Modification of species-based differential evolution for multimodal optimization

• Main Authors: Al Idrus, Said Iskandar, Syahputra, Hermawan, Firdaus, Muliawan
• Format: Conference or Workshop Item
• Language: English
• Published: 2015
• Subjects: QA Mathematics
• Online Access: http://repo.uum.edu.my/16809/1/14.pdf; http://repo.uum.edu.my/16809/; http://doi.org/10.1063/1.4937031

At this time optimization has an important role in various fields as well as between other operational research, industry, finance and management. Optimization problem is the problem of maximizing or minimizing a function of one variable or many variables, which include unimodal and multimodal functions. Differential Evolution (DE), is a random search technique using vectors as an alternative solution in the search for the optimum. To localize all local maximum and minimum on multimodal function, this function can be divided into several domain of fitness using niching method. Species-based niching method is one of method that build sub-populations or species in the domain functions. This paper describes the modification of species-based previously to reduce the computational complexity and run more efficiently. The results of the test functions show species-based modifications able to locate all the local optima in once run the program.