Two Improved Algorithms For Multi-objective Optimization Based On Cuckoo Search

The recently proposed Cuckoo search algorithm is an evolutionary algorithm based on probability.It surpasses other algorithms in solving the multi-modal discontinuous and nonlinear problems.Searches made by it are very efficient because it adopts Levy flight to carry out random walks.This paper proposes two improved versions of Cuckoo search for multi-objective problems(PMOCS and EMOCS).Combined with non-dominated sorting,crowding distance and Levy flights,Elite Strategy and Pareto Strategy are applied to improve the algorithm.Then numerical studies are conducted to compare the algorithm with standard MOCS,DEMO and NSGA-? against multiple test functions.Here is the final conclusion:The two improved Cuckoo search algorithms proposed by this paper convergence rapidly and have good diversity among the solutions.They have a lot of potentials.