Thesis (M.Sc.) - Cairo University - Faculty of Computers and Information - Department of Operation Research and Decision Support

Bio - inspired optimization algorithms are set of algorithms that imitate natural phenomena aiming to {uFB01}nd the optimal solution for a complex problem. They play a signi{uFB01}cant role in many di{uFB00}erent applications. One of the most e{uFB00}ective global search optimization algorithms in bio - inspired set of algorithms is particle swarm optimization (PSO) algorithm. PSO is known by its fast convergence comparing to many global search optimization algorithms. The main disadvantage of PSO is its dependency on many control parameters; Wrong choice for any of these parameter values may lead to the divergent of the algorithm. Thus, searching for another global search optimization algorithm that doesn{u2019}t have this problem is required. Di{uFB00}erential Evolution (DE) algorithm is one of the candidates. DE is a stochastic, population based optimization algorithm that depends on few numbers of parameters. On the other hand, least squares support vector machine (LS - SVM) is a machine learning algorithm that is used for classi{uFB01}cation by {uFB01}nding the optimal hyper-plane that separates various classes. LS - SVM is a parameters dependent algorithm, which means that it is so sensitive to the changes in the values of its parameters

Issued also as CD