Differential evolution; adaptation; control parameters; real-world problems

The problem of global optimisation is necessary to solve in many research and commercial areas. There are many methods to achieve and acceptable solution, very popular are techniques inspired by nature ecosystems called Evolutionary Algorithms (EA). In 1996, a simple and powerful EA known as Differential Evolution (DE) was introduced. The DE algorithm achieves very good results in artificial and also real-world problems. DE employs several control parameters, many adaptive DE variants were proposed. In 2006, and adaptive variant of jDE employing an adaptation of F, CR parameters was proposed. In jDE, values of F are sampled for each member of population from uniform distribution U (0.1, 0.9), similary, values of CR are sampled from U (0, 1). In this experimental study, new adaptive approaches of controlling F, CR parameters in jDE are proposed. The vaules of F and CR are generated from more specific intervals whice are changed during the search process. The intervals are oriented by linear or trigonometric function. Newly proposed variants of the jDE algorithm are compared with the original jDE when solving 22 real-world problems CEC 2011