Topics

hongtao
ai
Wednesday, October 3, 2018 10:13 PM
AI学习笔记——动态规划(Dynamic Programming)解决MDP(2)
求解最优MDP实际上就是找到最佳策略(Policy)π来最大化来最大化V函数(Value Function)。 1. 策略估算(Policy Evaluation) 在MDP问题中，如何评估一个策略的好坏呢？那我们就计算这个策略的V函数（值函数），这里我们又要用到之前文章中提到的Bellman Equation了。 这个等式可以通过下一个状态的值函数来求得当前状态的值函数。如果我们对上面这个Bellman
$ 0.000
1
hongtao
ai
Monday, October 1, 2018 8:39 AM
AI学习笔记——动态规划(Dynamic Programming)解决MDP(1)
我们介绍过MDP(Markov Decision Processes马可夫决策过程)以及什么是最优MDP，甚至从强化学习的角度介绍了DQN，Q-learning, Sarsa 等求解最优MDP方法，但是要深入理解强化学习，必须了解背后支持的理论基础。动态规划(Dynamic programming)就是这些算法为什么能够求解最优MDP的理论基础。
$ 0.000
2
newsrx
news
Monday, April 9, 2018 4:53 PM
Real-time implementation of optimal control considering gear shifting and engine starting for parallel hybrid electric vehicle based on dynamic programming
By a News Reporter-Staff News Editor at Journal of Engineering -- Investigators publish new report on Dynamic Programming. According to news reporting originating in Luoyang, People’s Republic of China,
$ 0.000
0
1
justyy
cn
Monday, August 7, 2017 6:11 PM
Software Engineer Interview Question – How to Improve Dynamic Programming Space Complexity? 进一步改进动态规化的空间复杂度
In yesterday's post, we present the final ‘optimized’ solution using Dynamic Programming, which is implemented using a 2 dimensional array with iteration. Let’s review this code:  昨天发了这个帖子讲到动态规化的2种改进,
$ 21.130
37
15
justyy
cn
Sunday, August 6, 2017 10:09 PM
Software Engineer Interview Question - How Many Ways from A to C via B? 软件工程师面试技巧之 从A到B再到C有多少种方法?
Let’s see this question: If @kenchung starts at A, and he can only travels one move at a time to the north or east, and he cannot go back. If he needs to travel through B, and finally reach C, how many
$ 55.954
78
10
justyy
cn
Sunday, August 6, 2017 3:37 PM
Software Engineer Interview Question - Dynamic Programming - Integer Break 软件工程师面试技巧之 动态规化 - 整数拆分
Dynamic Programming (DP)  is one of the very important algorithm in Computer Science. Here is the simplest example of demonstrating the concept of DP if you want to tell you 5-year-old son. Give 5
$ 62.114
45
10
1
justyy
cn
Saturday, August 20, 2016 8:54 PM
错位排列 的 R语言实现
 递归终止条件: F(1) = 0, F(2) = 1. 具体的推导可以看这里.英文: Derangement Permutation Implementation using R Programming    XXX原文::
$ 0.000
6
Top communities