At present,many experts and scholars use natural computing to deal with multi-objective optimization problems and get the desired results.This paper proposes an improved algorithm to deal with multi-objective optimization problems.Membrane calculation is a new research direction of natural calculation.Based on the theory of membrane calculation,this paper combines it with the emerging swarm intelligence algorithm fireworks explosion algorithm,and introduces the elite reverse learning mechanism to improve the accuracy of the algorithm.At the same time,it introduces the non-dominated sorting in NSGA-II algorithm.The two mechanisms of crowded distance calculation are used to enhance the diversity of the population and the speed of the algorithm.Moreover,an external archive set is created outside the population.The archive set is used to store the non-dominated solutions obtained by the algorithm in the search process,thereby increasing the rate of the algorithm.The fireworks explosion algorithm is a swarm intelligence algorithm that has emerged in recent years.The fireworks explosion algorithm was discovered by Tan et al.in 2010.By simulating the mechanism of explosion point explosion during the fireworks explosion,a new fireworks search was proposed.algorithm.Since the initial parameters of the algorithm are relatively small,the execution process is relatively simple,so it has certain feasibility in dealing with multi-objective optimization.Since the fireworks explosion algorithm has been applied to many times to deal with multi-objective optimization problems and achieved certain results,it has attracted the attention of many scholars.After combining the membrane calculation with the fireworks explosion algorithm,in order to fully verify the feasibility of the algorithm,two target test functions are used: ZDT1,ZDT2 and ZDT3 and three target test functions: DTLZ1,DTLZ3 and DTLZ6 to simulate the algorithm.Experiments and proved that the algorithm is feasible.At the same time,in order to verify the performance of theproposed algorithm,six algorithms(such as MOPSO algorithm,NSGA-II algorithm,PESA-II algorithm and SPEA-II algorithm)are selected here to compare with the proposed algorithm.IGD is used as an evaluation index and the IGDs of seven multi-objective optimization algorithms are compared.Experiments show that the proposed algorithm has better performance than the other six algorithms in dealing with multi-objective optimization problems.The algorithm also has a good performance in terms of convergence speed and diversity of non-dominated solutions.Compared with the other six algorithms,the non-dominated solution obtained by this algorithm can better approximate the real Pareto front,which lays a theoretical foundation for the processing of multi-objective optimization problems.Feature selection is an important technique in radar signal recognition.Research on feature selection has always been a key and difficult research direction in electronic countermeasures.This paper briefly describes the feature selection of radar signals.A multi-objective optimization function for radar signal feature selection is constructed by correlation function and redundancy function.The above proposed algorithm is applied to the constructed feature selection objective function,then the radar signal is selected.Finally,the correct rate of the algorithm on the signal data set is tested by FCM(Fuzzy C-means)clustering algorithm. |