Scribd
Upload
55 views10 pages

Unconstrained Optimization Methods: Amirkabir University of Technology Dr. Madadi

The simplex method is an optimization technique that involves (1) forming a geometric simplex with n+1 vertices in an n-dimensional space, (2) comparing the objective function values at each vertex and moving the simplex toward the optimum point iteratively, and (3) applying reflection, expansion, and contraction steps to zigzag toward the minimum value. The method converges when the standard deviation of the objective function values at the simplex's vertices is less than a small prescribed value.

Uploaded by

ss
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
55 views10 pages

Unconstrained Optimization Methods: Amirkabir University of Technology Dr. Madadi

The simplex method is an optimization technique that involves (1) forming a geometric simplex with n+1 vertices in an n-dimensional space, (2) comparing the objective function values at each vertex and moving the simplex toward the optimum point iteratively, and (3) applying reflection, expansion, and contraction steps to zigzag toward the minimum value. The method converges when the standard deviation of the objective function values at the simplex's vertices is less than a small prescribed value.

Uploaded by

ss
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 10

Unconstrained Optimization

Methods
Amirkabir University of Technology
Dr. Madadi
Simplex method
 The geometric figure formed by a set of n + 1
points in an n-dimensional space is called a
simplex.
Compare the values of the objective function
at the n + 1 vertices of a general simplex and
move the simplex gradually toward the
optimum point during the iterative process.

(1)

(2)
Reflection
 is the vertex corresponding to the highest
value of the objective function among the
vertices of a simplex.
We can expect the point obtained by
reflecting the point in the opposite face to
have the smallest value.
Reflection
If the objective function does not have
steep valleys, repetitive application of the
reflection process leads to a zigzag path in
the general direction of the minimum
Reflection
  (3)
except (4)

is the reflection coefficient defined as


(5)
Thus will lie on the line joining and from with If lies between and
where is the vertex corresponding to the minimum function value,
(6)

is replaced by and a new simplex is started.


Reflection
Difficulties
Straddle a valley
No return can be made to points that have
just been left.
Reflection

 Difficulties
 Reject the vertex corresponding to the second worst
value instead of the vertex corresponding to the
worst function value.
 The final simplex may again straddle the minimum.
 It may lie within a distance of the order of its own
size from the minimum.
 The pattern may lead to a cyclic process.
Expansion
If
  the reflection produces a new minimum
Expect to decrease the function value further
by moving along the direction pointing from to
.
Check if expansion is successful or not.
Continue reflection process
(7)

Where is called the expansion coefficient,


defined as
(8)
Contraction
If
the
  reflection process gives a point for which for all except
, and we replace point by . Thus the new will be .
In this case we contract the simplex as follows:

(9)

where is called the contraction coefficient and is defined as


(10)

If ,we still use Eq. (9) without changing the previous point . If the
contraction process produces a point for which we replace the
point in ,…, by and proceed with the reflection process again. On
the other hand, if , the contraction process will be a failure, and
in this case we replace all by and restart the reflection process.
Convergence
The
   method is assumed to have converged whenever
the standard deviation of the function at the vertices
of the current simplex is smaller than some
prescribed small quantity , that is,

(11)

You might also like