| 1. | The iterative process may be continued until some criterion for convergence is met .
这种迭代步骤可继续下去，直到符合某种收敛要求为止。 |
| 2. | Repeat steps 6 and 7 for the element
，请迭代步骤6和步骤7 。 |
| 3. | Moreover , we show that after finite iterations , the unit step is always accepted
此外，我们还证明：经一定的迭代步后，单位步长总可以取到。 |
| 4. | The step size for coefficient update in the mf is controlled by the estimated snr
通过sf估计输入信号信噪比( snr )来控制mf系数更新的迭代步长。 |
| 5. | A solution that minimizes the model function within the trust region is solved as a trial step
信赖域方法并不要求目标函数的二次模型凸性，在信赖域内求得迭代步使得二次模型最小。 |
| 6. | The iteration formula and determination of iteration step are set up according to a priori information of rock and concrete
建立了考虑先验信息的修正高斯-牛顿算法的迭代格式和迭代步长的确定方法。 |
| 7. | Operand of map - drift is large , so 2 ways of reducing the operand are put forward , which are increasing the iterative step and decreasing the numbers of range cells
由于md算法的运算量很大,采用了增加迭代步长、减少距离单元两种减少运算量的方法。 |
| 8. | Furthermore , in the last part of this paper , based on the simulation results , some problems of artificial neural network algorithm are addressed
收敛性是神经网络校正算法的关键问题，本文最后对此进行了理论探讨并根据仿真结果对其迭代步长进行选择。 |
| 9. | Based on double dogleg path , the iterative direction is always obtained in the intersection of double dogleg path and bound of trust region by solving the affine scaling trust region subproblem
一般地，基于双折线路径方法可以在双折线路径和信赖域边界相交点得到迭代步。 |
| 10. | Moreover , we prove that after finite iterations , the unit step is always accepted and the proposed method essentially reduces to the modified bgs method with the unit steplength
此外，还证明了经一定的迭代步后，单位步长总可以取到。因此算法在解的局部还原为用单位步长的bfgs算法。 |
