Scribd
Upload
429 views2 pages

Good Will Hunting - 1st Problem

1) The document summarizes a graph theory problem posed in the movie "Good Will Hunting" where the main character solves a challenge problem on a blackboard. 2) The problem involves finding the adjacency matrix, generating functions for walks on a 4-node graph, and the matrix giving the number of 3-step walks. 3) The document provides solutions for finding the adjacency matrix, the matrix for 3-step walks, and generating functions for all walks and walks from node 1 to 3.

Uploaded by

Al
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
429 views2 pages

Good Will Hunting - 1st Problem

1) The document summarizes a graph theory problem posed in the movie "Good Will Hunting" where the main character solves a challenge problem on a blackboard. 2) The problem involves finding the adjacency matrix, generating functions for walks on a 4-node graph, and the matrix giving the number of 3-step walks. 3) The document provides solutions for finding the adjacency matrix, the matrix for 3-step walks, and generating functions for all walks and walks from node 1 to 3.

Uploaded by

Al
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

ICE: A BLACKBOARD PROBLEM Math 21b, O.

Knill

In the movie
”Good Will
Hunting”, the
main character
Will Hunting
(Matt Damon)
solves a black-
board challenge
problem, which
is given as a
challenge to a
linear algebra
class.

THE ”WILL HUNTING” PROBLEM.


4
G is the graph
1 2 3
Find.
1) the adjacency matrix A.
2) the matrix giving the number of 3 step walks.
3) the generating function for walks from point i → j.
4) the generating function for walks from points 1 → 3.
This problem belongs to linear algebra and calculus even so the problem origins from graph
theory or combinatorics. For a calculus student who has never seen the connection between
graph theory, calculus and linear algebra, the assignment is actually hard – probably too hard
- as the movie correctly indicates. The problem was posed in the last part of a linear algebra
course. An explanation of some terms:
THE ADJACENCY MATRIX. The structure of the graph can be encoded with a 4 × 4 array
which encodes how many paths of length 1, one can take in the graph from one node to an
other:
0 1 0 1
 
no one no one which can more conve-  1 0 2 1 
one none two one niently be written as L=
 
0 2 0 0

no two no no an array of numbers  
one one no no called a matrix: 1 1 0 0
Problem 2 asks to find the matrix which encodes all possible paths of length 3.
GENERATING FUNCTION. To any graph one can assign for every pair of nodes i, j a series
(ij) n (ij)
f (z) = ∞ n=0 an z , where an is the number of possible walks from node i to node j with n
P

steps. Problem 3) asks for an explicit expression of f (z) and problem 4) asks for an explicit
expression in the special case i = 1, j = 3.
Linear algebra has many relations to other fields in mathematics. It is not true that linear
algebra is just about solving systems of linear equations.
SOLUTION TO 2). [L2 ]ij is by definition of the matrix product the sum Li1 L1j + Li2 L2j +
. . . + Li4 Lhj . Each term Lik Lkj is 1 if and only if there is a path of length 2 going from i to j
passing through k. Therefore [L2 ]ij is the number of paths of length 2 going from node i to j.
Similarly, [Ln ]ij is the number of paths of length n going from i to j. The answer is

2 7 2 3
 
 7 2 12 7 
L3 = 
 
2 12 0 2

 
3 7 2 2

SOLUTION TO 3). The geometric series formula

xn = (1 − x)−1
X

holds also for matrices:


∞ ∞
n n
Ln z n )]ij = (1 − Lz)−1
X X
f (z) = [L ]ij z = [
n=0 n=0

Cramer’s formula for the inverse of a matrix which involves the determinant using the adjugate
matrix [A] which is the transpose of the minors multiplied by (−1)i+j :
A−1 = adj(A)/det(A) .
This leads to an explicit formula (−1)i+j det(1 − z[L]ji )/det(1 − zL).

SOLUTION TO 4).
 
1−5z 2 z 2 +z 2z 3 +2z 2 −4z 3 +z 2 +z
4z 4 −2z 3 −7z 2 +1 4z 4 −2z 3 −7z 2 +1 4z 4 −2z 3 −7z 2 +1 4z 4 −2z 3 −7z 2 +1
z 2 +z 1−z 2 2z−2z 3 z 2 +z
 
 
(1 − Lz) −1
=
 4z 4 −2z 3 −7z 2 +1 4z 4 −2z 3 −7z 2 +1 4z 4 −2z 3 −7z 2 +1 4z 4 −2z 3 −7z 2 +1 

 2z 3 +2z 2 2z−2z 3 −2z 3 −3z 2 +1 2z 3 +2z 2 

 4z 4 −2z 3 −7z 2 +1 4z 4 −2z 3 −7z 2 +1 4z 4 −2z 3 −7z 2 +1 4z 4 −2z 3 −7z 2 +1 
−4z 3 +z 2 +z z 2 +z 2z 3 +2z 2 1−5z 2
4z 4 −2z 3 −7z 2 +1 4z 4 −2z 3 −7z 2 +1 4z 4 −2z 3 −7z 2 +1 4z 4 −2z 3 −7z 2 +1

The entry i = 1 and j = 3 is


2z 3 + 2z 2
(1 − zL)−1
13 =
4z 4 − 2z 3 − 7z 2 + 1
because  
−z 1 −z
adj(1 − zL)13 = (−1)1+3 det( 0 −2z 0 ) = 2z 2 + 2z 3
 
−z −z 1
and
1 −z 0 −z
 
 −z 1 −2z −z  4 3 2
det(1 − zL) = det   = 4z − 2z − 7z + 1 .
 
 0 −2z 1 0 
−z −z 0 1

You might also like