Home > Machine Learning, Projects > The project ‘PGASystem’ (Parallel Genetic Algorithms System).

The project ‘PGASystem’ (Parallel Genetic Algorithms System).

October 12, 2011 Leave a comment Go to comments

The project “PGASystem” (Parallel Genetic Algorithms System) is an under development system based on the client / server architecture and can be used to implement and study of parallel genetic algorithms.

Essentially, it is a system composed of a client and a server and can be used to implement the model of parallel evolving populations with different evolutionary behaviors, by simulating the parallel evolution of populations in nature.

According to this model, the system retains a number of parallel evolving populations which evolve through genetic algorithms with different evolutionary behaviors, i.e. with different selection methods, different operators, genetic mechanisms and parameters.

It is worth mentioning that parallel genetic algorithms were born out of the need to overcome some disadvantages of simple genetic algorithms and to boost their capacity to tackle difficult problems.

The disadvantage of genetic algorithms as stochastic algorithms is the uncertainty of finding the optimal solution. This fact creates the need to perform several experiments in genetic algorithms for each problem so as to increase the chances of any of the populations to find the area of optimal solution and converge to it.

On the other hand, genetic algorithms are relatively time-consuming algorithms, due to the iterations required for convergence, so the serial execution of many experiments would require too much time. Parallel genetic algorithms, however, went one step further with more accurate modeling of parallel evolving populations that occur in nature.

For example, if we consider two cities with populations which evolve in parallel will see that these populations are not isolated from each other but there are several interactions between them due to the existence of communication and transportation between cities. This information exchange occurs between the two subpopulations with either simple communication or segments of the population migrating from one city to another.

Modeling this interaction of the populations, parallel genetic algorithms adopted processes of communication between people as well as migration, i.e. transferring solutions from one population to another. This exchange of information between the populations proves itself positive for the evolution of individual populations and contributes to better performance.

  1. No comments yet.
  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: