A Monge array lets us find row minima faster than scanning every entry.
Monge数组能让我们比逐个扫描元素更快地找到每一行的最小值。
By exploiting the Monge array property, the dynamic programming transition can be optimized from quadratic time to near-linear time in some cases.
利用Monge数组的性质,在某些情况下可以把动态规划的转移从二次时间优化到接近线性时间。
SMAWK: A Parallel Algorithm for Finding Row Minima of Totally Monotone Matrices(Aggarwal, Klawe, Moran, Shor, Wilber,1987)——经典论文,常与Monge数组/全单调矩阵的行最小值问题并列讨论。
The Shortest Path Problem with Monge Properties and the Transportation Problem(相关运筹与优化文献中常见主题)——Monge结构在运输问题、最短路变体等中被系统使用。