the push-cart system with N=20

Optimal control vector for the push-cart system with N=20

This function is implemented in the m-file objpush.m.


GEATbx: Main page  Tutorial  Algorithms  M-functions  Parameter/Options  Example functions 

This document is part of version 3.3 of the GEATbx: Genetic and Evolutionary Algorithm Toolbox for use with Matlab - www.geatbx.com.
The Genetic and Evolutionary Algorithm Toolbox is not public domain.
© 1994-2000 Hartmut Pohlheim, All Rights Reserved, (support@geatbx.com).