In today’s video, we will be discussing the following:
1. How to solve a system of equations using both Gaussian Elimination and Gauss-Jordan Elimination (including getting a matrix into Row Echelon Form (REF) and Reduced Row Echelon Form (RREF)).
2. What is the rank of a matrix and how to find it.
3. What are pivot positions in a matrix and how to find them.
4. Linear independence and how to find the number of linearly independent rows or columns in a matrix (you will need to know this to find the rank).
5. Converting a system of linear equations into Ax = B form
Goal of Gaussian Elimination
This method is used to solve a system of linear equations. Note that these can only be LINEAR equations (nothing with log, square root, exponents, trig, etc. Only y=mx+b type equations). The goals of the Gaussian elimination method is to:
1. Put your matrix into REF. This means reduce the matrix with matrix operations (which can be swapping rows, adding one row to another, or multiplying a row by a constant) until you have 1’s along the diagonal of the matrix, and 0’s below the diagonal (the diagonal starts from the top left). Sometimes, REF form does not have 1’s down every spot of the diagonal, but there will be 0’s below every leading 1 in each row (please see 17:07 in the video for an example of a matrix like this)
2. Convert the matrix back into a system of equations. Use back substitution to solve for remaining variables once you have solved for the first variable (which should be the variable in the last row of your matrix once you reduce it to REF).
Goal of Gauss-Jordan Elimination
This method is also used to solve a system of equations. The goal of this method is to:
1. Put your matrix into RREF. This is the same as all the steps for Gaussian Elimination, however you need to keep reducing the matrix until there are ALL zeroes in the matrix EXCEPT 1’s down the diagonal of the matrix, starting from the top left.
2. Convert the matrix back into a system of equations. This will automatically give you the answers to the system.
Rank and Pivot Points
The rank of a matrix has many definitions. It is denoted as rank(A), where A is the matrix. It can be defined as all of the following:
1. The dimension of the vector space generated or spanned by its columns.
2. A measure of the “nondegenerateness” of the system of linear equations for the matrix.
3. The maximum number of linearly independent column vectors (number of non-zero columns) or the maximum number of linearly independent row vectors (number of non-zero rows).
To find the rank, the matrix must first have a matrix in REF. Then:
1. Check for the number of linearly independent rows or columns. Check either rows or columns, depending on which is lower! Rows/columns are linearly independent if no two rows can add to give a third row, and if no two rows are multiples of one another.
2. Note that the rank is only zero if you have the zero matrix (a matrix made up of only zeros!)