Introduce nonnegative slack variables , and , and write the objective as . With nonbasic, the initial simplex dictionary is
This is feasible at the origin. Enter . The simplex ratio test selects to leave, since its bound is , compared with from . Pivoting gives
Now enter . The ratio test chooses to leave: its bound is smaller than from and from . The resulting dictionary is
All basic values are nonnegative and all nonbasic objective coefficients are nonpositive. The simplex algorithm therefore certifies
The negative objective coefficients also make this continuous optimum unique.
For integer , all three original slacks are integers. Apply a Gomory fractional cut to the row
The fractional part of is , so the cut is , or . To justify it directly, rewrite the row as
The parenthesized term is an integer and is at most , since the other term is nonnegative. It is therefore at most , proving the cut. This retains every feasible integer solution and excludes the current fractional optimum. Since , the cut is equivalently .
Introduce the integer-valued nonnegative slack . Its dictionary row is , so primal feasibility fails while the objective row remains dual feasible. In the dual simplex algorithm, leaves; tie in the dual ratio test, each with ratio . Choose to enter. Substituting gives
This dictionary is feasible at and certifies the cut relaxation's optimum , achieved at the integer point .
To find every optimal integer point, the objective row forces . Nonnegativity of all the remaining rows then reduces to . On this optimal edge, . Since must be integral, only its two endpoints are possible:

Articles by others on the same topic (0)

There are currently no matching articles.