| Evolutionary algorithms are new kinds of modern optimization algorithms that are inspired by Principle of nature evolution. As new kinds of random search algorithms, they have some advantages over the traditional optimization algorithms, and are of the great importance and have a wide range of applications. In the paper, firstly the basic mathematical theory and application fields of the genetic algorithms are summarized, and some improved evolutionary algorithms for global optimization are proposed.For detail, we conclude them as follows:1. Firstly, a genetic algorithm for global optimization based on uniform distribution encoding and an orthogonal compact genetic algorithm are presented.The former algorithm adopts a novel encoding strategy,which each chromosome represents a multivariate uniform distribution region in the search space,and the algorithm searches the space in a region-by-region manner,thus the efficiency is improved.The population of the orthogonal compact genetic algorithm is represented using a probability vector, and the probability vector is always adjusted toward the elitism.This algorithm uses much less memory and is very simple.Through the numerical experiments,we find that the two proposed algorithms are effective and very efficient.2. A self-adaptive genetic algorithm for the problem of assignment model for fruitful discussion is proposed. Orthogonal design and matrix encoding are used to design the algorithm.Compared to the method based on integer programming, it is simple and easy to implement.The efficiency and stability of the algorithm are show by the numerical results.3. Particle swarm optimization is simple, effective and has a few parameters,but the key pamareters are of importance to the efficiency of the algorithm.Based on the standard particle swarm optimization algorithm, a new algorithm using uniform design to determining parameters is presented in this paper. The feasibility and high efficiency of the algorithm are show by the numerical results.
|
PDF Full Text Request