摘要
对TACO算法进行了改进,引入最大最小蚁群算法,并提出正实数编码方法和添加常数项的信息素更新技术,以增大算法搜索范围,简化搜索过程,降低挥发系数、信息素上下限等参数和优化函数值对算法的影响程度,改善算法的性能,提高算法的普遍适应性,并通过3个函数的优化求解,证明了其有效性。对于目标函数为隐式的复杂结构优化问题,提出将改进的蚁群算法与有限元方法相结合的方法,发展用于航空发动机涡轮盘的结构形状优化分析,结果表明所提出的方法是成功的。
An improved ant colony algorithm is developed based on TACO. MAX-MIN ant system is introduced. A new method of positive-real-number coding is put forward for enlarging search range and simplifying search process. In order to reduce the influence of the evaporation coefficients, the most and least value of pheromone and the value of optimized function, a new technology of constant-pheromone update is put forward. Thus the algorithmic performance and universality are improved. Through optimizing the solutions of three fanctions, their validity is proved. A new method of combining the improved ant colony algorithm and FEM is presented for the complicated structure whose objective function can＇t be showed by formula directly. The method is applied to the shape optimization of turbine disk. The results show that the method of combining the improved ant colony algorithm and FEM is successful.
出处
《航空学报》
EI
CAS
CSCD
北大核心
2007年第5期1110-1115,共6页
Acta Aeronautica et Astronautica Sinica
