Constructing Augmented Matrices

From a Linear System to an Augmented Matrix

A system of linear equations records several equations that must be considered together. Its coefficient matrix collects the coefficients of the variables, one equation per row. To keep the right-hand sides with those coefficients, we append their values as one more column. The resulting matrix is called an augmented matrix. Constructing it accurately depends on keeping the equation order and the chosen variable order fixed throughout.

Suppose a system has \(m\) equations and uses the ordered variables \(x_1,\dots,x_n\). Write its \(i\)th equation in the form \(a_{i1}x_1+\cdots+a_{in}x_n=b_i\). The number \(a_{ij}\) is the coefficient of variable \(x_j\) in equation \(i\), and \(b_i\) is that equation’s constant on the right-hand side. If a variable is absent from an equation, its coefficient is \(0\). The coefficient matrix \(\mathbf{A}\) has \(m\) rows and \(n\) columns, and the constants form the column vector \(\mathbf{b}\).

The vertical bar is a visual divider, not an extra column of numbers. It separates the coefficient columns from the constants column. The bar does not change the entries or the order of the rows. The construction follows the same order as the original system: equation 1 gives row 1, equation 2 gives row 2, and so on.

A Reliable Construction Process

Before writing entries, record the variables in their specified order. Then process each equation separately: enter its coefficients from left to right in that order, followed by its right-hand-side constant. In particular, do not omit a slot when a variable is missing; enter \(0\). A negative coefficient remains negative, and a fraction remains a fraction. The constant is copied from the right-hand side, not inserted among the variable coefficients.

This process is a way to record a system, not a method for solving it. At this stage, the main task is accurate placement: every coefficient belongs to a particular equation and a particular variable, while every constant belongs to a particular equation.

Worked Constructions

What the Construction Preserves

The entries are not simply a collection of numbers. Each row belongs to one equation, each coefficient column belongs to one variable in the fixed order, and the last column belongs to the right-hand sides. This organization gives a precise connection between the original system and a matrix equation.

Proof. Fix any assignment \(\mathbf{x}=(x_1,\dots,x_n)^{\top}\). By the definition of matrix multiplication, the \(i\)th entry of \(\mathbf{A}\mathbf{x}\) is \(a_{i1}x_1+\cdots+a_{in}x_n\). Thus the \(i\)th entry of \(\mathbf{A}\mathbf{x}=\mathbf{b}\) states \(a_{i1}x_1+\cdots+a_{in}x_n=b_i\), which is precisely equation \(i\) of the system. This holds for every \(i\) from \(1\) to \(m\). Therefore the matrix equation holds exactly when every equation in the system holds. \(\square\)

The theorem explains why the constants column is appended in the same row as its equation’s coefficients: the \(i\)th matrix equation must compare the expression formed from row \(i\) with the constant \(b_i\) from that same equation. Putting a constant in a different row would pair it with the wrong equation.

Proof. For each row index \(i\), the \(i\)th equation fixes its coefficient of each ordered variable \(x_j\), so it fixes every entry \(a_{ij}\) for \(1\leq j\leq n\). The same equation fixes its right-hand-side constant \(b_i\), which occupies the final entry of row \(i\). Since this determines every entry in every row, there is only one matrix formed by the stated construction. \(\square\)

The condition about order matters. Changing the variable order changes which coefficient belongs in each column. Changing the equation order changes which equation belongs in each row. Either change can produce a different displayed matrix, even though the same equations and variables are involved. In a calculation, write down the orders first rather than relying on an assumed default.

Common Pitfalls

Using a familiar variable order instead of the stipulated one. In the second worked example, the first equation is \(3x-2z+y=5\), but the specified order is \(z,x,y\). A tempting incorrect first row is \((3,-2,1\mid 5)\). That row records coefficients in the order \(x,z,y\), not \(z,x,y\). The correction is \((-2,3,1\mid 5)\): the coefficient of \(z\) comes first, followed by the coefficients of \(x\) and \(y\).

Omitting a zero coefficient. In the first example, the second equation has no \(z\)-term. Leaving that position out would give a row with too few coefficient entries and would shift the constant into the wrong position. The correct row is \((-1,4,0\mid 0)\): a zero records the absent \(z\)-term, and the final zero records the equation’s right-hand side.

Treating the constant as a variable coefficient. In the third example, the second equation is \(-2y=5\). Its coefficient row, in the order \(x,y,w\), is \((0,-2,0)\); the \(5\) is appended after the divider. It is not a coefficient of a fourth variable. For a quick audit, count \(n\) entries before the divider and exactly one constant after it in every row.