嵌入法地下水管理模型的分段决策技术

以地下水流的支配方程(有限元形式)作为线性规划的部分约束条件构成的地下水管理模型,称之为嵌入法;以各时段内务节点的水头之和最大作为目标函数在各时段内寻优,就是本文探讨的地下水管理模型的分段决策技术。该模型所形成的庞大的约束矩阵,往往需要超大容量计算机。本文利用线性变换消去约束条件中的有限元方程,并将决策变量转化为该时段的抽水量和补给量。将该模型运用于一个实际水源地,其结果令人满意,说明此模型是具有实用意义的。

STEPWISE-DECISION TECHNIQUE FOR EMBEDDING MODEL OF GROUNDWATER MANAGEMENT

By using the FEM ( Finite Element Method ) approximate partial differential equation, the co-operation of the linear programming with the linear equation group constitutes the embedding model of groundwater management, starting with the allocation equation of the groundwater flow. Because the model scale is very large, it is difficult to find a solution with the existing computers. The writer inquires into the stepwise decision technique, which maximizes the sum of hydaulic heads at all nodes in each interval. In this way, optimally pumping and recharging are achieved by the linear programming. Since the constrained matrix of the technique is sizable in common, the linear transformator is employed to define hydraulic heads by the pumpage and the recharge in the linear relationship and to eliminate the finite element equations in the constrains so that the decisive variales become the pumpage and recharge from the hydraulic head, thus reducing the computer memory considerably. Applying the stepwise-decision model to a water-supplying field where the groundwater extraction will be desirable, the optimal policy (the optimal pumpage and recharge in each interval) for the groundwater exploitation will be computed. Averaging the policy according to the time, the stable pumpage from each well and the stable recharge in each spreading basin ( the artificial recharge sinking basin ) will be obtained. By putting the stable amount in the forecasting model of groundwater, the calculating result will be completely identical with the planning one. This shows that the method recommended here is of practical usefulness.