Adaptive Cuckoo Search Algorithm Based on Dynamic Adjustment Mechanism

Abstract

Cuckoo Search (CS) algorithm, a simple and effective global optimization algorithm, has been widely used to deal with practical optimization problems. So as to improvethe standard cuckoo search algorithm, such as slow convergence and easy convergence to local optimal value, an Adaptive Cuckoo Search algorithm on the basis of Dynamic Adjustment Mechanism (ACSDAM) has been proposed. Based on exponential function and logarithmic function, the dynamic adjustment is made for updating step size and discovering probability. During the optimization process, updating step size and discovering probability of each nest are adjusted according to the number of iterations of each nest, so as to equilibrate the global detection and local capacity of the algorithm. Then 23 standard test functions will be selected for a simulation experiment, and compared with other CS variant algorithms, ACS-DAM effectively improved the rate of convergence and the algorithmic precision. ACS-DAM algorithm was employed to optimize the Support Vector Machine (SVM). The experiment proves that the convergence rate with ACSDAM is better than that with CS obviously and ACS-DAM has stronger optimization ability and higher efficiency than CS.