The Primal-Dual Correspondence Table

Use the primal-dual correspondence table to systematically translate any primal LP (with mixed constraint types and mixed variable sign restrictions) into its dual LP.

Step 1 of 157%

Tutorial

The Correspondence Table for a Max Primal

Every primal LP has a corresponding dual LP. When the primal has mixed constraint types (,=,\leq, =, \geq) and mixed variable sign restrictions, the primal-dual correspondence table tells us, line by line, the form of the dual.

For a maximization primal with constraints indexed i=1,,mi = 1, \ldots, m and variables indexed j=1,,n,j = 1, \ldots, n, the correspondence is:

Primal (max)Dual (min)i-th constraint is biyi0i-th constraint is =biyi freei-th constraint is biyi0xj0j-th constraint is cjxj freej-th constraint is =cjxj0j-th constraint is cj\begin{array}{|c|c|}\hline \textbf{Primal (max)} & \textbf{Dual (min)} \\\hline i\text{-th constraint is } \leq b_i & y_i \geq 0 \\ i\text{-th constraint is } = b_i & y_i \text{ free} \\ i\text{-th constraint is } \geq b_i & y_i \leq 0 \\\hline x_j \geq 0 & j\text{-th constraint is } \geq c_j \\ x_j \text{ free} & j\text{-th constraint is } = c_j \\ x_j \leq 0 & j\text{-th constraint is } \leq c_j \\\hline\end{array}

The top half maps primal constraints to dual variables. The bottom half maps primal variables to dual constraints.

A helpful pattern: for a max primal in canonical form (all constraints ,\leq, all variables 0\geq 0), every dual variable is 0\geq 0 and every dual constraint is .\geq. Any deviation from canonical on the primal side relaxes (to free/equality) or reverses (to the opposite sign) the corresponding dual restriction.

navigate · Enter open · Esc close · ⌘K/Ctrl K toggle