Genetics algorithms are bionspired metaheuristics in the field of genetics used to solve problems where an efficient algorithm is usually not known; is very applied in optimization problems, where sub-optimal soluction to the problem needs to be found. There are many ways to use the concepts of genetic algorithms in solving problems, such as NP-Difficult; due to is many customizable parameters, being possible to configure them according to the problem in question. Parallel genetic algorithms are commonly used to solve problems in a way that is closer to the evolutionary analogy of species, since several populations, sometimes isolated, develop independently and can therefore exploit a larger search space than the serial genetic algorithms; and can also Interact. One such interaction mechanism is known as migrations techniques or policies. In this work, migrations policies will be presented in the literature, a proposed policy and some other policies resulting from the variation of these, the objective is to compare the performance of the migrations policies presented here, in the optimization of four multimodal function in order to determine the most eficiente and suggest future improvements for them.
|
