算法的

In this paper a genetic algorithm four-dimension data assimilation system was applied to the more complex MM5 model, we verify the performance of the genetic algorithm assimilation system, we take the a heavy rainfall process during 04-05, July, 2003 as the experimental research object, and carry out data assimilation experiment of actual observation data, compare the experimental results of the genetic algorithm assimilation system and MM5 adjoint model assimilation system with the experimental results without assimilation, the following is the result:genetic algorithm assimilation system is used not only in a barotropic primitive equation but also in the complex model, MM5, for example, in the ideal experiment the result of the genetic algorithm assimilation system is better than the MM5 adjoint model assimilation system, in the assimilation experiment of actual observation data, whichever assimilation system is used, the initial field of numerical prediction model is effectively improved and the prediction effect of physical field and rainfall are somewhat enhanced.

本文将遗传算法的四维同化系统运用到更复杂的MM5模式中，在理想实验中验证遗传算法同化系统的性能，并以2003年7月4日"南京暴雨"作为研究对象，进行了实际观测资料的同化，并将遗传算法同化系统和MM5伴随模式同化系统得到的结果跟未同化的结果进行了比较，结果如下：遗传算法同化系统不仅可以用在正压原始方程模式上，同样也可以用在MM5等复杂模式上，理想实验证明了遗传算法同化系统优于MM5伴随同化系统，在实际观测资料同化中，采用两种方法的同化系统都能有效地改善数值预报模式的初始场，能在一定程度上提高物理量场和降水量的预报效果。

AGA is a kind of bionic optimization algorithm based on Genetic Algorithm.

该算法是基于遗传算法的一种仿生优化算法。

And it comparatives and analyses the efficiency of three algorithm's cliping through an example, the example proved: Liang－Barsky algorithm efficiency is highest, Cohen-Sutherland next best, and the midpoint of division algorithm efficiency is lowest.

并通过一个实例对三种算法的裁剪效率比较分析，实例证明：梁友栋一Barsky算法效率最高，Cohen—Sutherland次之，而中点分割算法裁剪效率最低。

The same result can also be achieved when optimizing F8 function (Rastrigin's function). Comparing to EGA, KEGA is easy to realize and the increase of computation load can be neglected, but the performance of KEGA can be improved greatly. 2. To enhance the speed of AIA, two new methods of computing the concentration of antibody were developed. Simulation results for F8 and F15 (needle in haystack: type I) functions show that the two methods can improve the running speed of AIA more than 8. 5 times. 3. As to the drawback in defining antibody concentration, a new definition based on Euclidean distance between two antibodies and fitness is proposed. On this basis, a new AIA was constructed. The convergence performance are more better than that of AIA. 4. Combining king crossover with AAIA and DBAIA, two new AIAs, KAAIA and DKBAIA, were proposed.

仿真实验结果表明KEGA实现简单，与EGA相比，其增加的计算开销可以忽略，但性能改善十分明显； 2、针对基于信息熵的免疫算法运行速度慢的缺点，提出了二种能够加快AIA运行速度的新的抗体浓度计算方法，得到两种新的免疫算法，即加速的人工免疫算法(Accelerated Artificial Immune Algorithm，AAIA)和加速的人工免疫算法-2（Accelerated Artificial Immune Algorithm-2，AAIA-2），对测试函数F15(needle in haystack：type I)、F8函数的仿真研究表明，这两种算法的运行速度都是AIA的8.5倍以上； 3、针对AIA的浓度定义存在的缺陷，提出了一种新的基于抗体间欧氏距离和适应度的新的抗体浓度定义，并依据该定义构造了一种基于欧氏距离的人工免疫算法(Euclidean Distance-based Artificial Immune Algorithm，DBAIA)。

Firstly,the author analyzes principles and general methods of lustering- based Data mining.

本文作者首先认真研究分析了基于聚类的数据挖掘的基本原理，提出了一个将单亲遗传算法与Kohonen聚类算法相结合的混合聚类算法，克服了Kohonen 算法的局部最优问题以及采用普通遗传算法聚类时的搜索速度和聚类精度的矛盾。

Dijkstra algorithm and matrix iterative algorithm are important algorithms of the shortest route problem.In order to find which of the two algorithms is more efficient,the two algorithms were discussed,and their program flows were introduced.

Dijkstra算法和矩阵迭代算法都是最短路径问题的经典算法，为了研究这2种算法的计算效率，介绍了2种算法的基本思路和程序流程；通过对时间复杂度、结果和时间的关系指标进行多次计算，表现2者之间的效率差异。

In this thesis, the theories and methods of GA are analyzed in detail , the trend of GA in recent times is briefly discussed. Because of the importance of image segmentation, two methods of color image segmentation based on GA are proposed. One is the surface of thresholds image segment method, the other is entropic threshold method. Two algorithms are realized through VC++ 6.0. After that, we compare the characteristics of two algorithms and prove that GA can be very well applied to solving image optimal problems.

本篇论文比较详细地论述了遗传算法的基本理论、方法，指出了算法的研究现状，由于图象分割的重要性，提出了两种基于遗传算法的彩色图象分割方法，即阈值曲面分割方法和最佳熵阈值分割方法，通过VC++6.0编程得到了实验验证，比较了两种算法的性能，证明遗传算法完全适用于图象的优化问题。

Constraint-based MiniCon algorithm is proposed. This algorithm adds a step of view selection. As a result, the problems of missing query rewritings and generating redundant query rewritings in traditional MiniCon algorithm in presence of domain constraints are solved, and the accuracy and completeness of the algorithm is ensured.

该算法在传统的MiniCon算法中增加了一步视图选择，解决了传统的MiniCon算法中丢失查询重写和生成冗余查询重写的问题，保证了算法的正确性和完备性，并给出了相应的算法分析和实例分析。

The double decomposition polygon clipping method is brought forward based on polygon trapezoidal decomposition. By analyzing its time complexity, the case it applicable to and its accuracy, it is pointed out that the new clipping method has a lower complexity than the old methods and does not make mistakes in the case SutherLand-Hodgman algorithm would.

在多边形梯形剖分的基础上提出了双剖分多边形裁剪算法，并对算法的时间复杂度、适用范围和处理的正确程度进行分析，指出该算法的时间复杂度低于以往算法，在处理非自交叉的多边形时结果正确，不存在类似SutherLand-Hodgman等算法处理出错的问题。

Simulated anneal and Tabu are extensively applied in the field, in which, Genetic Algorithms is most general one. In the former study of the optimizing for motor, people often focused on cross operator and variation operator, which was based on the G.A.

在以往的以遗传算法为优化设计主要算法的研究中，偏向于对交叉算子与变异算子进行改进以提高遗传算法的效率，即对遗传算法的本身进行改进，从而产生了很多种类的混合遗传算法。

The split between the two groups can hardly be papered over.

这两个团体间的分歧难以掩饰。

This approach not only encourages a greater number of responses, but minimizes the likelihood of stale groupthink.

这种做法不仅鼓励了更多的反应，而且减少跟风的可能性。