Genetic Algorithm (GA)
Genetic Algorithm has turned out to be excellent optimization tool. It does not require an exact mathematical model of the physical system. Furthermore, speed and robustness are prominent properties of this scheme. The proposed design strategy for optimizing the transformer core parameters is based on GA. It is a stochastic search algorithm that emulates biological evolutionary theories to solve optimization problems. It enables parallel search from a population of points. Therefore, it has the ability to avoid being trapped in local optimal solution.
Basic terms used in GA:
It can be consider as all possible solutions for a given problem and, in fact, is analogues to the population of human beings.
It can be defined as the function which takes a desired solution as an input and provides us with the suitability of this solution as an output. In some cases, it is same as objective function while could be different in other cases though.
These are two fundamental genetic operators which actually alter the genetic composition of the new solutions related to fitness function.
It is kind of a test. We set up some conditions and if the solution satisfies these conditions then we consider it as the best one else process starts again and we tend to find new solution.
Flow Diagram of GA:
A simple block diagram that explains the operation of GA optimization is given in Figure below:
Pros and Cons of Genetic Algorithm:
- No derivatives needed
- Easy to parallelize
- Can escape local minima
- Works on a wide range of problems
- Need much more function evaluations than linearized methods
- No guaranteed convergence even to local minimum
- Have to discretize parameter space