Global vs. local optimization animations


The following animated films visualizes the difference between local and global optimization strategies.

Mathematically, the problem is as follows: Given a continous function f(x,y) with real variables, find its minimum on a given domain. If the function is nice enough, then one can find local optima by classical methods. However, finding the optimum on the whole domain is usually difficult.


The story: Suppose that the deepest point on the surface on an unknown planet should be found. We send out robots, capable of sensing the altitude , but they cannot see further than 3 feet. 

The robot                                              

The robots land on the planet                

A lucky robot finds a deep hole                      

In a larger picture, we will see that the hole is far from being the deepest one!

A possible solution: climbing out of the hole and trying to find a deeper one.

Good bye, robot!












Note on use:  Save, Use, Link, but do not mirror:

These videos are copyrighted materials.

You are allowed
If you present these videos, you must  acknowledge this site as their source.

However, you are not allowed (without our explicit, written permission):