20 Systems of Linear Equations

1.3 Homogeneous Equations

A system of equations in the variables x1 , x2 , . . . , xn is called homogeneous if all the constant terms are
zero—that is, if each equation of the system has the form

a1 x1 + a2 x2 + · · · + an xn = 0

Clearly x1 = 0, x2 = 0, . . . , xn = 0 is a solution to such a system; it is called the trivial solution. Any

solution in which at least one variable has a nonzero value is called a nontrivial solution. Our chief goal
in this section is to give a useful condition for a homogeneous system to have nontrivial solutions. The
following example is instructive.

Example 1.3.1
Show that the following homogeneous system has nontrivial solutions.

x1 − x2 + 2x3 − x4 = 0
2x1 + 2x2 + x4 = 0
3x1 + x2 + 2x3 − x4 = 0

Solution. The reduction of the augmented matrix to reduced row-echelon form is outlined below.
     
1 −1 2 −1 0 1 −1 2 −1 0 1 0 1 0 0
 2 2 0 1 0 → 0 4 −4 3 0  →  0 1 −1 0 0 
3 1 2 −1 0 0 4 −4 2 0 0 0 0 1 0

The leading variables are x1 , x2 , and x4 , so x3 is assigned as a parameter—say x3 = t. Then the

general solution is x1 = −t, x2 = t, x3 = t, x4 = 0. Hence, taking t = 1 (say), we get a nontrivial
solution: x1 = −1, x2 = 1, x3 = 1, x4 = 0.

The existence of a nontrivial solution in Example 1.3.1 is ensured by the presence of a parameter in the
solution. This is due to the fact that there is a nonleading variable (x3 in this case). But there must be
a nonleading variable here because there are four variables and only three equations (and hence at most
three leading variables). This discussion generalizes to a proof of the following fundamental theorem.

Theorem 1.3.1
If a homogeneous system of linear equations has more variables than equations, then it has a
nontrivial solution (in fact, infinitely many).

Proof. Suppose there are m equations in n variables where n > m, and let R denote the reduced row-echelon
form of the augmented matrix. If there are r leading variables, there are n − r nonleading variables, and so
n − r parameters. Hence, it suffices to show that r < n. But r ≤ m because R has r leading 1s and m rows,
and m < n by hypothesis. So r ≤ m < n, which gives r < n.
 1.3. Homogeneous Equations 21

Note that the converse of Theorem 1.3.1 is not true: if a homogeneous system has nontrivial solutions,
it need not have more variables than equations (the system x1 + x2 = 0, 2x1 + 2x2 = 0 has nontrivial
solutions but m = 2 = n.)
Theorem 1.3.1 is very useful in applications. The next example provides an illustration from geometry.

Example 1.3.2
We call the graph of an equation ax2 + bxy + cy2 + dx + ey + f = 0 a conic if the numbers a, b, and
c are not all zero. Show that there is at least one conic through any five points in the plane that are
not all on a line.

Solution. Let the coordinates of the five points be (p1 , q1 ), (p2 , q2 ), (p3 , q3 ), (p4 , q4 ), and
(p5 , q5 ). The graph of ax2 + bxy + cy2 + dx + ey + f = 0 passes through (pi , qi ) if

ap2i + bpi qi + cq2i + d pi + eqi + f = 0

This gives five equations, one for each i, linear in the six variables a, b, c, d, e, and f . Hence, there
is a nontrivial solution by Theorem 1.3.1. If a = b = c = 0, the five points all lie on the line with
equation dx + ey + f = 0, contrary to assumption. Hence, one of a, b, c is nonzero.

Linear Combinations and Basic Solutions

As for rows, two columns are regarded as equal if they have the same number of entries and corresponding
entries are the same. Let x and y be columns with the same number of entries. As for elementary row
operations, their sum x + y is obtained by adding corresponding entries and, if k is a number, the scalar
product kx is defined by multiplying each entry of x by k. More precisely:
       
x1 y1 x1 + y1 kx1
 x2   y2   x2 + y2   kx2 
       
If x =  ..  and y =  ..  then x + y =  ..  and kx =  ..  .
 .   .   .   . 
xn yn xn + yn kxn

A sum of scalar multiples of several columns is called a linear combination of these columns. For
example, sx + ty is a linear combination of x and y for any choice of numbers s and t.

