中文摘要：

现代电力系统恢复控制中的机组优化启动问题是一个复杂的多阶段决策优化问题。从全过程最优化的角度出发进行系统建模，以恢复过程中确定时间段内系统发电功率对时间的积分最大为优化目标，同时计及各种约束条件。通过将恢复过程按时间离散化，将此连续的全过程最优化问题分解为每一时步的优化计算，每一时步的优化问题构成一个典型的“背包”问题。采用经典回溯算法逐步求解每一时步中的“背包”问题，确定每一时步启动的机组，从而形成整个恢复过程的机组优化启动策略。求解方法较好地权衡了对所求解问题的快速性和正确性要求。以改进的15机系统为例进行测试，结果表明利用所提方法确定的机组恢复顺序可以获得比传统方法更好的恢复效果。

英文摘要：

The optimal unit start-up strategy of modern power system restoration is a complex, multi-stage decision optimization problem. An optimal unit start-up strategy model is proposed for global optimization, which maximizes the total active power output within a certain period of system restoration under various constraints. The restoration process is divided into a series of time steps, each of which is optimized as a knapsack problem. The classic backtracking search method is employed to solve these knapsack problems and the units to be started in each time step are decided, the overall optimal start-up strategy of power system restoration is thus formed. The solution balances the precision and speed well. The case study of a modified 15-generator system shows that the determined unit start-up sequence by the proposed model is better than that by the traditional methods.