Engineering Optimization

14Genetic Algorithms and Evolutionary Computation

14.1 Introduction

Genetic algorithms (GA) are mimetic approaches to the “intelligence” behind natural evolution embodied by random selection and survival of the fittest, which seems to direct evolution in biological species. These algorithms make progress toward an optimum in a logic that mimics our understanding of genetic evolution. Hence, the term evolutionary computation, or evolutionary optimization, is often used. However, in some disciplines evolutionary optimization means incremental process set point or controller coefficient adjustment in a manner similar to a CHD search. Accordingly, I prefer the term GA over “evolutionary.”

The concept from genetics is that genes in the DNA of an individual define attributes or traits. Genes are functions within the DNA. Chromosomes are sequences of genes that define a trait such as height or eye color. To illustrate the concept, the cell entries in Table 14.1 represent genes, and the grouping of the first four comprises the chromosome that will relate to height. Similarly, the last two genes that are indicated comprise the chromosome for eye color. This is a concept and not the biological reality.

Table 14.1 Example of two chromosomes.

With two genes for tall and two for short this individual would be of medium height.

For brevity, the gene labels will just use the leading letters ...