Multi-objective Cuckoo Search Algorithm And Its Applications

C uckoo search algorithm(referred as CS) is a novel heuristic algorithm, which is designed by professor Xin-she Yang in 2009 through simulating the cuckoo’s behavior of laying eggs. The CS has been applied in solving practical problems more and more successfully because of its efficient ability in exploring the search domain and converging to the global optimal solution. However, CS algorithm is suitable for solving the single objective optimization problem, but in reality, most of the questions are based on multiple and conflicting objectives.With the proposing of the evolutionary algorithms, the evolutionary algorithms for solving multi-objective problems are then developed, such as the classical Genetic Algorithm, Particle Swarm Optimization Algorithm and Bat Algorithm, even the newer Firefly Algorithm have been used to solve the multi-objective problems. This paper designed a Multi-objective Cuckoo Search Algorithm, referred as MOCS, to solving the multi-objective problems with using the high parallel searching ability and fast convergence characteristics of the CS algorithm. The results of the simulation experiment and the performance test shows that the MOCS algorithm is improved in the performance of convergence, diversity and uniformity compared with classical NSGAII algorithm. In this paper, the main research results are as follows:(1) The enhancement of the fitness function: designed a new fitness calculation function through bring the niche technology and the sharing mechanism into the basic fitness function of non-dominated sorting of the Pareto optimal solutions to enhance the fitness of non-dominated solutions which is sparse.(2) Gradually archives reduction based on the niche technology: proposed the gradually archives reduction method based on the niche technology, which ensured the uniformity and diversity of the Pareto solution set of the algorithm obtained.(3) The MOCS algorithm: applied CS algorithm to solving multi-objective optimization problems, with using the enhanced fitness function and gradually archives reduction based on the niche technology to design the MOCS algorithm. Then test the performance of the MOCS through the numerical simulation tests of 9 typical test functions.(4) The application of the MOCS algorithm: applied the MOCS algorithm to the multi-objective optimization problem of comprehensive transportation network management. And compared with MOPSO algorithm, the MOCS algorithm can find better solutions.