粒子群混合算法在变导向器面积计算中的应用

求解非线性方程组经典方法具有严格的局部收敛性;粒子群等进化算法解决了全局收敛的问题,但计算效率偏低,存在最优解不稳定的问题。结合经典Newton-Raphson法的超线性收敛速度和粒子群算法全局收敛能力的粒子群混合算法具备2类算法的优点。在迭代初期采用粒子群算法获得的近似全局解作为Newton-Raphson算法的初始值,以确定高精度的解。利用粒子群混合算法在发动机变导向器面积的大偏离计算中获得了较好的收敛效果,解决了常规Newton-Raphson法不收敛的问题。

Application of Hybrid Particle Swarm Algorithm in Variable Guide Vane Area Computation

The classical method for solving nonlinear equation group has a strict local convergence limitation. The whole convergenceproblem was solved by particle swarm algorithm, which is low computation efficiency and optimum solution unsteady. The hybrid particle swarm algorithm contains high liner convergence speed of Newton-Raphson method and whole convergence ability of particle algorithm. The whole solutions obtained by particle algorithm in the iterative initial period were concerned the high accuracy solution. The better convergence results were obtained by hybrid particle swarm algorithm in the high deviation calculation of engine variable guide vane area,which solve the non-convergence problems in normal Newton-Raphson method.