Pre-Calculus: Solving Systems Using Matrices

Solving systems of linear equations is a cornerstone of algebra, but as the number of equations and variables grows, substitution and elimination become messy and inefficient. Matrices provide a powerful, systematic, and programmable method to handle systems of any size. Mastering this technique is essential for future studies in linear algebra, engineering, computer science, and economics, where it forms the basis for solving complex real-world problems.

Representing a Linear System as an Augmented Matrix

A system of linear equations is a set of equations with multiple variables where each term is either a constant or a product of a constant and a single variable. The first step in the matrix method is to translate this system into a compact, rectangular array of numbers called a matrix.

Consider the system: $$2x+y-z=8$$ $$-3x-y+2z=-11$$ $$-2x+y+2z=-3$$

We extract the coefficients of the variables and the constants from the right-hand side to form an augmented matrix. Each row corresponds to an equation, and each column (except the last) corresponds to a variable's coefficients. The final column, separated by a line, holds the constants.

The augmented matrix for this system is: $$\left[\begin{array}{rrrr}2& 1& -1& 8\\ -3& -1& 2& -11\\ -2& 1& 2& -3\end{array}\right]$$

The variables $x$, $y$, and $z$ are implied by the column positions. This representation strips away the variables and focuses purely on the numerical relationships, setting the stage for efficient manipulation.

Row Operations and Row-Echelon Form

We solve the matrix using row operations, which are analogous to the algebraic operations you use to solve systems by elimination. These operations are legal because they produce equivalent systems with the same solution set.

The three elementary row operations are:

1. Swap: Interchange two rows.
2. Scale: Multiply all entries in a row by a non-zero constant.
3. Replace: Add a multiple of one row to another row.

Our goal is to use these operations to transform the matrix into a simpler, triangular form. This simpler form is called row-echelon form (REF), which has these characteristics:

• All non-zero rows are above any rows of all zeros.
• The leading entry (first non-zero number from the left) of a row is always strictly to the right of the leading entry of the row above it.
• All entries in a column below a leading entry are zeros.

The process of achieving REF is called Gaussian elimination. Let's apply it to our example matrix.

We'll perform operations to get zeros below the leading entries.

1. Get a 1 in the top-left position (Row 1, Column 1). We can scale Row 1 by $1/2$, but working with integers is often easier. Let's start by creating a zero below the 2 in column 1.

• $R_2=3R_1+2R_2$: This eliminates the $-3$ in $R_2$.
• $R_3=R_1+R_3$: This eliminates the $-2$ in $R_3$.

Result: $$\left[\begin{array}{rrrr}2& 1& -1& 8\\ 0& 1& -1& 2\\ 0& 2& 1& 5\end{array}\right]$$

1. Now, get a zero below the new leading 1 in $R_2$, Column 2.

• $R_3=-2R_2+R_3$

Result (Row-Echelon Form): $$\left[\begin{array}{rrrr}2& 1& -1& 8\\ 0& 1& -1& 2\\ 0& 0& 3& 1\end{array}\right]$$

This matrix is in REF. Think of it like a neatly sorted bookshelf where each step down is indented, making it easy to start solving from the bottom up.

Reduced Row-Echelon Form and Back-Substitution

While REF allows for solution via back-substitution, an even cleaner form exists. Reduced row-echelon form (RREF) adds two more requirements to REF:

• The leading entry in every non-zero row is 1.
• Each leading 1 is the only non-zero entry in its column.

The process of achieving RREF is called Gauss-Jordan elimination. Starting from our REF matrix, we continue:

1. Scale the third row by $1/3$ to get a leading 1: $R_3=(1/3)R_3$.

$$\left[\begin{array}{rrrr}2& 1& -1& 8\\ 0& 1& -1& 2\\ 0& 0& 1& 1/3\end{array}\right]$$

1. "Clear" the column above each leading 1.

