Define the Model
Transportation Model Definition Dialog
After clicking the transportation menu item, a dialog is presented with inputs for the number of suppliers and demanders. Either input can range from 3 to 10. A checkbox indicates whether random data is to be supplied or the data form left with 0 entries. The former case is useful for demonstrations or practice with the transportation algorithm. The latter is appropriate if the student has specific data to enter.
The solution options provide different amounts of information regarding the algorithm and different levels of interaction with the student. As the algorithm progresses it is possible to shift between the options.
The Example Problem
The figure below shows the transportation data form constructed by the add-in. Data has been entered for the example of this section. The yellow region in column B holds information necessary for the program. This information should not be changed by the student. Buttons control the progress. Click on the Start button when the data is prepared.
The transportation algorithm requires a balanced problem in which the total supply equals the total demand. Our example does not satisfy this requirement in that the supply exceeds the demand. The program presents the dialog below prior to addressing the situation.
With a response of OK, the add-in creates a dummy demander that has a demand equal to the excess supply. Cell costs are zero in the new column. Flow assigned to the dummy column represents supply that is not shipped. The new model is shown below.
The model is now complete. The program constructs a new matrix with cells provided for a variety of computed quantities necessary to solve the problem. The remainder of the pages in this section describe the primal simplex applied to the transportation problem. That algorithm is summarized below.
The Simplex Algorithm
In the following we show the formal simplex algorithm for the transportation problem. We also illustrate the steps of the algorithm using the Teaching Transportation add-in. Click on a link to see the article.
Research Models and Methods
by Paul A. Jensen and Jon Bard, University of Texas, Copyright by the Authors