Fuzzy system optimization using CUCKOO algorithm

Cuckoo Algorithm (COA) which has been improved by Ramin Rajabioun in 2011 is a new heuristic searching algorithm. In this thesis study, it has been studied on investigating performance of COA and adapting it to various applications. Performance of the algorithm has been evaluated in the framework of function optimization and dynamic system identification problem based on fuzzy logic. In the study, COA has been used while modeling of sample dynamic systems are being done. Modifing COA algorithm, two new algorithms D-COA and Ü-COA have been defined and these three different COA algorithms have been compared statistically on dynamic system identification problems. ANFIS fuzzy model has been used as a modeling tool. Obtained ANFIS model for each system has been evaluated with different data set (test set) in order to investigate more detailed optimization performance of COA on fuzzy network structure. At the second stage of this thesis study, the results obtained by COA, D-COA and Ü-COA have been compared with the result obtained by ABC and PSO for each problem. According to the evaluation results, it is identified that D-COA run approximately 5% more speed than COA.