Introduction to Systems of Linear Equations

You are probably familiar with the concept of a system of linear equations and with some methods for solving such systems. In this section, we will look at the algebra and geometry of finding and interpreting solutions of systems of linear equations. We will start with two-variable and three-variable systems, then move on to systems involving more variables.

Algebra of Linear Systems

When you were first introduced to systems of equations, you learned to solve for one variable in terms of the other(s), then substitute. Here, we will introduce another method. This alternative method involves adding multiples of one equation to another equation in order to eliminate one of the variables. This method will form the foundation for an algorithm we will develop for solving linear systems and performing other computations related to systems. Exploration init:systwoeqs1 illustrates how the second method works.

The purpose of this problem is to formalize what you may already know (perhaps under a different name) about elementary row operations as means of solving systems of linear equations. Consider the system

We will begin by adding twice the first row to the second row, and replacing the second row with the sum.

Note that this step eliminates from the second equation.

Next we divide both sides of the second equation by .

We now know what is. Our next goal is to eliminate from the first equation. To this end, we subtract twice the second row from the first row and replace the first row with the difference.

Next we multiply both sides of the first equation by .

Finally, we can switch the order of equations in order to display in the top row.

This solution can be written as an ordered pair .

In Exploration init:systwoeqs1 we introduced elementary row operations and the notation associated with them. We now make these definitions formal.

As we applied elementary row operations to the system in Exploration init:systwoeqs1, the system changed, but a quick check will convince you that all six systems have the same solution: . The six systems are said to be equivalent.

It turns out that if a system of equations is transformed into another system through a sequence of elementary row operations, the new system will be equivalent to the original system, in other words, both systems will have the same solution(s). We will formalize this statement as Theorem th:elemRowOpsEquivSys at the end of this section.

At this point you may be wondering whether it will always be possible to take a system of three equations and three unknowns and use elementary row operations to transform it to a system of the form The short answer to this question is NO. The existence of an equivalent system of this form implies that the original system has a unique solution . However, it is possible for a system to have no solutions or to have infinitely many solutions. We will study these different possibilities from an algebraic perspective in subsequent sections. For now, we will attempt to gain insight into existence and uniqueness of solutions through geometry.

Geometry of Linear Systems in Two Variables

Exploration Problem init:systwoeqs1 offers an example of a linear system of two equations and two unknowns (variables) with a unique solution.

Geometrically, the graph of each equation is a line in . The point is a solution to both equations, so it must lie on both lines. The graph below shows the two lines intersecting at .

Given a system of two equations with two unknowns, there are three possible geometric outcomes.

  • First, the graphs of the two equations intersect at a point. If this is the case, the system has exactly one solution. We say that the system is consistent and has a unique solution.

  • Second, the two lines may have no points in common. If this is the case, the system has no solutions. We say that the system is inconsistent.

  • Finally, the two lines may coincide. In this case, there are infinitely many points that satisfy both equations simultaneously. We say that the system is consistent and has infinitely many solutions.

Given a linear system in two variables and more than two equations, we have a variety of geometric possibilities. In terms of the number of solutions, there are three possibilities.

  • First, it is possible for the graphs of all equations in the system to intersect at a single point, giving us a unique solution.

  • Second, it is possible for the graphs to have no points common to all of them. If this is the case, the system is inconsistent.

  • Finally, it is possible for all of the lines to coincide, giving us infinitely many solutions.

Geometry of Linear Systems in Three Variables

In Example ex:threeeqthreevars1 we solved the following linear system of three equations and three unknowns We found that the system has a unique solution . The graph of each equation is a plane. The three planes intersect at a single point, as shown in the figure.

Given a linear system of three equations in three variables, there are three ways in which the system can be consistent.

  • First, the three planes could intersect at a single point, giving us a unique solution.

  • Second, the three planes can intersect in a line, forming a paddle-wheel shape. In this case, every point along the line of intersection is a solution to the system, giving us infinitely many solutions.

  • Finally, the three planes can coincide. If this is the case, there are infinitely many solutions.

There are four ways for a system to be inconsistent. They are depicted below.

General Systems of Linear Equations

An -tuple is a solution to the equation provided that it turns the equation into a true statement. The set of all -tuples that are solutions to a given equation is called the graph of the equation. The graph of a linear equation in two variables is a line in . The graph of a linear equation in three variables is a plane in . In , for , we say that the graph of a linear equation is a hyperplane. A hyperplane cannot be visualized, but we can still talk about intersections of hyperplanes and their other attributes in algebraic terms.

A linear system of equations and unknowns is typically written as follows

A solution to a system of linear equations in variables is an -tuple that satisfies every equation in the system. All solutions to a system of equations, taken together, form a solution set.

Recall that to solve systems of equations in this section, we utilized three elementary row operations. These operations are:

(a)
Switching the order of two equations
(b)
Multiplying both sides of an equation by the same non-zero constant
(c)
Adding a multiple of one equation to another

Proof
Clearly, the order of equations does not affect the solution set, so item:rowswap produces an equivalent system. Next, you learned years ago that multiplying both sides of an equation by a non-zero constant does not change its solution set, which establishes that item:constantmult produces an equivalent system. To see that item:addrow produces an equivalent system, note that if we add a multiple of an equation to another equation in the system, we are adding the same thing to both sides, which does not change the solution set of that equation, nor of the system.

Practice Problems

Give a graphical illustration of each of the following scenarios for a system of three equations and two unknowns:
(a)
The system of three equations is inconsistent, but a combination of any two of the three equations forms a consistent system.
(b)
The system is consistent and has a unique solution.
(c)
The system is consistent and has infinitely many solutions.
(d)
The system is inconsistent and no two equations form a consistent system.
Solve each system of linear equations or demonstrate that a solution does not exist, and interpret your results geometrically.

Solution:

Solution:

Consider the following system of equations.
Find all possible values of k such that this system has no solution.

Solution:

Find all possible values of k such that this system has infinitely many solutions.

Solution:

Why is there a non-zero provision in Part item:constantmult of Definition def:elemrowops? Why is there not a non-zero provision in Part item:addrow?
Suppose the following system was obtained from system by adding twice the second row of to the first row. Find system .

Answer:

The following figures show a geometric depiction of two equivalent systems. (The systems are equivalent because they have the same solution set.) Can the first system be transformed into the second system by elementary row operations? If so, how?

Begin by carrying the first system to Then carry this system to the second system. (If you can figure out how to carry the second system to this one, you should be able to reverse the process.)
Consider the system of equations

Show that if is a solution to this system, and if we apply elementary row operation item:addrow to the system, then will be a solution to the new system of equations.

Demonstrate that elementary row operations are reversible by answering the following questions. Be specific about the elementary row operation that you would use.
(a)
Suppose we obtained system (B) from system (A) by swapping two equations. How would we obtain system (A) from system (B)?
(b)
Suppose we obtained system (B) from system (A) by multiplying one of the equations of (A) by a non-zero constant . How would we obtain system (A) from system (B)?
(c)
Suppose we obtained system (B) from system (A) by adding a multiple of one of the equations of (A) to another. How would we obtain system (A) from system (B)?