Genetic optimization is the use of programming algorithms to find the best solution to a problem. This has its origins in the work of mathematicians starting as early as the 1950s who took models they saw in biology and applied them to nonlinear problems that were difficult to solve by conventional means. The idea is to mimic biology, which evolves over the course of generations to create the fittest possible population. In programming, it is possible to simulate this process to come up with a creative solution to a problem.
Nonlinear problems can be challenging for mathematicians. An example can be seen in securities trading, where there may be a number of possible decisions which quickly branch off to create a tree of choices. To independently calculate the probabilities associated with each choice would be very time consuming. The mathematician might also miss an optimal solution by failing to combine possible choices to explore new permutations. Genetic optimization allows researchers to perform calculations of this nature in a more efficient way.
The researcher starts with a subject of interest, known as a “population,” which can be divided into individuals, sometimes known as creatures, organisms, or chromosomes. These terms, borrowed from biology, reflect the origins of this approach to programming. A computer can start to run a simulation with the population, selecting individual organisms within a generation and allowing them to intermix to create a new generation. This process can be repeated through several generations to combine and recombine possible solutions, ideally arriving at the most fit option for the given conditions.
This can be extremely resource heavy. The calculations used in genetic optimization require significant computing power to quickly compare and select a number of options and combinations simultaneously. Early research into genetic optimization was sometimes limited by available processing power, as researchers could see the potential applications, but couldn't execute complex programs. As computer power increases, the utility of this method does as well, although large and complex calculations may still require a highly specialized computer.
Researchers in the field of mathematics can work with genetic optimization in a variety of settings. Ongoing development of new formulas and approaches illustrates evolutions in mathematics as people learn about new ways to consider complex problems. Some simple genetic optimization can be seen at work in settings like software for securities traders and programming for games and virtual reality where the programmers want users to have a more natural experience.
