Simplex Starting From a Given BFS
Simplex Starting From a Given BFS
Many exam problems give a candidate vector $\bar x$ and ask what to do next.
Steps
- Convert the problem to standard form.
- Lift $\bar x$ by computing slack and surplus variables.
- Check that $\bar x$ is feasible.
- Identify a basis associated with $\bar x$.
- Compute reduced costs.
- If not optimal, choose an entering variable and perform the requested number of iterations.
Important
If the given point is not a BFS, the simplex method cannot start from it directly. You must either find an initial BFS or use a phase-one method. In this course, exam questions usually give a BFS or ask you to verify that it is not one.
Mini checklist
- Are there exactly $m$ basic columns?
- Is $B$ invertible?
- Is $B^{-1}b\ge0$?
- Are nonbasic variables zero?
- Are reduced costs nonnegative for minimization?
Exam checkpoint
For simplex questions in minimization form, negative reduced costs indicate possible improvement. Use $u=B^{-1}A_j$, apply the ratio test only to positive components of $u$, then update the basis.