Back to the on-screen lesson ·

Bases and basic feasible solutions

A corner written as a choice of columns: basic variables, the zeros that pin the point, and the sign test.

Paper packet. Every task here also exists on screen, where it is checked automatically; answers written on paper are not assessed by Nydus. When you are back at a device, enter your answers there.

1. What you will learn

By the end of this lesson you will be able to form the basic solution belonging to a chosen basis by zeroing the non-basic variables and solving the square system, write the basis matrix of a named basis, decide whether a basic solution is feasible from the signs of its values, and say why basic feasible solutions and corners of the feasible region are the same points.

2. What you already have

You can put a program into standard form, and you know from unit 2 that an optimum sits at a corner. This lesson gives the corner an algebraic description, so that an algorithm can hold one without drawing anything.

3. Basis, basic, feasible

A basis is a set of $m$ columns of the constraint matrix whose square matrix is invertible. The variables of those columns are basic; every other variable is non-basic and set to zero. Solving the remaining square system gives the basic solution, and it is a basic feasible solution when every value comes out non-negative.

4. A corner, written as a choice of columns

In standard form there are $m$ equations and $n > m$ variables, so the system is underdetermined and its solutions form a space. To pick a single point out of it, choose $m$ columns whose matrix $B$ is invertible, set the other $n - m$ variables to zero, and solve $Bx_B = b$. The result is a basic solution, and it is determined entirely by which columns were chosen.

Each variable set to zero is one constraint held with equality — a decision variable at zero means an axis, a slack at zero means its constraint is binding. Holding $n - m$ of them at once is what pins the point to a corner. Basic feasible solutions and corners of the feasible region are the same thing, seen algebraically and geometrically, and that identity is what lets an algorithm search a shape.

Basic does not imply feasible. The algebra will happily return a negative value, and that is a crossing of boundary lines outside the region.

Another way: steps

  1. Choose $m$ columns with a non-zero determinant.
  2. Set the other variables to zero.
  3. Solve the square system for the basic ones.
  4. Check the signs: non-negative means feasible.

Another way: example

$2x_1 + 5x_2 + s_1 = 100$, $3x_1 + 2x_2 + s_2 = 60$. The basis $\{s_1, s_2\}$ gives $x = 0$, $s_1 = 100$, $s_2 = 60$: the origin, feasible, and free of any work at all — which is why the slacks are where the method starts.

5. Where this usually goes wrong

Basic and feasible are run together, so a negative basic value looks like an arithmetic error rather than a point outside the region. The second habit is thinking a basis is a set of values; it is a set of columns, and the values follow from it with no freedom at all.

6. The same corner, two ways

  1. $x + y \le 6$, $x, y \ge 0$, standard form $x + y + s = 6$. Take the basis $\{x\}$.

    One equation, one basic variable.

  2. Then $y = 0$ and $s = 0$, so $x = 6$: the basic solution $(6, 0, 0)$.

    Algebra.

  3. $y = 0$ is the $x$ axis and $s = 0$ is the line $x + y = 6$. The point is where they cross — the corner $(6, 0)$.

    Geometry, same point.

7. Your turn: in $x + y + s = 6$, is the basis $\{y\}$ with $x = 8$ a basic solution?

  1. No. A basic solution sets every non-basic variable to zero, and $x$ is non-basic here, so $x$ must be $0$.

    The zeros are not optional.

  2. Your turn: work this step out. Its working is at the end of the packet.

    The basis $\{y\}$ gives $x = 0$, $s = 0$ and $y = 6$.

8. Guided practice

In standard form the program is $x + y + s = 11$ with $x, y, s \ge 0$: one equation, three variables. Fill in the basic solution for each choice of basic variable.

$x$$y$$s$
Basic variable $x$
Basic variable $y$
Basic variable $s$

9. Guided practice

The program $x + y \le 13$, $x, y \ge 0$ has standard form $x + y + s = 13$. Plot its basic feasible solutions in the $x$-$y$ plane.

Plot your answer on the grid:

24681012141618202468101214161820xy

10. Guided practice

Standard form gives $2x + 2y + s_1 = k$ and $7x + 7y + s_2 = m$. Write the basis matrix for the basis $\{x, s_2\}$, in that order.

This task has no paper form; do it on a device.

11. Practice

A choice of basis solves the equations and gives the basic variable in the second row the value $-2$. What is that solution?

12. Practice

A standard-form program has $5$ equations and $7$ variables. Complete the description of one of its basic solutions.

A basic solution sets z variables to zero and solves for the remaining b.

13. Somewhere new

A standard-form program has $2$ equations and $9$ variables. At most how many bases does it have?

Answer:

14. Lesson test

Lesson test: one question per skill, one attempt each, no hints. Your answers are checked when you submit.

15. Test question

In standard form the program is $x + y + s = 11$ with $x, y, s \ge 0$: one equation, three variables. Fill in the basic solution for each choice of basic variable.

$x$$y$$s$
Basic variable $x$
Basic variable $y$
Basic variable $s$

16. What you can do now

You can form a basic solution from a basis and decide whether it is feasible. Say in your own words why a variable set to zero corresponds to a constraint held with equality.

Working for the steps left to you

7. Your turn: in $x + y + s = 6$, is the basis $\{y\}$ with $x = 8$ a basic solution?, step 2