基于混合遗传算法求解飞机定常状态  

Solution of steady state of aircraft based on mixture genetic algorithm

作　　者：栾志博[1] 郑淑涛[1] 李洪人[1] 

机构地区：[1]哈尔滨工业大学机电工程学院,哈尔滨150001

出　　处：《吉林大学学报（工学版）》2011年第1期165-170,共6页Journal of Jilin University:Engineering and Technology Edition

基　　金：教育部新世纪优秀人才支持计划项目(NCET-04-0325)

摘　　要：为保证对飞机定常状态准确求解,在分析了定常状态的基本特性和相应约束条件的基础上,提出了一种新的混合遗传算法。算法基于'学习潜能'的概念将遗传算法中的拉马克学习与鲍德温学习两种学习机制有机结合在一起,将局部搜索次数在群体中进行合理分配,使学习的优势得到充分发挥,使其不足得到有效抑制。本文算法不依赖于飞机动力学模型的具体形式,只利用状态参数进行计算,可以求解任何预先指定的定常状态。仿真结果表明了本文算法的有效性。To obtain the accurate solutions of the steady state of aircraft, a hybrid genetic algorithm was proposed according to the analysis of the basic characteristics and constraints of the steady state. Based on the concept of ＂individual learning potentiality＂, this algorithm logistically integrates the Lamarckian learning and Baldwinian learning mechanisms. It rationally distributes the number of local search among the population to make the advantage of the learning into full play, meanwhile inhibit its disadvantage. The algorithm can solve any preassigned steady state on the basis of the state variables rather than the real form of the aerodynamic model. Simulation results validate the proposed algorithm that combines the two learning mechanisms.

关 键 词：人工智能 遗传算法 定常飞行状态 个体学习潜能 拉马克学习 鲍德温学习 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]