Example 1.3.3
         
3 −1 6 −5 1
If x = and then 2x + 5y = + = .
−2 1 −4 5 1
22 Systems of Linear Equations

Example 1.3.4
         
1 2 3 0 1
Let x = 0 , y = 1 and z = 1 . If v = −1 and w = 1 , determine whether v
        
1 0 1 2 1
and w are linear combinations of x, y and z.

Solution. For v, we must determine whether numbers r, s, and t exist such that v = rx + sy + tz,
that is, whether
         
0 1 2 3 r + 2s + 3t
 −1  = r  0  + s  1  + t  1  =  s+t 
2 1 0 1 r +t

Equating corresponding entries gives a system of linear equations r + 2s + 3t = 0, s + t = −1, and

r + t = 2 for r, s, and t. By gaussian elimination, the solution is r = 2 − k, s = −1 − k, and t = k
where k is a parameter. Taking k = 0, we see that v = 2x − y is a linear combination of x, y, and z.
Turning to w, we again look for r, s, and t such that w = rx + sy + tz; that is,
         
1 1 2 3 r + 2s + 3t
 1  = r 0 +s 1 +t  1  =  s+t 
1 1 0 1 r +t

leading to equations r + 2s + 3t = 1, s + t = 1, and r + t = 1 for real numbers r, s, and t. But this

time there is no solution as the reader can verify, so w is not a linear combination of x, y, and z.

Our interest in linear combinations comes from the fact that they provide one of the best ways to
describe the general solution of a homogeneous system of linear equations. When  solving such a system

x1
 x2 
 
with n variables x1 , x2 , . . . , xn , write the variables as a column6 matrix: x =  .. . The trivial solution
 . 
xn
 
0
 0 
 
is denoted 0 =  .. . As an illustration, the general solution in Example 1.3.1 is x1 = −t, x2 = t, x3 = t,
 . 
0
and x4 = 0,where t is a parameter, and we would now express this by saying that the general solution is
−t
 t 
x= 
 t , where t is arbitrary.
0
Now let x and y be two solutions to a homogeneous system with n variables. Then any linear combi-
nation sx + ty of these solutions turns out to be again a solution to the system. More generally:

Any linear combination of solutions to a homogeneous system is again a solution. (1.1)

6 The reason for using columns will be apparent later.
 1.3. Homogeneous Equations 23

In
 fact,suppose
 thata typical equation in the system is a1 x1 + a2 x2 + · · · + an xn = 0, and suppose that
x1 y1
 x2   y2 
   
x =  .. , y =  ..  are solutions. Then a1 x1 + a2 x2 + · · · + an xn = 0 and a1 y1 + a2 y2 + · · · + an yn = 0.
 .   . 
xn y
 n 
sx1 + ty1
 sx2 + ty2 
 
Hence sx + ty =  ..  is also a solution because
 . 
sxn + tyn

a1 (sx1 + ty1 ) + a2 (sx2 + ty2 ) + · · · + an (sxn + tyn )

= [a1 (sx1 ) + a2 (sx2 ) + · · · + an (sxn )] + [a1(ty1 ) + a2 (ty2) + · · · + an (tyn )]
= s(a1x1 + a2 x2 + · · · + an xn ) + t(a1y1 + a2 y2 + · · · + an yn )
= s(0) + t(0)
=0

A similar argument shows that Statement 1.1 is true for linear combinations of more than two solutions.
The remarkable thing is that every solution to a homogeneous system is a linear combination of certain
particular solutions and, in fact, these solutions are easily computed using the gaussian algorithm. Here is
an example.

Example 1.3.5
Solve the homogeneous system with coefficient matrix
 
1 −2 3 −2
A =  −3 6 1 0 
−2 4 4 −2

Solution. The reduction of the augmented matrix to reduced form is

 
  1 −2 0 − 5 0 1
1 −2 3 −2 0  
 −3   3 
6 1 0 0 → 0 0 1 −5 0 
 
−2 4 4 −2 0
0 0 0 0 0

so the solutions are x1 = 2s + 15 t, x2 = s, x3 = 53 , and x4 = t by gaussian elimination. Hence we can

write the general solution x in the matrix form
       1 
x1 2s + 15 t 2 5
 x2   s   1   0 
x=       
 x3  =  3 t  = s  0  + t  3  = sx1 + tx2 .
5 5
x4 t 0 1
24 Systems of Linear Equations

   1 
2 5
 1   0 
Here x1 =    
 0  and x2 =  3  are particular solutions determined by the gaussian algorithm.
5
0 1

The solutions x1 and x2 in Example 1.3.5 are denoted as follows:

Definition 1.5 Basic Solutions

The gaussian algorithm systematically produces solutions to any homogeneous linear system,
called basic solutions, one for every parameter.

