Genetic and Evolutionary Algorithms for PSP - Digital Signal Processing in Power System Protection and Control - page 271

Chapter 13
Genetic and Evolutionary Algorithms
for PSP

In this chapter the procedures originating from biological evolution are presented
that can be applied e.g. for intelligent searches and optimization in power systems.
The procedures belong to the so-called ''biological programming'' family, which is
not limited to the genetic algorithms (GA) only. In wider sense the neural networks
described in Chap. 12 , being an analogy of human brain, are also good examples of
this family.
The procedures presented below follow from the Darwin theory of evolution
of species. Many examples prove that Darwinian mechanism generates a kind of
optimization process. Simplifying, one can say that in real world only the best
individuals from given population have a chance to survive, and this is forwarded
to the next and further generations of offsprings, which become better and better
adjusted to the surrounding world. As such the mechanism should deliver the best
solutions to given problem (here—survival), which is now tried to be transferred to
some other domains, including physics, economy and natural and technical

sciences.
It can be stated that in many domains one has to deal with the classical problem
of optimization. Economy particularly has become a specialist of that field.
Generally speaking, a large part of mathematical development during the XVIII
century dealt with that topic (remember those always repeated problems where you
had to obtain the derivative of a function to find its extremes). One can surely
admit that purely-analytical methods widely proved their efficiency; nevertheless,
they suffer from an insurmountable weakness: reality rarely obeys to those won-
derful differentiable functions that you try to use to describe the world around.
John Holland, from the University of Michigan began his work on (GA) at the
beginning of the 1960s. A first achievement was the publication of Adaptation in
Natural and Artificial System in 1975 [ 12 ]. Both genetic and evolutionary tech-
niques deliver their solutions performing statistical search within a pre-defined
population. It is important to understand that the functioning of such an algorithm
does not guarantee success. We are in a stochastic system and a genetic pool may
be too far from the solution, or for example, a too fast convergence may halt the