• Practice
  • About 22 min
  • 5 practice questions

Consistency of linear systems by the rank method, and homogeneous systems

Use the Rouché-Capelli theorem to decide whether a system is consistent, find one- and two-parameter families of solutions, find conditions on parameters, and test homogeneous systems for non-trivial solutions.

If something looks wrong, please tell us.

By the end of this lesson you can

  • State the Rouché-Capelli theorem: AX = B is consistent if and only if ρ(A) = ρ([A | B]).
  • Decide between a unique solution, infinitely many solutions (a k-parameter family) and no solution.
  • Find the values of parameters such as λ and μ for each case.
  • Explain why a homogeneous system always has the trivial solution and when it has a non-trivial one.
  • Balance a chemical equation by solving a homogeneous system.

Matrix inversion and Cramer's rule fail when |A| = 0, and they say nothing about systems with more or fewer equations than unknowns. The rank method answers every case: it tells us whether solutions exist and, if they do, how many free choices they allow.

The Rouché-Capelli theorem

A system AX = B is consistent if and only if ρ(A) = ρ([A | B]).

— Rouché-Capelli theorem (stated without proof in the textbook)

Reduce [A | B] to row-echelon form; the same steps reduce A (just ignore the last column). Count the non-zero rows of each to get the two ranks. With n unknowns:

Deciding the type of solution (n unknowns)

  • ρ(A) = ρ([A | B]) = n Consistent
    • Unique solution
  • ρ(A) = ρ([A | B]) = n − k, k ≠ 0 Consistent
    • Infinitely many solutions
    • They form a k-parameter family
  • ρ(A) ≠ ρ([A | B]) Inconsistent
    • No solution
    • The echelon form has a row 0 = non-zero

Worked example 1: unique solution, more equations than unknowns

Test x + y = 3, 2x − y = 0, x + 2y = 5. Working: [A | B] = [1 1 | 3; 2 −1 | 0; 1 2 | 5]. R₂ → R₂ − 2R₁: [0 −3 | −6]; R₃ → R₃ − R₁: [0 1 | 2]. R₂ → (−1/3)R₂: [0 1 | 2]; R₃ → R₃ − R₂: [0 0 | 0]. So ρ(A) = ρ([A | B]) = 2 = number of unknowns: consistent with a unique solution. From y = 2 and x + y = 3, x = 1. The extra equation x + 2y = 5 is also satisfied: 1 + 4 = 5. Note that A here is not square, so inversion and Cramer could not be used.

Worked example 2: a one-parameter family

Test x + y + z = 2, x + 2y + 3z = 5, 2x + 3y + 4z = 7. Working: R₂ → R₂ − R₁: [0 1 2 | 3]; R₃ → R₃ − 2R₁: [0 1 2 | 3]; R₃ → R₃ − R₂: [0 0 0 | 0]. So ρ(A) = ρ([A | B]) = 2 < 3: consistent with infinitely many solutions, a 3 − 2 = 1 parameter family. Put z = t (any real number). Then y = 3 − 2t and x = 2 − y − z = 2 − 3 + 2t − t = t − 1. Solution: (x, y, z) = (t − 1, 3 − 2t, t), t ∈ R. Check in the second equation: (t − 1) + 2(3 − 2t) + 3t = t − 1 + 6 − 4t + 3t = 5.

Worked example 3: a two-parameter family and an inconsistent system

(i) x + 2y − z = 4, 2x + 4y − 2z = 8, −x − 2y + z = −4: every equation is a multiple of the first, so the echelon form has one non-zero row and ρ(A) = ρ([A | B]) = 1 = 3 − 2. Put y = s and z = t; then x = 4 − 2s + t, a two-parameter family.

(ii) x + y + z = 2, x + 2y + 3z = 5, 2x + 3y + 4z = 8: the same steps as example 2 give R₃ = [0 1 2 | 4] and then [0 0 0 | 1]. Now ρ(A) = 2 but ρ([A | B]) = 3. The last row says 0 = 1, so the system is inconsistent.

Worked example 4: conditions on parameters

For what values of λ and μ does x + y + z = 5, x + 2y + 2z = 8, x + 3y + λz = μ have (i) no solution, (ii) a unique solution, (iii) infinitely many solutions? Working: R₂ → R₂ − R₁: [0 1 1 | 3]; R₃ → R₃ − R₁: [0 2 λ − 1 | μ − 5]; R₃ → R₃ − 2R₂: [0 0 λ − 3 | μ − 11]. (i) If λ = 3 and μ ≠ 11, ρ(A) = 2 and ρ([A | B]) = 3: no solution. (ii) If λ ≠ 3 (any μ), ρ(A) = ρ([A | B]) = 3: unique solution. (iii) If λ = 3 and μ = 11, ρ(A) = ρ([A | B]) = 2 < 3: infinitely many solutions (one parameter).

Homogeneous systems

A homogeneous system AX = O is always satisfied by x₁ = x₂ = ... = xₙ = 0, the trivial solution, so it is always consistent (adding a zero column cannot raise the rank). The question is whether it has a non-trivial solution too.

