不等式约束下的广义几何规划的一种有效算法

A algorithm for generalized geometric programming with nonequality constraints

利用增广拉格朗日方法、牛顿方法,给出了不等式约束下的广义几何规划的一类有效算法.该算法是对A.GONEN和M.AVR IEL提出的算法的推广,它把不等式约束转化为等式约束,再利用乘子罚函数法得到此算法.最后在适当条件下证明了该算法的收敛性.

An algorithm for generalized geometric programming with nonequality constraints is given by using augmented Lagrangian method and Newton method. The algorithm is the expansion compared to the algorithm which was proposed by A. GONEN and M. AVRIEL. It firstly emphasis to turn the nonequality constraints into the equality constraints, then obtains the algorithm by using penalty function method. Finally, the convergence of the algorithm can be proved under appropriate conditions.