非线性全局优化问题的进化算法研究.docVIP

硕 士 学 位 论 文 论文题目： 专业名称： 单位年级： 完成日期： 渤海大学研究生学院非线性全局优化问题的进化算法研究 [摘 要] 全局优化问题在许多领域中都有广泛的应用和发展。特别是带有约束条件的全局优化问题。本文对求解全局优化问题的常用方法以及发展状况进行了阐述，对带有约束的全局优化问题进行研究，提出了一种改进的进化规划算法和遗传算法在实际中的应用。本文主要内容如下： 1．在进化规划算法中其突变过程中，我们可以把可行域内的点看成是不同大小的体。根据函数值定义其，其越大函数值越小，然后基于聚类思想，提出一种改进的进化规划算法。通过计算大小，来增加新的搜索方向。从而沿着函数值可能下降的方向进行再以一定的概率P接收这种方法，以概率1—P接收高斯变异算子的搜索方法，来保证算法的随机性和全局性最后几个典型的数值例子，来说明算法的可行性和有效性。 2．算法应用非常广泛，选择适当的算法可以节约社会资源，大大提高工作效率，可以避免人为的失误和一些主观因素的影响，保证其科学性、随机性、客观性。预测选举问题实际上可以看成一个多约束条件的优化问题，有很多求解的办法这里将用遗传算法来搜索满足条件的可行解，其准确性大大提高。 [关 键 词]： 全局优化；进化规划；随机搜索；非线性规划；变异算子 THE STUDY OF EVOLUTIONARY ALGORITHM ON NONLINEAR GLOBAL OPTIMIZATION ABSTRACT The issue of global optimization has been applied and developed widely in various fields, especially the one that is inhibited. This paper focus on researching such issue, elaborating its development and commonly used methods. On this basis, the author puts forward a newly-improved evolutionary and genetic programming method to put into practical use. There are two main parts as follows : 1 . In the mutational process of genetic programming method, we can consider the points of feasible region as different volume individuals. Usually, we define the volume according to its function value, then based on the theory of clustering. we put forward an improved genetic programming method. Considering the individual in feasible region as astrals of different volumes, Can the feasible region of different size of individual as defined by the astral body, increase inflation E new search directions, thus may decline along the function value direction in a certain probability method, in order to receive the P probability 1 - P receiving gaussian mutation operator search method, to ensure that the algorithm convergence and overall, finally several typical application of numerical example to show the feasibility and effectiveness of the algorithm. 2. Optimizing algorithm is used widely in our daily life, and if we choose the prope