一类大系统的稳态递阶优化控制算法

充分利用线性稳态大系统这一特点，采用一种新方法在不破坏可分性的前提下对求解问题进行凸化．这种两级算法是：上级（协调器）固定协调变量；下级各局部决策单元并行求解Ｎ个子问题，再把结果送回协调器．这样，通过上、下级不断的信息交换就可得到原问题的最优解．由于该方法消去了子系统等式约束相应的Ｌａｇｒａｎｇｅ乘子，故没有扩大求解空间，提高了优化效率．仿真结果表明算法是有效的．

Hierarchical Optimization Control for a Class of Large Scale Steady State Systems

Making full use of the distinguishing feature of the linear constraints of large scale steady state systems with linear constraints,the anthors put forward a new method of hierarchical optimization control.The two level algorithm is that,the upper level (coordinator) fixes coordination variables,the all units of the lower level solve parallel N subproblems respectively,then the solutions of the subproblems are sent back to coordinator.By the continuous information exchange between the upper level and the lower level,an optimal solution can be finally obtained.Because the algorithm eliminates the Lagrange multipliers corresponding to the equality constraints of subsystems,the solving space is not extended,and the efficiency of optimization is increased.Simulation result shows that the algorithm is effective.