Distributed genetic algorithm implementation by means of Remote Methods Invocation technique – Java RMI

Abstract

The aim of this work is distributed genetic algorithm implementation (so called islandalgorithm) to accelerate the optimum searching process in space of solutions. The distributedgenetic algorithm has also smaller chances to fall in local optimum. This conception depends onmutual cooperation of the clients who perform separate work of genetic algorithms on localmachines. As a tool for implementation of distributed genetic algorithm, created to produce netapplication Java technology was chosen. In Java technology, there is a technique of remotemethods invocation – Java RMI. By means of invoking remote methods, objects between theclients and the server RMI can be sent.To test the work of genetic algorithm, searching for maximum function of two variables whichpossess a lot of local maxima and can be written by means of mathematical formula was chosen.The work of the whole system depends on existence of the server on which there are registeredremote services (methods) RMI and clients, each one on a separate machine. Each of the clientshas two threads, one of them accomplishes the work of local genetic algorithm whilst the otheraccomplishes the communication with the server. It sends to the server a new best individualwhich was found by the local genetic algorithm and takes the server form with the individuals, leftthere by other clients.To sum up there was created an engine of distributed genetic algorithm which searches themaximum of function and after a not large modification can be used to solve every optimizationproblem.