毕业论文遗传算法研究.doc

　毕业设计（论文） 题目 遗传算法研究 摘　要 遗传算法（GA）是一类借鉴生物界自然选择和自然遗传机制的随机化搜索算法。作为一种有效的全局优化搜索算法，它具有简单、通用、鲁棒性（Robust）强和适于并行分布处理的特点。基本遗传算法提供了一种通用的算法框架，方便根据具体的问题提供改进策略或者和其它算法混合使用，因此具有广泛的应用潜力。本文中介绍的改良遗传算法改进了基本遗传算法的选择和变异操作，选择操作使用了最优保存策略，保证了当前产生最优个体不会因为交叉、变异而丢失，目的是提高算法的收敛速度；变异操作采用交换、倒序、插入三种变异算子混合的方式，增加种群的多样性，防止算法过早的陷入局部最优解而出现“早熟”现象，目的是提高解的质量。为了验证算法改进策略的有效性，以求解TSP为例，针对不同规模的TSP（eil51、eil76、eil101）做了大量的数据统计。通过对比交换、倒序、插入变异算子以及三种算子混合时解的质量来说明使用多种变异算子混合的优势。同时通过求解相同的问题（att48）与其它智能优化算法（基本遗传算法、模拟退火算法、基本蚁群算法）做了对比，验证算法改进策略的有效性。 关键词： 改良遗传算法 旅行商问题 最优保存策略 混合变异算子  Abstract The genetic algorithm (GA) is a random search algorithm learn from biological natural selection and natural genetic mechanisms. As a kind of effective global parallel optimization search algorithm, It has a simple, universal, robustness strong and suitable for the parallel distributed processing characteristics. The basic genetic algorithm provides a general algorithm framework to facilitate the improvement strategies or mixed with other algorithms, depending on the problem, and therefore has broad application potential. Improved genetic algorithm described in this article to improve the basic genetic algorithm selection and mutation operations, select the operation using the optimal preservation strategy to ensure that the current to produce the best individual will not be lost because of the crossover, and mutation, the purpose is to improve the convergence speed ; mutation operation using exchange, reverse, insert three variation operator mixed methods, increasing the diversity of the population to prevent the algorithm into a local optimal solution to the premature to improve the quality of the solution, in order to verify the effectiveness of the strategy to solve the TSP, for example, for different sizes of TSP (eil51, eil76, eil101) have done a lot of statisti. By contrast, exchange, reverse, insert the quality of the solution when the mutation operator and the mixing of the three operators to use a variety of mutation op