Dynamic Programming

Dynamic Programming
From An
Excel Perspective
Dynamic Programming From An Excel Perspective
Dynamic Programming
From An Excel Perspective
Ranette Halverson, Richard Simpson

Catherine Stringfellow
Department of Computer Science

Midwestern State University
Dynamic Programming
 A popular method for solving problems by breaking

them down into overlapping sub-problems that display
optimal substructure
Can be thought of as a top-down approach utilizing a
bottom-up evaluation
 Normally used to solve optimization problems
Excel
Generally taught in the freshman application classes

 Seldom taught to computer science majors
 In reality CS majors need to be able to use spreadsheets
 So what do we do?
Solution
 Do not need to teach the spreadsheet AT ALL
 Include spreadsheet usage in a few of their projects
and/or homework
 Spreadsheet usage includes
Graphing data collected via empirical analysis of two
algorithms.
 Rapidly construct mathematical tables for applications
 Simulating wave-front parallel algorithms
 Evaluating dynamic programming tables (the point of this
talk)
A Very Simple Example (used in Computer Science for

Science Majors)
The memo-ization of the recursive Fibonacci function.
Remember the complexity of the following?
int Fib( int n)
{
if (n<3) return 1
else return ( Fib(n-1)+Fib(n-2) );
}
Two well-known O(n) solutions
int FibMemo(int n,int * A){

int Fib( int n) if (A[n]!=0) return A[n];
{ else {
A=new int[n+1]; A[n]= FibMemo(n-1,A) + FibMemo(n-2,A);
A[1]=A[2]=1; return A[n];
for(int i=3 ; i<=n ; i++) }
A[i] = A[i-1] + A[i-2]; };
return A[n]; int Fib(int n)
{
}
int * A = new int[n+1] ;
// Pure Bottom up calculation using
for (int i=1;i<n+1;i++){ A[i]=0;}
// an array. The non array version is
A[1]=A[2]=1;
// not relative to our discussion.
return FibMemo(n,A);
} // A recursive Memoized version
Excel’s simple approach
=A1+B1 Kand copy cell to the right

Pascal's triangle is constructed

in Excel in the bottom-up
approach.
The programmed solution can

be handled via DP as in the
Fibonacci example, either using
an array with or without
memoized recursion.
formula =B1+A2 is
copied from B2 to The pure recursive version is
the remaining cells. computationally unacceptable.
There are many DP algorithms that appear throughout our

curriculum.
 Longest Common Subsequence Bioinformatics class.
 Sequence Alignment: Bioinformatics
 Optimal Binary Search Tree: Algorithm Analysis
 Matrix Chain Multiplication: Algorithms Analysis
 Graph Algorithms
Longest Common Subsequence (LCS)

Definition: find the longest subsequence that is common
to two (or more) sequences.
Example
Seq1 = B D C A B A Seq2 = A B C B D A B
LCS = BCBA
Note: The LCS is not a substring!

Longest Common Subsequence (LCS)

DP leads to the following recursive approach.
Let z=z1 z2 … xk be the LCS of x1 x2 … xi-1 xi

y1 y2 … yj-1 yj
0 𝑖𝑓 𝑥 = 0 𝑜𝑟 𝑗 = 0
𝑐 𝑖, 𝑗 = ൞ 𝑐 𝑖 − 1, 𝑗 − 1 + 1 𝑖𝑓 𝑖, 𝑗 > 0 𝑎𝑛𝑑 𝑥𝑖 = 𝑦𝑗
max(𝑐 𝑖, 𝑗 − 1 , 𝑐 𝑖 − 1, 𝑗 𝑖𝑓 𝑖, 𝑗 > 0 𝑎𝑛𝑑 𝑥𝑖 ≠ 𝑦𝑗
Where c[ib ,j] is the length of the LCS of x1..xi and y1..yj
The initialized LCS Table
Copy the following cell

formula to all grey cells.
These represent the c(i,j)’s
IF(D$2=$B4,C3+1,MAX(D3,C4))
Cell formula
And the solution is
Note diagonal increments
Length of LCS
DP Problems with complex table manipulation

 Optimal Binary Search Tree (in paper)
 Matrix Chain Multiplication
Question: What do you do with problems that require the

processing of rows or columns in such a way that the usual
cell function approach is not adequate?
Excel does not allow cell function side effects! Hmm.

Write the code in the include macro language (VB?)
Summarizing
 CS students can benefit from work with Excel
 Excel can support many projects in the CS curriculum.
 Table processing available in Excel supports some
algorithm visualization quite well.
 This approach works particularly well with the simpler
DP problems.
THE END

Dynamic Programming

Uploaded by

Dynamic Programming

Uploaded by

Dynamic Programming

Ranette Halverson, Richard Simpson

Department of Computer Science

 A popular method for solving problems by breaking

Generally taught in the freshman application classes

A Very Simple Example (used in Computer Science for

Two well-known O(n) solutions

int FibMemo(int n,int * A){

Excel’s simple approach

=A1+B1 Kand copy cell to the right

Pascal's triangle is constructed

The programmed solution can

There are many DP algorithms that appear throughout our

Longest Common Subsequence (LCS)

Note: The LCS is not a substring!

Longest Common Subsequence (LCS)

Let z=z1 z2 … xk be the LCS of x1 x2 … xi-1 xi

The initialized LCS Table

Copy the following cell

These represent the c(i,j)’s

And the solution is

Note diagonal increments

DP Problems with complex table manipulation

Question: What do you do with problems that require the

Excel does not allow cell function side effects! Hmm.

You might also like