概率dp题，掷骰子

Little Pony and Dice

Time Limit: 2000/1000 MS (Java/Others)    Memory Limit: 32768/32768 K (Java/Others)
Total Submission(s): 40    Accepted Submission(s): 8

Problem Description

  
  

   
   Twilight Sparkle was playing Ludo with her friends Rainbow Dash, Apple Jack and Flutter Shy. But she kept losing. Having returned to the castle, Twilight Sparkle became interested in the dice that were used in the game.
   
   

   
   
The dice has m faces: the first face of the dice contains a dot, the second one contains two dots, and so on, the m-th face contains m dots. Twilight Sparkle is sure that when the dice is tossed, each face appears with same probability. Also she knows that each toss is independent from others.
   
   

   
   
There is n + 1 cells in total, number from 0 to n, Twilight Sparkle race her token from 0 to n according to die rolls. The game end once she move to n or over it. Help Twilight Sparkle calculated the probability she end the game exactly at n.
  
  

 

Input

  
  

   
   Input contains multiple test cases (less than 200). 
   
   
For each test case, only one line contains two integers 
   
   
   
   
    
    
     
     m
    
     and 
   
   
   
   
    
    
     
     n
    
     (
   
   
   
   
    
    
     
     1≤m,n≤109
    
    ).
  
  

 

Output

  
  

   
   For each case, output the corresponding result, rounded to 5 digits after the decimal point.
  
  

 

Sample Input

  
  

   
   3 5
6 10
  
  

 

Sample Output

  
  

   
   0.49794
0.28929
  
  

 

分析：看到n和m的取值范围时，就有点儿蒙。其实，当n变大的时候，相应的误差会变得越来越小，忽略误差，就可以应用动态规划求解。具体：dp[i]表示的是恰好在i处终结的概率，而隐含的一维状态是指走了1~k步的所有情形。那么，dp[i]=dp[i-1]/m+dp[i-2]/m+dp[i-3]/m+ . . .  + dp[i-m]/m，而dp[i-1]=dp[i-2]/m+dp[i-3]/m+ . . . + dp[i-m-1]/m。所以，化简后便有dp[i] = dp[i-1]/m + dp[i-1]。如果当i>m时，便减去dp[i-m-1]*m这一项。