Improving Cuckoo Algorithm And Particle Swarm Algorithm Fusion Research

The optimization problems are always closely related to the real world life. Inrecent years, optimization problems have been widely applied to the field of daily lifeand engineering calculation. Modern optimization problems with growing complexitypose a great challenge to the traditional numerical calculation methods. Therefore,researchers turn to solve complex optimization problems by relying on heuristicalgorithm, such as genetic algorithm (GA), particle swarm optimization algorithm(PSO), differential evolution algorithm (DE) and so on. For the optimization problems,this paper made some modifications to the Cuckoo Search (CS) algorithm and PSO. Themain work is as follows:1. A new Self adaptive Cuckoo Search Algorithm. The cuckoo algorithmcompletely depends on the random walk strategy, so that it does not have a fastconvergence speed, and with the increase of the iterative steps, the search activity willdecrease. Thus, we proposed a new adaptive cuckoo search algorithm. In the algorithm,a self adaptive parameter control strategy is used to adjust the core parameters of the CSalgorithm, thereby enhancing the search ability of CS algorithm. In addition, a novelmutation technique is utilized to guarantee the CS diversity. Finally, several representivetest functions show that, the algorithm not only improves the search accuracy tests, butalso speeds up the convergence speed of the CS algorithm.2. Particle Swarm Optimization with Levy Flights. In order to improve the problemof premature in particle swarm optimization, the paper proposed a technique byintroducing levy flights. After finishing the update of the particle position, PSO does notdirectly calculate the objective function value, but uses levy flights to change thedirection of the particle and prevent the particle from local optimal value. Simulationresults show that, the algorithm has excellent optimization.