  • If there are fewer equations than unknowns (m < n), ρ(A) < n, so there is always a non-trivial solution.
  • If m = n: when |A| ≠ 0, ρ(A) = n and the trivial solution is the only one; when |A| = 0, ρ(A) < n and there are non-trivial solutions.
  • If m > n, reduce and compare ρ(A) with n in the same way.

A homogeneous system of n equations in n unknowns has a non-trivial solution if and only if |A| = 0.

— Homogeneous systems

Worked example 5: trivial or non-trivial?

(i) x + y + z = 0, x − y + z = 0, x + y − z = 0: |A| = 1(1 − 1) − 1(−1 − 1) + 1(1 + 1) = 0 + 2 + 2 = 4 ≠ 0, so only the trivial solution x = y = z = 0.

(ii) x + 2y + 3z = 0, 2x + 3y + 4z = 0, 3x + 4y + 5z = 0: |A| = 1(15 − 16) − 2(10 − 12) + 3(8 − 9) = −1 + 4 − 3 = 0, so non-trivial solutions exist. R₂ → R₂ − 2R₁ gives [0 −1 −2], R₃ → R₃ − 3R₁ gives [0 −2 −4], and R₃ → R₃ − 2R₂ gives a zero row. From −y − 2z = 0, put z = t: y = −2t and x = −2y − 3z = 4t − 3t = t. Solution: (t, −2t, t), a one-parameter family. Check in the third equation: 3t − 8t + 5t = 0.

Worked example 6: balancing a chemical equation

Balance x₁C₃H₈ + x₂O₂ → x₃CO₂ + x₄H₂O (burning propane). Working: carbon gives 3x₁ = x₃; hydrogen gives 8x₁ = 2x₄, that is 4x₁ = x₄; oxygen gives 2x₂ = 2x₃ + x₄. This is a homogeneous system of 3 equations in 4 unknowns, with rank 3 < 4, so it has a one-parameter family of solutions. Put x₁ = t: x₃ = 3t, x₄ = 4t, and 2x₂ = 6t + 4t, so x₂ = 5t. The smallest positive integers come from t = 1: C₃H₈ + 5O₂ → 3CO₂ + 4H₂O. Check atoms: C 3 = 3; H 8 = 4 × 2; O 5 × 2 = 10 = 3 × 2 + 4.

Key takeaways: compare ρ(A) with ρ([A | B]); equal ranks mean consistent, and the solution is unique when the rank equals the number of unknowns, otherwise a family with (n − rank) parameters; unequal ranks mean no solution. A homogeneous square system has non-trivial solutions exactly when |A| = 0.

Check your understanding

  1. 1. A system has 3 unknowns and ρ(A) = ρ([A | B]) = 2. Which statement is correct?

    1. It has a unique solution.
    2. It is inconsistent.
    3. It has infinitely many solutions forming a one-parameter family.
    4. It has infinitely many solutions forming a two-parameter family.
    Show answer

    Answer: C. Equal ranks mean the system is consistent. The number of parameters is n − rank = 3 − 2 = 1, so there are infinitely many solutions in a one-parameter family.

  2. 2. For a system AX = B, ρ(A) = 2 and ρ([A | B]) = 3. What does this tell you?

    1. The system is inconsistent and has no solution.
    2. The system has a unique solution.
    3. The system has a two-parameter family of solutions.
    4. The system is homogeneous.
    Show answer

    Answer: A. By the Rouché-Capelli theorem a system is consistent only if ρ(A) = ρ([A | B]). Here 2 ≠ 3, so there is no solution: the echelon form contains a row reading 0 = a non-zero number.

  3. 3. A homogeneous system of 3 equations in 3 unknowns has |A| = 7. What are its solutions?

    1. Infinitely many
    2. None
    3. Exactly two
    4. Only the trivial solution x = y = z = 0
    Show answer

    Answer: D. A homogeneous system always has the trivial solution. Since |A| ≠ 0, ρ(A) = 3 = number of unknowns, so that solution is the only one.

  4. 4. For what value of λ does the system x + y = 0, 2x + λy = 0 have a non-trivial solution?

    1. 0
    2. 2
    3. −2
    4. 1
    Show answer

    Answer: B. A square homogeneous system has a non-trivial solution exactly when |A| = 0: |1 1; 2 λ| = λ − 2 = 0, so λ = 2. Then both equations say y = −x, and (1, −1) is a non-trivial solution.

  5. 5. The augmented matrix of a system in x, y, z reduces to [1 2 3 | 4; 0 1 −1 | 2; 0 0 k − 2 | k − 2]. For which k does the system have infinitely many solutions?

    1. k = 0
    2. k = 1
    3. k = 2
    4. Every k other than 2
    Show answer

    Answer: C. If k = 2 the last row is 0 0 0 | 0, so ρ(A) = ρ([A | B]) = 2 < 3: infinitely many solutions. If k ≠ 2 both ranks are 3, giving a unique solution (z = 1). There is no value giving no solution.

Privacy · Terms · Refunds · Delivery