摘要
针对遗传算法易早熟和收敛速度慢的不足提出自适应遗传算法,引入自适应变异算子自动调整自变量步长v使收敛加速,以个体适应值计算交换率和交叉点位置增强了算法的智能性,同时在进化过程中加入特殊的个体保证基因的完备性。水轮机叶片型面复杂,因此铸件铸造成形后难以设置测量基准点和坐标系,需要通过测量数据与设计曲面的坐标变换实现最佳匹配。将自适应遗传算法引入最佳匹配问题的求解,算例分析表明该方法与标准遗传算法相比具有运算速度快和稳定性好等特点。
A new algorithm named self adapting genetic algorithm is presented to avoid the bugs such as the premature problem and the low convergence speed of the standard genetic algorithm. The self adapting mutation operator automatically adjusts the step size of variable so that the convergence speed is improved. Crossover probability and location is calculated by units applicability which enhanced the intelligence of the algorithm. Two special units U0 and U1 ensure the integrality of gene. Duo to the complexity of blade surface of water turbine, the reference mark and coordinate is difficult to establish on the blade roughcast. According to this, the coordinates of the measured data and design surface need transforming to achieve the optimal matching. The self adapting genetic algorithm is applied to solve the optimal matching question and the results suggest that this algorithm has higher operation speed and stability compared with the standard genetic algorithm.
出处
《铸造技术》
EI
CAS
北大核心
2006年第2期101-104,共4页
Foundry Technology
基金
国家"863"高技术研究发展计划燃气轮机重大专项资助项目(2002AA503020)
国家自然科学基金资助项目(59879023)
关键词
叶片铸件
最佳匹配
自适应遗传算法
余量
Blade roughcast
Optimal matching
Self adapting genetic algorithm
Tolerance
