基本思想与分治法类似，也是将待求解的问题分解为若干个子问题（阶段），按顺序求解子阶段，前一子问题的解，为后一子问题的求解提供了有用的信息。在求解任一子问题时，列出各种可能的局部解，通过决策保留那些有可能达到最优的局部解，丢弃其他局部解。依次解决各子问题，最后一个子问题就是初始问题的解。



由于动态规划解决的问题多数有重叠子问题这个特点，为减少重复计算，对每一个子问题只解一次，将其不同阶段的不同状态保存在一个二维数组中。

-The basic idea with the divide and conquer method is similar to the problem is to be solved is decomposed into several sub-questions (stages), in order to solve sub-stages, before a sub-problem solutions for problem solving after a child has provided useful information. In any sub-problem solving, list all possible partial solutions, through making reservations that are likely to achieve optimal local solution, discarding other local solution. Order to solve each sub-problem, and finally a sub-problem is the initial solution to the problem. Dynamic programming to solve the problem because the majority of sub-problems that have overlapping characteristics, in order to reduce double counting, for each sub-problem solution only once, at different stages of its different states stored in a two-dimensional array.