Finding General Solution of a System- Linear Algebra Methods

What Is a General Solution in Linear Algebra?

A general solution represents all possible solutions to a system of linear equations. Not just one answer—you get a formula that describes every point satisfying every equation in the system.

That's the whole point. You're not hunting for a single intersection anymore. You're mapping the entire solution space.

Why Linear Algebra Methods Work Better

Traditional substitution and elimination break down when systems get messy. Three equations with five variables? Good luck with substitution. Linear algebra handles any size, any configuration.

You represent the system as Ax = b, where:

• A is the coefficient matrix
• x is the vector of unknowns
• b is the constants vector

Then you solve using matrix operations. Clean. Systematic. Scalable.

Method 1: Gaussian Elimination

This is your workhorse. Transform the augmented matrix into row echelon form, then back-substitute.

Step-by-Step Process

1. Write the augmented matrix [A|b]
2. Use row operations to get zeros below pivot positions
3. Continue until you reach row echelon form
4. Identify leading variables and free variables
5. Express leading variables in terms of free variables

The solution you get at the end is your general solution. Every free variable becomes a parameter.

Example

System:

x₁ + 2x₂ - x₃ = 4
3x₁ + 6x₂ + 2x₃ = 18
2x₁ + 4x₂ - 3x₃ = 5

Augmented matrix:

[1 2 -1 | 4]
[3 6 2 | 18]
[2 4 -3 | 5]

After row reduction, you'll identify free variables and express the solution set parametrically.

Method 2: Gauss-Jordan Elimination

Same idea, but you reduce all the way to reduced row echelon form (RREF). No back-substitution needed. The solution sits right there in the matrix.

RREF gives you the solution directly. Each row shows one variable equal to a combination of parameters. Faster, but more row operations.

Method 3: LU Decomposition

Break A into L (lower triangular) and U (upper triangular). Then solve Ly = b, then Ux = y.

This method shines when you have multiple right-hand sides or need to solve the same system repeatedly. You factor once, solve many times.

Method 4: Matrix Inverse

If A is invertible, x = A⁻¹b. Simple. Clean. But computing A⁻¹ costs O(n³) operations, same as Gaussian elimination, and you rarely need the full inverse.

Most problems don't require it. Stick with elimination unless you have a specific reason to invert.

Comparing the Methods

How to Find the General Solution: Getting Started

Here's your practical workflow:

Step 1: Set Up the Augmented Matrix

Write coefficients on the left, constants on the right. Check your arithmetic here—errors compound fast.

Step 2: Choose Your Elimination Strategy

Use Gaussian elimination for speed. Use Gauss-Jordan if you want the final matrix readable. Use LU decomposition only if you're solving the same A with different b vectors.

Step 3: Identify Pivot Positions

Each pivot gives you a leading variable. Variables without pivots are free variables. The number of free variables tells you the dimension of your solution space.

Step 4: Express the Solution

Write each leading variable as a function of the free variables. Add the free variables as parameters (usually t, s, u...).

Example general solution:

x₁ = 2 - 3t
x₂ = t
x₃ = 1 + 2t

Where t is any real number. That's your general solution.

What About Inconsistent Systems?

Sometimes no solution exists. Row reduction reveals this immediately—a row like [0 0 0 | 1] means trouble. Stop. No general solution exists.

The rank of A versus the rank of [A|b] tells you everything:

• Rank(A) = Rank([A|b]): solution exists
• Rank(A) = Rank([A|b]) = n: unique solution
• Rank(A) = Rank([A|b]) < n: infinite solutions (general solution)
• Rank(A) ≠ Rank([A|b]): no solution

Homogeneous Systems

When b = 0, you have a homogeneous system. There's always at least the trivial solution x = 0. The general solution gives you the null space of A.

Finding the general solution means finding all vectors perpendicular to the rows. Useful for eigenvalues, least squares, and more.

When Software Is the Better Choice

For systems larger than 4×4, do yourself a favor—use Python (NumPy), MATLAB, or Mathematica. Manual elimination on a 10×10 system will eat your afternoon.

NumPy example:

import numpy as np
np.linalg.solve(A, b)

That's it. Done. The computer handles the elimination.

The Bottom Line

Gaussian elimination covers 90% of what you need. Learn it properly. Master row operations until they're automatic. The concepts transfer directly to Gauss-Jordan and LU decomposition when you need them.

General solutions exist when you have more variables than independent equations. The free variables parameterize everything. That's the whole game.