genetic algorithms in artificial intelligence

genetic algorithms in artificial intelligence

genetic algorithms:

Genetic Algorithms initiated and developed in the early 1970s by John Holland (1973; 1975) are worthodar search and optimization algorithms, which mimic some of the processes of natural evolution. GAs perform directed random searches through a given set of alternatives with the air of finding the best alternative with respect to the given criteria of goodness. These criteria are required to be expressed in terms of an objective function which is usually referred to as a fitness function.

Fitness is defined as a figure of merit, which is to be either maximized or minimized. It is further required that the alternatives be coded in some specific finite length which consists of symbols from some finite alphabet. These strings are called chromosomes and the symbols that form the chromosomes are known as genes. In the case of binary alphabet (0, 1) the chromosomes are binary strings and in the case of real alphabet (0-9) the chromosomes are decimal strings.

.. Starting with an initial population of chromosomes, one or more of the genetic inheritance operators are applied to generate offspring that competes for survival to make up the next generation of population. The genetic inheritance operators are reproduction, cross over, mutation, inversion dominance, deletion, duplication, translocation, segregation, speciation, migration, sharing, and mating.

However, for most common applications, reproduction, mating (cross-over), and mutation are chosen as the genetic inheritance operators. Successive generations of chromosomes improve in quality provided that the criteria used for survival is appropriate. This process is referred to as Darwinian natural selection or survival of the fittest.

Reproduction which is usually the first operator applied on a population selects good chromosomes in a population to form the mating pool. A number of reproduction operators exist in the literature (Goldberg and Deb, 1991). Cross over is the next operator applied. Here too, a . number of cross over operators have been defined (Spears and De Jong, 1990). But in almost all cross over operators, two strings are picked from the mating pool at random and some segments of the strings are exchanged between the strings. Single point cross over, two point cross over, matrix cross over are some of the commonly used cross over operators. It is intuitive from the construction that good substrings from either parent can be combined to form better offspring strings.

en compared to cross over is used sparingly. The operator changes a I to a 0 and vice versa with a small probability P - The need for the operator is to keep the diversity of the population.

Though most GA simulations are performed by using a binary.coding of the problem parameters, real coding of the parameters has also been propounded and applied (Rawlins, 1990). GĀS have been theoretically and empirically proven to provide robust search in complex space