• First, use the 1 in $R_3$ to clear column 3 above it: $R_2=R_2+R_3$ and $R_1=R_1+R_3$.

$$\left[\begin{array}{rrrr}2& 1& 0& 25/3\\ 0& 1& 0& 7/3\\ 0& 0& 1& 1/3\end{array}\right]$$

• Now, use the 1 in $R_2$ to clear column 2 above it: $R_1=R_1-R_2$.

$$\left[\begin{array}{rrrr}2& 0& 0& 6\\ 0& 1& 0& 7/3\\ 0& 0& 1& 1/3\end{array}\right]$$

1. Finally, scale $R_1$ by $1/2$ to get a leading 1: $R_1=(1/2)R_1$.

Final RREF: $$\left[\begin{array}{rrrr}1& 0& 0& 2\\ 0& 1& 0& 7/3\\ 0& 0& 1& 1/3\end{array}\right]$$

This matrix translates directly back to the solution: $x=2$, $y=7/3$, $z=1/3$. The RREF makes the solution immediately obvious without any further algebra.

Interpreting Solutions: Unique, Dependent, and Inconsistent Systems

Not every system has a single, unique solution. The final RREF of the augmented matrix reveals three possible outcomes.

1. Unique Solution (Consistent & Independent): This is what we just saw. The RREF has a diagonal of 1's in the coefficient columns (left of the bar), giving one explicit solution for each variable.

1. Infinitely Many Solutions (Consistent & Dependent): This occurs when the system has at least one free variable. In the RREF, a column corresponding to a variable lacks a leading 1. That variable is free and can be assigned any real number value (e.g., $t$), with the other variables expressed in terms of it.

Example RREF: $$\left[\begin{array}{rrrr}1& 2& 0& 4\\ 0& 0& 1& 1\\ 0& 0& 0& 0\end{array}\right]$$ Here, $y$ is a free variable (let $y=t$). The solution is: $x=4-2t$, $y=t$, $z=1$.

1. No Solution (Inconsistent): This happens when the equations contradict each other. In the RREF, this is shown by a row where all the coefficient entries are 0, but the constant entry (right of the bar) is non-zero. This row translates to the false statement $0=c$ (where $c\ne 0$).

Example RREF: $$\left[\begin{array}{rrr}1& 3& 5\\ 0& 0& 2\\ 0& 0& 0\end{array}\right]$$ The second row, $0x+0y=2$, is impossible. The system is inconsistent.

Common Pitfalls

Incorrectly Applying Row Operations: A frequent error is performing an operation that is not one of the three elementary row operations. For example, multiplying a row by zero or swapping columns is not allowed, as it changes the solution set of the system. Always ensure your operation is a Swap, Scale (by a non-zero number), or Replace.

Misinterpreting Free Variables: When you see a column without a leading 1 in RREF, it's tempting to think the system has no solution. The opposite is true—it has infinite solutions. The key is to correctly identify which variable is free and then solve the other variables in terms of it. Failing to do so leads to an incomplete answer.

Arithmetic Errors in Sequential Steps: Row operations are sequential; an arithmetic mistake in an early step propagates through all subsequent steps, leading to a wrong final matrix. Work methodically, check each step, and consider using the original matrix to verify your solution by plugging it back into the original equations.

Confusing REF and RREF: Row-echelon form is a valid stopping point from which you can use back-substitution. Reduced row-echelon form goes further to provide the solution directly. Be clear on what form your problem requires. In many computational contexts, RREF is the expected final answer.

Summary

• An augmented matrix provides a compact, numerical representation of a linear system, separating coefficient data from constants.
• Row operations (swap, scale, replace) are used in Gaussian elimination to achieve row-echelon form (REF) and in Gauss-Jordan elimination to achieve reduced row-echelon form (RREF), the latter giving the most direct solution.
• The final matrix reveals the system's nature: a diagonal of 1's indicates a unique solution, a column without a leading 1 indicates a dependent system with free variables and infinite solutions, and a row of zeros equaling a non-zero constant proves the system is inconsistent with no solution.
• Accuracy in arithmetic and a strict adherence to allowed row operations are critical to reliably solving systems using matrices.