A Hybrid Method Of Global Optimization Based On Tabu Search And Differential Evolution

Evolutionary algorithms have emerged as a result of observing and trying to copy natural processes occurring in the world of living organisms, in particular evolution associated with it of selection(natural selection) populations of living beings. The idea of evolutionary algorithms was made in the late sixties- early seventies of the twentieth century. It was founded on the desire to be realized in the form of a computer program algorithm that will solve complex problems the way nature does through evolution.The aim of this work is the study of differential evolution(DE) algorithm, combining it with traditional local search techniques as a way to optimize their performance and complexity. Global optimization was chosen as the problem being solved, because it is very important in the actual production.In this paper we would like to introduce a new hybrid differential evolution tabu search(TS) method and test it on global optimization problems. The test results would be presented in graphs and tables. Each aspect of hybrid differential evolution tabu search method would be described, optimal parameters would be retrieved.The test results show improvement comparing to the parent methods, which means this algorithm could be researched and improved to get even better performance.The main instrument for practical research was chosen MATLAB because it has many builtin functions and toolbars for solving problems in inherit programming and their parallel execution.