We have seen that chaos is filled
with cycles, but all of them are unstable. 
The idea of control of chaos is
very simple: 
By a small change of a
system parameter,
force the system near the desired cycle. 
Because the cycle is unstable, eventually the system will wander
away from the cycle. 
When the system has wandered too far from the cycle, apply another parameter change
to force it back to the cycle. 

How can this idea be implemented, and has it been applied to any real systems? 
First, we illustrate this method by stabilizing a fixed point of the tent map.  
Next, we describe the method of Ott, Grebogi, and Yorke
for implementing control in physical situations where no model equations are known.  
Here is an example of applying this method to control an
oscillating magnetic ribbon.  
Here is an example of applying this method to control
rabbit heart arrhythmias.  
Here is an example of applying this method to control
atrial fibrillation in human hearts.  

One possible problem is that typically parameter changes
should be small, and so we can only stabilize a cycle if an iterate lands near that
cycle. This may take some time, but sensitivity to initial conditions can be
harnessed, in a method called targeting, to force the iterates near the desired cycle. 
New applications of this idea appear frequently. Control of chaos may be the
most important impact of chaos on engineering. 