Lesson Explainer: Solution Cases of System of Linear Equations Mathematics

In this explainer, we will learn how to determine whether a linear system of equations has a unique solution, no solution, or an infinite number of solutions.

When working with a system of linear equations, it is easy to assume that there is a solution. This solution might be a unique solution, where all of the variables have only one possible value. Alternatively, there may be infinitely many solutions, meaning that each variable can be expressed in terms of some independent parameters. Although there are fundamental differences in these two types of solutions, they are similar to each other in the sense that they represent systems of linear equations which do actually have a solution. We broadly categorize these two solution types by saying that they correspond to systems of linear equations that are “consistent.” It may also be the case that it is not possible to obtain a solution to a system of linear equations, in which case the problem is known as “inconsistent.” For example, if we were to take the system of linear equations then we would quickly find that it is not possible to solve this problem. Thinking in a geometric sense, it is because the two lines encoded by the two equations are parallel, meaning that they will never intersect. In this case we would say that the system is “inconsistent,” as it is not possible to find a solution. This can be more clearly shown by rearranging both of the above equations for to read

Now we have as the subject of both equations, and we can equate them to find which cancels to give , which is blatantly absurd. Given this logical inconsistency, we say that the whole system of linear equations is inconsistent.

On the basis of this preliminary example, it is probably worth asking ourselves what we might expect to see if we are working with an inconsistent system of linear equations. Suppose we take, for example, the system of linear equations

It is not obvious, at this stage, whether the system of linear equations has a unique solution, has infinitely many solutions, or does not have any solutions at all. In this instance, any appropriate method would show that there is a unique solution with the values

We have formatted the answer in a very particular way, the reasons for which will become apparent later on in the explainer. Now suppose that, instead of the above set of equations, we had begun with a system of linear equations which led us to the values

The last line is obviously ridiculous, as it suggests that . In this case, we would have said that the system of linear equations was inconsistent, as it produces an absurdity. We will now take the time to properly define what it means for a system of equations to be inconsistent, before preparing a second definition that will allow us to more easily classify systems of equations as either consistent or inconsistent.

When attempting to solve a system of linear equations, the most common method is arguably the Gauss–Jordan method, where elementary row operations are used to manipulate the system of equations into an expression for the solution, without changing any properties of the system itself. Most often, this will be achieved by first writing the system of equations in a convenient form which encapsulates all of the information, which in this case is the corresponding augmented coefficient matrix.

In translating from a system of linear equations to its augmented coefficient matrix, we have essentially already understood one of the key steps for determining whether the system is consistent or inconsistent. To elucidate this method, we take the augmented coefficient matrix of a system of linear equations where there are equations with variables, writing these variables above the columns that they correspond to:

Suppose now that we used elementary row operations, or any other suitable method, to transform this system into the reduced echelon form

This corresponds to a unique solution, where may only take one possible value in order to solve the system of linear equations. In this case the system is consistent, as there is at least one valid solution. However, suppose we had arrived at the reduced echelon form

This clearly does not make any sense, as we cannot have if the latter quantity is nonzero. This would, therefore, represent a system that is inconsistent. The property is summarized by the following theorem.

A pivot is the first nonzero entry of a matrix. For an augmented coefficient matrix, if one of the pivots is in the right-most column, then all other entries in the row are zero. For variables , suppose this applied to row of the augmented coefficient matrix after some number of row operations had been performed. By having the first nonzero entry in the right-most column, this would imply that where . This is clearly impossible as the left-hand side is equal to zero and the right-hand side is not. We will demonstrate one example of this property in practice.

Suppose that we have the system of linear equations

This can be represented by the augmented coefficient matrix

We attempt to solve this system using the method of Gauss–Jordan elimination, by using elementary row operations to attempt to manipulate the matrix into reduced echelon form. We first highlight all of the pivot entries:

To begin working towards reduced echelon form, we must eliminate all of the nonzero entries which are below the pivot in the first row. We use the joint row operations and to give the matrix

Now we need to eliminate the nonzero entry which is below the pivot in the second row. This is achieved with the row operation , giving the matrix

This matrix is in an echelon form and yet there is a pivot in the right-most column. By the theorem above, this means that the system is inconsistent. Writing out this system of equations would give

The third equation contains a false statement, validating our assertion that the system is inconsistent.

We will now give 3 examples where we must recognize whether or not a given augmented matrix is representing a system of linear equations that is consistent, or inconsistent. If the system of equations is consistent, then we will also be expected to recognize whether or not the system of equations has a unique solution or infinitely many solutions.

The previous questions were based on recognizing whether an augmented matrix is consistent, by looking at the entries and the location of the pivots. In the two questions below we will examine specific augmented matrices containing unknown variables, and we will use row operations to show how these systems of linear equations can be classified with respect to these variables.

In most cases, it is impossible to immediately deduce whether a system of linear equations is consistent or inconsistent. It is therefore wisest to assume that a system is consistent and to begin finding the solution using the normal method of Gauss–Jordan elimination to achieve an echelon form. Provided that we know how to recognize that a system is inconsistent whenever it becomes apparent, there is no disadvantage to proceeding as though the system is consistent. Within the broader scope of linear algebra, there are multiple ways of interpreting what it means for a system of equations to be inconsistent. However, one common way is to try and understand this physically. Earlier in the explainer, we had the system of linear equations and we rearranged both of these straight line equations to get the two equations

These two straight lines are parallel, meaning that they will never intersect. This is the framework that is used to understand what it means when a system of linear equations is inconsistent but has more variables or more equations. Generally speaking, we think of an inconsistent system as referring to straight lines, planes, hyperplanes, and so on that do not intersect each other. There is clearly nothing wrong with a situation where any two geometric objects do not meet each other, but it is necessary to have terminology to distinguish this case from the those where there is either a unique meeting point or infinitely many points.