Improved Dynamic Programming for Reservoir Operation Optimization with a Concave Objective Function
Improved Dynamic Programming for Reservoir Operation Optimization with a Concave Objective Function
复制标题
DOI：
10.1061/(asce)wr.1943-5452.0000205
复制
发表时间：
2012-11
影响因子：
3.1
通讯作者：
T. Zhao;Ximing Cai;X. Lei;Hao Wang
中科院分区：
文献类型：
--
作者：
T. Zhao;Ximing Cai;X. Lei;Hao Wang
边际效用递减是水资源系统的一个重要特征。假设边际效用递减（即，在水库效用函数的基础上，推导了确定性和随机性条件下水库蓄水与最优放水决策之间的单调关系，并提出了一种提高凹目标函数下水库调度的确定性动态规划（DP）和随机动态规划（SDP）计算效率的算法。实际算例表明，改进的DP和SDP比传统的DP和SDP具有更高的计算效率。改进的DP和SDP的计算复杂度为O（n）（n阶，状态离散化的数目），而传统DP和SDP的计算复杂度为O（n2）。
AbstractDiminishing marginal utility is an important characteristic of water resources systems. With the assumption of diminishing marginal utility (i.e., concavity) of reservoir utility functions, this paper derives a monotonic relationship between reservoir storage and optimal release decision under both deterministic and stochastic conditions, and proposes an algorithm to improve the computational efficiency of both deterministic dynamic programming (DP) and stochastic dynamic programming (SDP) for reservoir operation with concave objective functions. The results from a real-world case study show that the improved DP and SDP exhibit higher computational efficiency than conventional DP and SDP. The computation complexity of the improved DP and SDP is O(n) (order of n, the number of state discretization) compared to O(n2) with conventional DP and SDP.