Aiming at six parameters(κ,γ,δ,μ,η and C_0) of the dynamic vibration absorber in deep sea,the optimal design of parameters was processed with Broyden-Flether-Goldfarb-Shanno(BFGS) method of variable metric method,penalty function method and the line-search method.

In this paper, we combine the extended projection type method given in 1 and uncontrained variable metric methodaand and give a new algorithm with n+m +1 parameters to solve linearly constrained nonlinear programming problems.

This paper prestents a class of reduced variable metric algorithm for programming problems with standard linear constraints with the help of reduced gradient and Huang s variable metric methods.

Adaptive method and variable scale method are introduced in this paper.

Variable scale method,a sort of optimal method,is used in design of optical system,meanwhile,the technique of using the ‘escape function’ for variable scale method in design is describle.

According to the problem of artillery control system and simulated annealing algorithm,a new improved simulated annealing algorithm,a simulated annealing algorithm based DFP is designed,and it is applied to PID control of artillery control system.

In the central level, the modified DFP method is adopted to calculate the optimal control of reactive power for each controllable node whose voltage is kept within required constraints, the active power loss at its minimum, and the results are transferred to the node level via SCADA/EMS system.

In accordance with the Duncan Chang nonlinear elastic soil model,that,the DFP method is used to draw up optimization programme for parameteric identification of the parameters in this model,it is verified combined with a calculation example of the practical engineering project,and its feed back analysis is carried out too.