Checking for optimality
So far we have only looked at ways of obtaining an initial basic feasible solution to the
balanced transportation problem. We now develop a method for checking whether the
current basic feasible solution is optimal. For illustrative purposes, we will start with the
initial basic feasible solution that was provided by the North-West Corner method. Usually,
initial basic feasible solutions obtained by the Least-Cost method (or other methods
given in many text-books, such as Vogel’s method) will give better starting configurations.
Suppose that the cost cij of transporting 1 unit from source i to destination j is made up
of a dispatch cost λi and a reception cost µj so that
λi + µj = cij
whenever xij is a basic variable.
Remarks
• The total number of λi and µj variables is n+m. However, there are only n+m−1
basic variables. Thus, we are free to choose one of the λi
’s or µj
’s arbitrarily. It is
usual to set λ1 = 0.
• These “costs” can take negative values if required.
Considering only these dispatch and reception costs, it would cost λi + µj to send 1 unit
from source i to destination j. For (i, j) not corresponding to a basic variable, it will
often be the case that λi + µj 6= cij . In particular, if λi + µj > cij for a particular (i, j)
not corresponding to a basic variable, then there would be a benefit from sending more
goods that way.
So let sij = cij − λi − µj
. The sij values are entered in the top right of the cells. Then
sij is the change in cost due to allocating 1 extra unit to cell (i, j) (in fact it is a shadow
price). If any sij is negative (so that λi + µj > cij ), then the total cost can be reduced
by allocating as many units as possible to cell (i, j). However, if all the sij are positive
then it will be more expensive to change any of the allocations and so we have found a
minimum cost.
Thus the procedure is as follows:
1. Assign values of λi and µj to the columns.
2. Enter the values sij = cij − λi − µj
in every cell.
3. If all the sij ’s are non-negative, we have an optimal solution.
Assigning values of λi and µj to our example with the initial basic feasible solution given
by the North-West Corner method, gives the following transportation tableau: