Linear Algebra › Section A · Linear equations in linear algebra › Topic A.5
Solution sets of linear systems
When a system has infinitely many solutions, listing them is impossible, but describing them is easy. Every solution set is a point, line, plane or higher flat piece, and vector notation captures it in one short formula.
Describing infinitely many solutions
Row reduction tells you when a system has infinitely many solutions, but a list such as \(x_1 = 2 + x_3\), \(x_2 = 3 - 2x_3\) hides the geometry. This topic writes solution sets as vectors, which shows at a glance that they are lines, planes or their higher-dimensional cousins. It also reveals a striking structure: every consistent system \(A\mathbf{x} = \mathbf{b}\) has a solution set that is a shifted copy of the solution set of \(A\mathbf{x} = \mathbf{0}\). Applications such as balancing chemical equations, in the next topic, rely on exactly this description.
Homogeneous systems
A system is homogeneous when it can be written \(A\mathbf{x} = \mathbf{0}\), with every right-hand side zero. Such a system is always consistent, because \(\mathbf{x} = \mathbf{0}\) is a solution, called the trivial solution. The real question is whether a non-trivial solution exists, meaning a non-zero vector \(\mathbf{x}\) with \(A\mathbf{x} = \mathbf{0}\). The answer comes from the existence and uniqueness theorem: the homogeneous equation has a non-trivial solution exactly when it has at least one free variable. A non-trivial solution may contain some zero entries, as long as not all of them are zero.
More unknowns than equations
If a homogeneous system has more variables than equations, it must have a non-trivial solution. The reason is a count: each pivot uses up a different row, so an \(m \times n\) matrix has at most \(m\) pivots, and when \(n > m\) at least \(n - m\) columns are left without one. Those columns give free variables. The same count does not guarantee anything for \(A\mathbf{x} = \mathbf{b}\) with \(\mathbf{b} \ne \mathbf{0}\): such a system can still be inconsistent, and the free variables only matter once a solution is known to exist.
Worked example: a line of solutions
Solve \(A\mathbf{x} = \mathbf{0}\), where \(A\) has rows \((1, 3, -1)\), \((2, 5, -4)\) and \((3, 8, -5)\). 1. Row 2 minus 2 times row 1, and row 3 minus 3 times row 1, both give \((0, -1, -2)\). 2. Row 3 minus row 2 gives a zero row, so \(x_3\) is free. 3. Scaling row 2 by \(-1\) gives \((0, 1, 2)\), and row 1 minus 3 times that gives \((1, 0, -7)\). 4. So \(x_1 = 7x_3\) and \(x_2 = -2x_3\), and
5. Check row 1: \(7 - 6 - 1 = 0\). The solution set is \(\mathrm{Span}\{(7, -2, 1)\}\), a line through the origin.
Parametric vector form
Writing \(\mathbf{x}\) as a combination of fixed vectors, with the free variables as weights, is called parametric vector form. Take the single equation \(2x_1 - 4x_2 + 6x_3 = 0\). Here \(x_1 = 2x_2 - 3x_3\), with \(x_2\) and \(x_3\) free, so
The solution set is the span of \((2, 1, 0)\) and \((-3, 0, 1)\), a plane through the origin. In general, the solution set of \(A\mathbf{x} = \mathbf{0}\) is the span of one vector for each free variable, and it is just \(\{\mathbf{0}\}\) when there are none.
Non-homogeneous systems
Suppose \(A\mathbf{x} = \mathbf{b}\) is consistent and \(\mathbf{p}\) is any one solution, called a particular solution. Then the solution set of \(A\mathbf{x} = \mathbf{b}\) is the set of all vectors
where \(\mathbf{v}_h\) runs through every solution of the homogeneous equation \(A\mathbf{x} = \mathbf{0}\). Linearity explains why. If \(A\mathbf{v}_h = \mathbf{0}\), then \(A(\mathbf{p} + \mathbf{v}_h) = \mathbf{b} + \mathbf{0} = \mathbf{b}\). Conversely, if \(A\mathbf{w} = \mathbf{b}\), then \(A(\mathbf{w} - \mathbf{p}) = \mathbf{0}\), so \(\mathbf{w} - \mathbf{p}\) is a homogeneous solution. Geometrically, the solution set is the homogeneous solution set translated by \(\mathbf{p}\): a line or plane parallel to it, usually missing the origin.
Worked example: a shifted line
Solve \(A\mathbf{x} = \mathbf{b}\) for the same \(A\) and \(\mathbf{b} = (3, 3, 6)\). 1. The same row operations on \([A \ \mathbf{b}]\) give the reduced rows \((1, 0, -7, -6)\), \((0, 1, 2, 3)\) and a zero row, so the system is consistent. 2. Then \(x_1 = -6 + 7x_3\) and \(x_2 = 3 - 2x_3\), with \(x_3\) free. 3. In vector form,
4. Check \(\mathbf{p} = (-6, 3, 0)\) in row 1: \(-6 + 9 - 0 = 3\). The solution set is the line of the previous example, shifted to pass through \(\mathbf{p}\). Taking \(x_3 = 1\) gives \((1, 1, 1)\), another particular solution.
A routine for any consistent system
1. Row reduce the augmented matrix to reduced echelon form. 2. Write each basic variable in terms of any free variables. 3. Write a typical solution \(\mathbf{x}\) as a vector whose entries depend on the free variables. 4. Split \(\mathbf{x}\) into a constant vector plus one vector for each free variable, with that free variable as its weight. The constant vector is a particular solution, and the rest is the general solution of the homogeneous equation. The number of free variables is the number of columns of \(A\) minus the number of pivots, a count the topic Dimension and rank develops further.
Common mistakes to avoid
First, saying a homogeneous system might have no solution; \(\mathbf{x} = \mathbf{0}\) always works. Second, thinking a non-trivial solution must have every entry non-zero. Third, applying the more-unknowns-than-equations rule to \(A\mathbf{x} = \mathbf{b}\) with \(\mathbf{b} \ne \mathbf{0}\); it guarantees solutions only for homogeneous systems. Fourth, forgetting the particular solution and giving only \(x_3(7, -2, 1)\) as the answer to \(A\mathbf{x} = \mathbf{b}\). Fifth, assuming the solution set of \(A\mathbf{x} = \mathbf{b}\) passes through the origin; it does so only when \(\mathbf{b} = \mathbf{0}\).
Key takeaways
A homogeneous system Ax = 0 always has the trivial solution, and has a non-trivial one exactly when there is a free variable, which is guaranteed when there are more unknowns than equations. Parametric vector form writes the solutions as a span, with one vector for each free variable. If Ax = b is consistent with particular solution p, every solution is p + v_h, where v_h solves Ax = 0, so the solution set is the homogeneous one translated by p. The number of free variables is the number of columns minus the number of pivots.
Flashcards
Flashcard
A system of the form Ax = 0
Test your understanding
Question 1 of 5
A homogeneous system has 3 equations in 5 unknowns. What can you conclude?
Question 2 of 5
The reduced echelon form of Ax = 0 gives x₁ = 4x₃, x₂ = −x₃, with x₃ free. What is the solution set in parametric vector form?
Question 3 of 5
p is one solution of Ax = b, and the solutions of Ax = 0 are all multiples of v. Which describes every solution of Ax = b?
Question 4 of 5
Which statement about homogeneous systems Ax = 0 is true?
Question 5 of 5
Ax = b is consistent and A has 4 columns and 2 pivots. How many free variables does its general solution have?
