**6. References**


**15** 

*South Africa* 

**Optimal Design of Power System Controller** 

Genetic Algorithms (GAs) have recently found extensive applications in solving global optimization problems (Mitchell, 1996). GAs are search algorithms that use models based on natural biological evolution (Goldberg, 1989). They are intrinsically robust search and optimization mechanisms and offer several advantages over traditional optimization techniques, including the ability to effectively search large space without being caught in local optimum. GAs do not require the objective function to have properties such as

In the last few years, Genetic Algorithms (GAs) have shown their potentials in many fields, including in the field of electrical power systems. Although GAs provide robust and powerful adaptive search mechanism, they have several drawbacks (Mitchell, 1996). Some of these drawbacks include the problem of "genetic drift" which prevents GAs from maintaining diversity in its population. Once the population has converged, the crossover operator becomes ineffective in exploring new portions of the search space. Another drawback is the difficulty to optimize the GAs' operators (such as population size, crossover and mutation rates) one at a time. These operators (or parameters) interact with one another in a nonlinear manner. In particular, optimal population size, crossover rate, and mutation rate are likely to change over the course of a single run (Baluja, 1994). From the user's point of view, the selection of GAs' parameters is not a trivial task. Since the 'classical' GA was first proposed by Holland in 1975 as an efficient, easy to use tool which can be applicable to a wide range of problems (Holland, 1975), many variant forms of GAs have been suggested often tailored to specific problems (Michalewicz, 1996). However, it is not always easy for the user to select the appropriate GAs parameters for a particular problem at hand because of the huge number of choices available. At present, there is a little theoretical guidance on how to select the suitable GAs parameters for a particular problem (Michalewicz, 1996). Still another problem is that the natural selection strategy used by GAs is not immune from failure. To cope with the above limitations, an extremely versatile and effective function optimizer called Breeder Genetic Algorithm (BGA) was recently proposed (Muhlenbein, 1994). BGA is inspired by the science of breeding animals. The main idea is to use a selection strategy based on the concept of animal breeding instead of "natural selection" (Irhamah & Ismail, 2009). The assumption behind this strategy is as follows: "*mating two individuals with high fitness is more likely to produces an offspring of high fitness than mating two randomly selected individuals*".

continuity or smoothness and make no use of hessians or gradient estimates.

**1. Introduction** 

**Using Breeder Genetic Algorithm** 

*University of Cape Town Private Bag., Rondebosch 7701* 

K. A. Folly and S. P. Sheetekela

