Population Growth Prediction with Genetic Algorithm (GA)

In this study, it is aimed to predict population growth by using software with visual interface C # programming language which was developed in the .NET environment by making use of genetic algorithms which is a sub-discipline of artificial intelligence. In this context, based on the counting information conducted by Turkish Statistical Institute was taken as basis and the developed software was able to estimate the population growth for the following years. In the scope of the study, in terms of speed and performance Genetic Algorithm with Real Coding was used instead of binary coded genetic algorithm; Tournament Selection Method was used as the selection method. Because Genetic Algorithms (GA) is binary coded algorithms; the expression of the parameters with "1" and "0" increases size of the chromosomes considerably and that results limited sensitivity. However, it is more advantageous to use genuine coded GA, which can code with real numbers instead. The real code GA is both more accurate and takes up less space in the PC memory. In addition, in the developed software in order to avoid the complexity and difficulties of use of existing software, population growth forecast is being made initially after user enters the number of populations, the number of iterations, the crossover rate and the population counts of previous years information obtained from Turkish Statistical Institute. Within the context of this study, data collected from Turkish Statistical Institute for 2016 estimates of population growth were made for Turkey and Konya provinces and estimated increase rate was compared with the actual data. The estimated population increase rate for 2020 for Turkey and Konya province is given thereafter. In addition to that, the developed software is designed for flexibility and structure that can predict population growth for continents, countries, cities, borders and even villages with available data