Scribd
Upload
230 views

Dynamic Programming

Dynamic programming (DP) breaks problems down into overlapping subproblems and solves each subproblem only once, storing the results in a table for future reference. The document provides examples of DP problems like finding the sum of numbers, calculating the nth Fibonacci number, and solving problems related to coin change, palindromes, and rock climbing. It emphasizes practicing DP problems is key to mastering the technique and being able to apply it to new problems.

Uploaded by

Talha Habib
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
230 views

Dynamic Programming

Dynamic programming (DP) breaks problems down into overlapping subproblems and solves each subproblem only once, storing the results in a table for future reference. The document provides examples of DP problems like finding the sum of numbers, calculating the nth Fibonacci number, and solving problems related to coin change, palindromes, and rock climbing. It emphasizes practicing DP problems is key to mastering the technique and being able to apply it to new problems.

Uploaded by

Talha Habib
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
You are on page 1/ 5

Dynamic Programming DP is a very important part of programming .

The basic theory of dp is pretty simple, but to apply the knowledge, we require a lot of practices, practicing == mastering, its extremely true for dp . And the fun is that too many problems can be solved with dp, its one kind of optimized brute force, which can solve many problems surprisingly fast . To learn dp, we start from the very beginning [ its not for advanced reader, so Im not differentiating between dp & Memoization, there is no problem to call both dp ] Given n numbers, you have to print their sum How would you do that ? scanf(%d,&n) ; int sum = 0 ; for(i=0;i<n;i++) { scanf(%d,&in[i]) ; sum = sum + in[i] ; } Simpe, isnt it ? you can say youve solved a dp problem :D

How ? dp means to solve a problem, break it into subproblems, solve them, & get your answer from them Let call dp[i] = sum of numbers indexed from 0 to I ; So, youll have to find dp[n-1], ok? Now, lets break it into subproblems, cant we say dp[i] = dp[i-1] + in[i] ; . ? so , we can find the solution from the recursive call bool visited[MAX] = {0} ; int cal( int i ) { If(i<0) If(visited[i])

return 0 ; return dp[i] ;

Visited[i] = 1 ; return dp[i] = in[i] + cal(i-1) ; } So, if we say, youll be given 100 numbers, & then 10000000 queries, each giving i as input, & you are to print in[0] + in[1] + .. + in[i] .. you can do this dp [ advanced coders call it memoization, but you can call dp ] If you dont do this memoization, youll get tle [ why?i think you know ]

.. so to solve dp , we need to do things 1. 2. 3. define state [ here i is state ] derive the recursive solution [ here dp[i] = in[i] + dp[i1] ] find base case

& its over :D What is state ? its the position where the problem is strictly defined and we can give a solution To define the state, be carefull, make sure itll end up to the base case, it must not be cyclic . Int cal(int i) { If(i<0 || i==n) return 0 ; If( isV[i] ) return dp[i] ;

dp[i] = in[i] + cal(i+1) + cal(i-1) ; isV[i] = 1 ; Return dp[i] ; } . What do you think about this dp ??? :D:P

Now . Find nth Fibonacci number modulo 1007 [ n<=10000 ] I know you can solve it with dp[ or memoization [ hint: dp[i] = dp[i-1] + dp[i-2], & you need two base cases as there are two terms in the recursive call ] ] and you know, (a+b)%M = (a%M + b%M ) % M Now solve it given n & r, find nCr[you know what it is] modulo 1007 n<=1000, r<=1000, there will be 1000000 test cases Its very basic idea of dp to start learning + practicing Types of DP : There are many types, you can master them with practising , & can apply the same idea in many problems You can learn some standard dp problem solutions like knapsack, coin change, matrix chain multiplication, rock climbing, tsp etc etc but I think you should think about the problem & make out a solution yourself, & its not tough if you know the raw idea of dp Solve the following problems

A. Memoization
1. http://uva.onlinejudge.org/external/107/10721.html 2. http://uva.onlinejudge.org/external/110/11069.html [a graph problem] http://uva.onlinejudge.org/external/109/10918.html [Tri Tiling] Probability/Expected Value
1. http://uva.onlinejudge.org/external/117/11762.html [Race to 1 ] 2. http://uva.onlinejudge.org/external/111/11181.html[Probability|

Given]s Simple dp[minimize/maximize] similar to lcs/edit distance

http://uva.onlinejudge.org/external/115/11584.html Palindromes ] http://uva.onlinejudge.org/external/111/11151.html [ Longest Palindrome ]

[Partitioning

by

Coin Change: http://uva.onlinejudge.org/external/1/147.html [ Dollars ] http://uva.onlinejudge.org/external/111/11137.html [Ingenuous Cubrency]

Rock Climbing : http://uva.onlinejudge.org/external/109/10913.html [Walking on a Grid]

You might also like