Charlie Hoang
Patterns Within systems of Linear Equations
Part A: Let us take into consideration the 2x2 system of linear equations: { We note that the constants in order in each equation take on an arithmetic sequence. The first being 1, 2, 3, and the other being 2, -1, -4. Thus it takes on the form: Ax + (A+n)y = A + 2n, where n is an integer. If we were to graph these two equations, we would get:
Hence the intersection point would be at (-1, 2). If we graph more equations with the parameter that -10 < A < 10 and -10 < n < 10, where both A and n are integers, that take on the form Ax + (A+n)y = A + 2n, we have:
Charlie Hoang
Again, we see that the intersection point of all of these lines is at (-1, 2), and this leads to the hypothesis that any two lines in the form Ax + (A+n)y = A + 2n will intersect at the point (-1, 2). Let us then have a 2x2 system of linear equations: { ( ( ) ) ,where n and m are integers.
We solve this system using matrices: We start with the system of equations: { Consider also the idea that the inverse of a matrix M, multiplied with M itself will yield the identity matrix. We define M to be: ( Such that M x ( ) = ( ) From here we multiply both sides by the inverse of M, thus leaving the left hand side with ( ) )
Charlie Hoang
( )= (
)( )
This yields ( ) =
x (
Solving for x in this manner yields: X= X=
( )( ( ) ( ) ( )( ) )
= -1
Likewise y can be solved in this manner: Y= Y=
( ( ( ) ) ) ( ( ) )
=2
Hence for a 2x2 system of equations in the form { ( ( ) ) , where n and m are integers,
they will intercept at point (-1, 2) QED If we extend this to a 3x3 system, where we have { ( ( ( ) ) ) ( ( ( ) ) )
, where m, n, and k are integers
A sample graph of this is shown below, with {
Charlie Hoang
As can be seen by this graph, the conjecture that arises in this situation is such that regardless the value of A, B, C, n, m, or k, as long as they are integers, the solution to these planes will be a straight line through 3D space, not just a point. If the solution is a straight line, intuitively it must encompass every point that makes up that line. When infinitely many solutions are possible, the determinant of that matrix must be equal to zero. Given a matrix M = [ ] , det M = a(ei-fh) + b(fg-di) + c(dh-eg), with this in mind, pairing up
these letters with the coefficients of the given system of equations { ( ( ( ) ) ) ( ( ( ) ) )
Charlie Hoang After some simplification, det M = a(bk-mc) + (2a+2n)(mc-bk) + (a+2n)(bk-mc), which will simplify to 0. Hence for a 3x3 in the form above, the resulting solution will always be a line. QED Part B: Let us now consider this 2x2 system {
If we now change these lines into slope-intercept form we have {
We note that it takes on the form y = ax -
which implies that a2x ay = 1
A graph of a family of equations in this form, with a variety of a-values is shown below
As more and more lines are plotted, the shape of the parabola gets more and more defined. From the graph we see that the parabola is symmetrical on the x-axis and has vertex at the origin and will thus take on the form y2 = hx, for some negative number h. It is then fitting to assume that regardless the variety of the gradients of lines in the form a2x ay = 1, there will always be a parabolic region where no lines will cross into. For a general system of equations which take on this form, we would have
Charlie Hoang
Since these are very similar, the only difference between them is the gradient value. With this we could arbitrarily assign another constant k as our gradient. y = kx which implies that k2x ky = 1 xk2 yk 1 = 0
Rewriting gives:
Now we solve for k using quadratic formula:
Looking at the discriminant of this, it is clear that y2+4x 0 Hence there will be no k value such that y2+4x < 0 This means that given any line in the form y = kx - , where k is the gradient, there will never exist a line that will cross into the region outlined by y2 < -4x From here it can be inferred that if the parabolic region will take on the form y2 = -4x QED