Moreover, the algorithm gives a routine way to express every solution as a linear combination of basic
solutions as in Example 1.3.5, where the general solution x becomes
   1     
2 5 2 1
 1   0   1  1  0 
x = s       
 0  + t  3  = s  0  + 5t  3 
5
0 1 0 5

Hence by introducing a new parameter r = t/5 we can multiply the original basic solution x2 by 5 and so
eliminate fractions. For this reason:

Convention:
Any nonzero scalar multiple of a basic solution will still be called a basic solution.

In the same way, the gaussian algorithm produces basic solutions to every homogeneous system, one
for each parameter (there are no basic solutions if the system has only the trivial solution). Moreover every
solution is given by the algorithm as a linear combination of these basic solutions (as in Example 1.3.5).
If A has rank r, Theorem 1.2.2 shows that there are exactly n − r parameters, and so n − r basic solutions.
This proves:

Theorem 1.3.2
Let A be an m × n matrix of rank r, and consider the homogeneous system in n variables with A as
coefficient matrix. Then:

1. The system has exactly n − r basic solutions, one for each parameter.

2. Every solution is a linear combination of these basic solutions.

 1.3. Homogeneous Equations 25

Example 1.3.6
Find basic solutions of the homogeneous system with coefficient matrix A, and express every
solution as a linear combination of the basic solutions, where
 
1 −3 0 2 2
 −2 6 1 2 −5 
A=  3 −9 −1 0

7 
−3 9 2 6 −8

Solution. The reduction of the augmented matrix to reduced row-echelon form is

   
1 −3 0 2 2 0 1 −3 0 2 2 0
 −2 6 1 2 −5 0   0 1 6 −1 0 
 → 0 
 3 −9 −1 0 7 0   0 0 0 0 0 0 
−3 9 2 6 −8 0 0 0 0 0 0 0

so the general solution is x1 = 3r − 2s − 2t, x2 = r, x3 = −6s + t, x4 = s, and x5 = t where r, s, and

t are parameters. In matrix form this is
         
x1 3r − 2s − 2t 3 −2 −2
 x2   r   1   0   0 
         
    
x =  x3  =  −6s + t  = r  0  + s    
 −6  + t  1 

 x4   s   0   1   0 
x5 t 0 0 1

Hence basic solutions are

     
3 −2 −2
 1   0   0 
     
x1 = 
 0  , x2 = 
  −6  , x3 = 
  1 

 0   1   0 
0 0 1
26 Systems of Linear Equations

Exercises for 1.3

   
Exercise 1.3.1 Consider the following statements about 2 1
a system of linear equations with augmented matrix A. In Exercise 1.3.3 Let x =  1  , y =  0 , and
each case either prove the statement or give an example −1 1
 
for which it is false. 1
z =  1 . In each case, either write v as a linear com-
−2
a. If the system is homogeneous, every solution is
bination of x, y, and z, or show that it is not such a linear
trivial.
combination.
b. If the system has a nontrivial solution, it cannot be    
0 4
homogeneous. a. v =  1  b. v =  3 
−3 −4
c. If there exists a trivial solution, the system is ho-    
mogeneous. 3 3
c. v =  1  d. v =  0 
d. If the system is consistent, it must be homoge- 0 3
neous.
Exercise 1.3.4 In each case, either express y as a linear
combination of a1 , a2 , and a3 , or show that it is not such
Now assume that the system is homogeneous.
a linear combination. Here:
     
−1 3 1
e. If there exists a nontrivial solution, there is no triv-  3   1   1 
ial solution. a1 =    
 0  , a2 =  2  , and a3 =  1 
 

