Before execution, we must do a correct configuration of the islands, that will intervene in the process, and of the parameters problem that we have to solve. All host used in problem execution have installed the JDK development package.
1. Then, we have to decide the number of islands that we will
use and, in which host we will make it each one of them. So, we
need the IP address for each host.
2. Later, we configure the algorithm in each island. We have
to specify the output and the configuration files, in case that
they are different for each island, others parameters, statisticals,
etc.
3. Finally, we run the islands in each host, assigning a number
for each one. The island number 1 will be the last one to execute.
Defining a correct configuration file is the most important step for solving the optimization problems. The configuration file contains all the algorithm information and the way in which the problem will be solved. The configuration file is divided in four parts:
· Main parameters:
In this part there is information on the configuration of the distributed
algorithm, for example: port address, number of islands, IP addresses...
· Problem parameters: Here it is the information on the type of problem that we going to solve (SAT ó RND) and others specific problem parameters.
· Algorithm parameters : This data defines the type of genetic algorithm used and the crossover, mutation, replacement and migration operators.
· Miscellaneous: Diverse additional parameters.