Linear systems and Gaussian elimination
Systems of linear equations appear everywhere, from balancing chemical reactions to fitting models to data. Gaussian elimination solves them systematically.
Linear systems
A linear system is a collection of equations like \(2x + y - z = 3\), where variables appear only to the first power, multiplied by constants. A solution assigns values to all variables satisfying every equation at once. Geometrically, each equation in two variables is a line, and solving means finding where the lines meet.
How many solutions
Every linear system has exactly one solution, infinitely many solutions, or none. Two lines in the plane may cross once, coincide (infinitely many), or be parallel (none). A system with at least one solution is consistent; one with none is inconsistent. There is no linear system with exactly two solutions.
Augmented matrices
A system is recorded compactly as an augmented matrix: the coefficients in rows, with the constants in a final column. The system \(x + 2y = 5\), \(3x - y = 1\) becomes the matrix with rows (1, 2 | 5) and (3, -1 | 1). Working with matrices avoids rewriting variable names and makes elimination mechanical.
Elementary row operations
Three row operations change a system without changing its solutions: swapping two rows, multiplying a row by a non-zero constant, and adding a multiple of one row to another. Each is reversible, so the new system is equivalent to the old one. Gaussian elimination uses them to simplify systematically.
Row echelon form
Gaussian elimination produces row echelon form: all zero rows are at the bottom, each row's first non-zero entry (its pivot) lies to the right of the pivot above it, and entries below each pivot are zero. The system can then be solved by back substitution, starting from the last equation. The method is named after Carl Friedrich Gauss, though Chinese mathematicians described it in the Nine Chapters on the Mathematical Art about two thousand years ago.
Reduced row echelon form
Continuing elimination upwards gives reduced row echelon form (RREF): every pivot equals 1 and is the only non-zero entry in its column. The RREF of a matrix is unique, and the solution can be read off directly. This full process is often called Gauss-Jordan elimination.
Detecting inconsistency
If elimination produces a row of the form (0 0 ... 0 | c) with c non-zero, the system says 0 = c, which is impossible, so it is inconsistent. This happens, for example, when equations describe parallel lines or planes that never meet.
Worked example
Solve \(x + y + z = 6\), \(2x + 3y + z = 11\), \(x - y + 2z = 5\). Subtract 2 times row 1 from row 2: \(y - z = -1\). Subtract row 1 from row 3: \(-2y + z = -1\). Add 2 times the new row 2 to row 3: \(-z = -3\), so \(z = 3\). Then \(y = -1 + 3 = 2\) and \(x = 6 - 2 - 3 = 1\). Check in the third equation: \(1 - 2 + 6 = 5\).
Key takeaways
Linear systems have exactly one, infinitely many, or no solutions; consistent systems have at least one. Augmented matrices record coefficients and constants. Row swaps, scaling by non-zero constants, and adding multiples of rows preserve solutions. Gaussian elimination reaches row echelon form for back substitution; Gauss-Jordan elimination reaches the unique reduced row echelon form. A row reading 0 = c with c non-zero signals inconsistency. The method is ancient, appearing in Chinese mathematics about two thousand years ago.
Flashcard
Variables to the first power with constant coefficients
Check your understanding
Elimination on a system produces the row (0 0 0 | 4). What does this mean?
