Big data is extremely beneficial. However, engineers and scientists are unable to utilize this data without the help of complex AI algorithms.
A genetic algorithm is based on the Darwinian principle of natural selection. Its purpose is to evolve and quickly solve optimisation problems.
There are many applications for Genetic Algorithms across many industries like air travel, trading and data security.
In air travel, a genetic algorithm optimises shape, minimises wing weight and optimises fuel weight. This all improves the overall efficiency of the airplane.
In security, GA’s are used for encrypting sensitive data and protecting copyrights. Hackers then create a more complex GA to beat that encryption. The cycle repeats. Possibly forever.
In robotics, a genetic algorithm can be programmed to search for a range of optimal designs for each specific use. It can also return results for entirely new types of robots, ones that can perform multiple tasks and have more general applications.
How Does A Genetic Algorithm Work?
So in Darwin’s theory of Natural Selection, three main principles necessary for evolution to happen are :
- Variation — There must be a variety of traits present in the population or a means with which to introduce a variation.
- Selection — There must be a mechanism by which some members of the population can be parents. Passing down their genetic information and some do not.
- Heredity — There must be a process in place by which children receive the property of their parent
So keep that in mind as we move forward.
In a simulation, let’s say we want the computer to generate a “creature” that can walk the furthest. Firstly, the initial population is randomly generated. Each creature has its own unique traits (variation).
In the simulation, the creatures are made to walk and the distance each creature walks is calculated. Two creatures that performed well are randomly selected (Selection). Passing on desirable traits. Survival of the fittest!
In the mating process, a new creature is generated (Heredity). The child gets some traits from one parent and some from the other. A random mutation occurs which slightly changes some of these traits.
By using this process again with several creatures, the next generation is created. The steps are repeated multiple times. Good traits are retained and bad ones are lost.
The creatures are able to walk further after each generation. The evolution stops when there is no significant increase in average distance travelled by creatures from two generations. The genetic algorithm has done its job!
So, I think its pretty easy to see how genetic algorithms can be applied to air travel. After each generation, the planes fly further. Perhaps, carrying more passengers and using less fuel. Eventually, you have a much more optimised aircraft that can save airlines millions of dollars a year in efficiency.