【摘要】 提出一种由信赖域方法和梯度法相结合的求解Fermat场址问题的新算法.该算法在迭代中先采用信赖域法,当出现内循环时,则改用不做线搜索的梯度法.算法运算简单,具有全局收敛性,并克服了信赖域算法产生内循环时造成的运算量大和梯度法收敛速度慢的缺陷.更多还原

【Abstract】 Fermat location is a class of special non-differentiable problem. A combined algorithm based on trust region method and gradient method is proposed. The algorithm tries the TR technique first and the gradient method with nonlinear search when the trial step is not successful. The global convergence of the algorithm is proved under a weaker condition. The method saves a considerable computations.更多还原