|
| "Optimization Solution" |
|
| Differential Evolution Optimization Method |
|
|
MIN: |
Iterations + BestValue
|
BestMember
|
|
MAX: |
|
|
IMPLEMENTATION
Differential Evolution Method Optimization
|
This method may be considered as a random-search technique. Stochastic methods are efficient techniques for
finding the global minimum of a function with multiple variables. In general, thse type of optimization techniques are
inspired by biological processes and are called evolutionary algorithms.
Several advantages of using evolutionary algorithms over the traditional optimization methods include:
- Optimize with continous or discrete variables.
- Usually do not require derivative information.
- Deal with a large number of variables.
- Work with numerically generated data, experimental data, or analytical functions.
Algorithm Creation
Using the "peaks" function, which is used extensively in Matlab:
|
f(x,y) = 3(1-x)2 e-x2 - (y + 1)2 - 10( 1/5x - x3 - y5)
e-x2
|
This function has several local minima and maxima. It is also possible to find the global maximum for the peaks function by
simply returning the negative value of the function.
Testing the Differential Evolution Method
To test it out we find the the minimum and maximum points of the peaks function. We ran the program using different iteration
parameters(4). The user can manipulate and test it further by manually entering new iterations.
Supporting code and methods are not shown.
|
|
|
|