Modifications And Application Research On Genetic Algorithm And Particle Swarm Optimization Algorithm

As optimization problems exist widely in all domains of scientific research and engineering application, research on optimization method is of great theoretical significance and practical value. Since traditional top-down research methods show up a lot of shortcomings, they are difficult to solve more and more complex problems in modern society. The new intelligent optimization algorithms based on creatures and natural ecological systems make the optimization problems with high degree of complexty solved perfectly. Under the background, biologically-inspired intelligent computations which are entirely different from the classed optimization methods have emerged. The paper gives a comprehensive study on genetic algorithm and particle swarm optimization from the aspects of algorithm mechanism, modifications and their applications. The main achievements of this dissertation include:(1) This paper proposed two improved evolutionary algorithms based on two different improved strategies for solving constrained optimization problems, and corresponding complexity and convergence were also analyzed. The two improved strategies are:(1) Search space of constrained dominance problems with high dimensions is compressed into two dimensions. A linear fitness function based on mathematic analysis is deduced in two dimension space so as to fast evaluate fitness value of each individual in population. A crossover operator based on density function and a new mutation operator is developed to extend the search space and extract the better solution. In addition, an average linkage based on hierarchical clustering method is introduced into the proposed method to maintain the number of individuals in Pareto set. (2) Neighborhood function criterion (NFC) is applied to selection process. By using NFC, good individuals can be distinguished from the population and ensure the diversity of the population. On the other hand, the preservation method for Pareto candidate solution set based on NFC is incorporated into the proposed algorithm. This method can maintain the diversity of Pareto candidate set effectively. The complexity of time and space in the proposed algorithm is analyzed. For a set of benchmark problems, the experimental results showed that the proposed algorithms could search more effectively and provide good performance both in convergence and in diversity of solutions.(2) Because genetic algorithm is of good parallel search, and this characteristic is suitable to for solving nonlinear programming problem. This paper proposed an improved genetic algorithm (IGA) to handle nonlinear programming problems. A novel selection strategy and local searching shceme are incoporated into the proposed method. The main idea on IGA:each individual in selection process is represented as a three-dimensional feature vector composed of objective function value, the degree of constraints violations and the number of constraints violations. We can distinguish excellent individuals through two indices according to Pareto partial order. Additionally, IGA incorporates a local search process into selection operation so as to find feasible solutions located in neighboring areas of some infeasible solutions. Experimental results over a set of benchmark problems demonstrated that proposed IGA had better robustness, effectiveness and stableness than other algorithm reported in literature. In addition, image segmentation problem is considerd as nonlinear probleming problems with constraint conditions. This paper proposed an improved scheme based on genetic algorithm for fastening threshold selection in multilevel MCET. Firstly, image segmentation problem is considered as an optimization problem by a recursive programming technique. Secondly, this scheme uses a recursive programming technique to reduce computational complexity of objective function in multilevel MCET. Then, a genetic algorithm is proposed to search several near-optimal multilevel thresholds based on this recursive programming technique. Empirically, the multiple thresholds obtained by our scheme are very close to the optimal ones via exhaustive search. The proposed method was evaluated on various types of images, and the experimental results show the efficiency and the feasibility of the proposed method on the real images.(3) This paper proposed a double center particle swarm optimization algorithm (DCPSO). Particle swarm optimization algorithm is a new promising swarm intelligence optimization technology, and it has been extensively applied to solve all kinds of complex optimization problems because of its advantage of simpler theory, less parameter and better performance. However, each particle's individual minimum and population's minimum are two major factors to affect the algorithm's convergence speed and precision. We analyze particle's flying trajectory and introduce general center particle and special center particle in the DCPSO, and consequently improves PSO algorithm's converging speed and precision without increasing computing complexity. The proposed algorithm was tested in two ways:a fixed iteration number test and fixed time length test. Experiment results showed that the proposed algorithm was feasible and efficient.(4) This paper proposed a modified differential evolutionary algorithm (MDE) to solve multiobjective optimization problem. Differential Evolutionary (DE) is a simple, fast and robust evolutionary algorithm for complex optimization problems. There are some different points between MDE and traditional DE:child individual mutation and its selection process for new child; MDE allow infeasible solutions of population to participle in mutation process that is similar to particle velocity updating scheme in PSO, and a fast nondominated sorting and ranking selection scheme of NSGA-â…¡proposed by Deb is incorporated in individual's selection process. We finally obtained a set of global optimal solutions by continuing optimization on infeasible and feasible individuals. Simulated experiments showed that the obtained solutions present good uniformity of diversity, and approach the true frontier of Pareto. Its convergence is also idea.