1 0 1
f. If there exists a solution, there are infinitely many
solutions.    
1 −1
g. If there exist nontrivial solutions, the row-echelon  2   9 
form of A has a row of zeros. a. y =  4 
 b. y =   2 


0 6
h. If the row-echelon form of A has a row of zeros,
there exist nontrivial solutions. Exercise 1.3.5 For each of the following homogeneous
systems, find a set of basic solutions and express the gen-
i. If a row operation is applied to the system, the new eral solution as a linear combination of these basic solu-
system is also homogeneous. tions.

a. x1 + 2x2 − x3 + 2x4 + x5 = 0
Exercise 1.3.2 In each of the following, find all values
x1 + 2x2 + 2x3 + x5 = 0
of a for which the system has nontrivial solutions, and
2x1 + 4x2 − 2x3 + 3x4 + x5 = 0
determine all solutions in each case.
b. x1 + 2x2 − x3 + x4 + x5 = 0
a. x − 2y + z = 0 b. x + 2y + z = 0 −x1 − 2x2 + 2x3 + x5 = 0
x + ay − 3z = 0 x + 3y + 6z = 0 −x1 − 2x2 + 3x3 + x4 + 3x5 = 0
−x + 6y − 5z = 0 2x + 3y + az = 0
c. x1 + x2 − x3 + 2x4 + x5 = 0
c. x + y − z = 0 d. ax + y + z = 0 x1 + 2x2 − x3 + x4 + x5 = 0
ay − z = 0 x+y− z=0 2x1 + 3x2 − x3 + 2x4 + x5 = 0
x + y + az = 0 x + y + az = 0 4x1 + 5x2 − 2x3 + 5x4 + 2x5 = 0
 1.4. An Application to Network Flow 27

d. x1 + x2 − 2x3 − 2x4 + 2x5 = 0 Exercise 1.3.8 The graph of an equation ax+by+cz = 0

2x1 + 2x2 − 4x3 − 4x4 + x5 = 0 is a plane through the origin (provided that not all of a,
x1 − x2 + 2x3 + 4x4 + x5 = 0 b, and c are zero). Use Theorem 1.3.1 to show that two
−2x1 − 4x2 + 8x3 + 10x4 + x5 = 0 planes through the origin have a point in common other
than the origin (0, 0, 0).
Exercise 1.3.6 Exercise 1.3.9

a. 
Does Theorem 1.3.1 imply that the system a. Show that there is a line through any pair of points
−z + 3y = 0 in the plane. [Hint: Every line has equation
has nontrivial solutions? Explain.
2x − 6y = 0 ax + by + c = 0, where a, b, and c are not all zero.]

b. Show that the converse to Theorem 1.3.1 is not b. Generalize and show that there is a plane ax+by+
true. That is, show that the existence of nontrivial cz + d = 0 through any three points in space.
solutions does not imply that there are more vari-
ables than equations. Exercise 1.3.10 The graph of

Exercise 1.3.7 In each case determine how many solu- a(x2 + y2 ) + bx + cy + d = 0

tions (and how many parameters) are possible for a ho-
is a circle if a 6= 0. Show that there is a circle through any
mogeneous system of four linear equations in six vari-
three points in the plane that are not all on a line.
ables with augmented matrix A. Assume that A has
nonzero entries. Give all possibilities. Exercise 1.3.11 Consider a homogeneous system of lin-
ear equations in n variables, and suppose that the aug-
a. Rank A = 2. b. Rank A = 1. mented matrix has rank r. Show that the system has non-
trivial solutions if and only if n > r.
c. A has a row of zeros.
Exercise 1.3.12 If a consistent (possibly nonhomoge-
d. The row-echelon form of A has a row of zeros.
neous) system of linear equations has more variables than
equations, prove that it has more than one solution.

1.4 An Application to Network Flow

There are many types of problems that concern a network of conductors along which some sort of flow
is observed. Examples of these include an irrigation network and a network of streets or freeways. There
are often points in the system at which a net flow either enters or leaves the system. The basic principle
behind the analysis of such systems is that the total flow into the system must equal the total flow out. In
fact, we apply this principle at every junction in the system.

Junction Rule
At each of the junctions in the network, the total flow into that junction must equal the total flow
out.

This requirement gives a linear equation relating the flows in conductors emanating from the junction.