论文部分内容阅读
首先介绍了遗传算法和模拟退火算法等全局优化算法,并针对遗传算法的早熟现象和容易陷入局部最优的缺点,将模拟退火算法引入到遗传算法中,提出了遗传模拟退火矢量量化码书设计(GSAKVQ)算法.此外,针对基于划分的染色体编码方式的特点,算法提出了新的有效的交叉算子和变异算子.同时,将算法从输入空间映射到特征空间,提出了相应的遗传模拟退火核矢量量化算法,改善了算法在某些数据集上的不足.最后,通过实验表明,GSAKVQ算法,在大部分的数据集上都能取得较好的结果,从而验证了算法在数据聚类问题上的有效性.
Firstly, the global optimization algorithms such as genetic algorithm and simulated annealing algorithm are introduced. Aiming at the prematurity of genetic algorithm and the disadvantage of being easily trapped in local optimum, the simulated annealing algorithm is introduced into genetic algorithm, and the genetic simulated annealing vector quantization codebook design (GSAKVQ) algorithm.In addition, a new effective crossover operator and mutation operator are proposed for the algorithm based on the division of the chromosome coding method.At the same time, the algorithm is mapped from input space to feature space and the corresponding genetic simulation Annealing vector quantization algorithm to improve the algorithm in some datasets deficiencies.Finally, experiments show that, GSAKVQ algorithm in most data sets can get better results, which verifies the algorithm in data clustering Validity on the issue.