IE 2100

Case Study (Phase 1)

Before answering the questions, we argue that in this rice movement case, there exists 2 major
internal inconsistencies:
Inconsistency 1:
Data from Figure 3. and Figure 4., i.e. from Distance Matrix and Transportation Cost Matrix,
don’t match each other accurately. For instance, the ratio between distance from 4 (Krishna) to
52
0.11
5 (Guntur) and distance from 4 (Krishna) and 6 (Kurnool), is 489 , however, the ration
212.7
0.29
between their respective transportation costs is 726.6 , which implies that the
assumption that transportation cost per ton from one location to another is proportional to the
distance between these two locations may not be true for every pair of locations. The
implication of this inconsistency is either we can just directly use the given cost matrix and
calculate the costs between the four near-Karimnagar points and Karimnagar by multiplying
their respective distances by Rs. 5.50 or we can resort to other sources to acquire the real non-
proportional road-transportation costs which is difficult in this case.

Inconsistency 2:
It is given that the cost per km.tonne by road is Rs. 5.50 and the total transportation cost to
meet Karimnagar demand of 4375 MT from Krishna by road is Rs. 229042.00, so the “road”
229042
52.35
transportation cost per ton is 4375 , which is significantly less than 425.3, the
“railway” transportation cost per ton from the same supply point, Krishna, to Warangal, a
demand point near Karimnagar. Furthermore, if we calculate the distance between Krishna and

This study source was downloaded by 100000840454458 from CourseHero.com on 03-07-2022 18:50:48 GMT -06:00

https://www.coursehero.com/file/45107139/IE-2100-Case-Study-Phase-1docx/
 Karimnagar using the Rs. 5.50 rate, the resulting distance is 9.52 km, this is highly not possible.
The implication of this inconsistency is that we have to adjust this cost to a level which is higher
than its corresponding railway cost.

To address these 2 inconsistencies, we propose 2 corresponding presumptions.

Presumption 2:
Rather than only 52.35 Rs per ton cost from Krishna to Karimnagar, we acquire a more reliable
number from a real-life scenario road transportation cost online calculator
(http://www.truckbhada.com/CalculateFreight), which is 1610 Rs per ton cost with total
distance 322 km from Krishna to Karimnagar, and the cost per ton per km is around 5, which is
approximate to Rs. 5.50, and significantly larger than the approximate railway cost 425.3. It
gives a more accurate optimal result than using the data given in the case.

Since we can solve the balanced disaggregated problem in the same manner, so in phase 1, we
just solve the balanced aggregated transshipment problem. Here are some details of
formulation.
Assign locations in the following way:
East Godavari, West Godavari, Krishna and Guntur as pure supply points,
Mancherial, Sanathnagar, Jangaon, and Nizambad as transshipment points,
Vishakapatnam, Anantpur, Hyderabad, Adilabad, Khammam, Karimnagar and Warangal as pure
demand points.
Then the constraints are given as below:
For a pure supply or demand point, the supply or demand can be directly from Figure 1.
Monthly Movement Plan of FCI in Andhra Pradesh, for a transshipment point, its supply or

This study source was downloaded by 100000840454458 from CourseHero.com on 03-07-2022 18:50:48 GMT -06:00

https://www.coursehero.com/file/45107139/IE-2100-Case-Study-Phase-1docx/
 demand equals to its current supply or demand plus the total transportation amount, which is
75250, and the constraints are all equations since the problem is balanced.
At last, the cost matrix is given by the following “10%” rule:
Since Mancherial is in district Adilabad, Sanathnagar is in district Hyderabad and Jangaon is in
district Warangal, we assign an extra 10% charge when transporting rice from pure supply
points plus Nizambad to these transshipment points or transporting rice from these
transshipment points to pure demand points plus Nizambad on the basis of the costs
corresponding to Adilabad, Sanathnagar and Jangaon, respectively, e.g., it takes Rs. 591 to
transport a ton of rice from East Godvari to Adilabad, so it will then take Rs. 591*(1+10%) =
650.1 to transport a ton of rice from East Godvari to Mancherial.
Presumption 3:
Costs between Mancherial and Adilabad, Sanathnagar and Hyderabad, Jangaon and Warangal
are assumed to be 0, cost between any 2 different transshipment points is inhibitively high, set
to be 10*8, cost from one transshipment point to itself is 0, cost from pure supply points other
than Krishna to Karimnagar is also inhibitibely high, 10*8.

With the complete cost matrix and constraints, we can formulate this balanced transshipment
problem and solve using simplex method to have the result as below:

This study source was downloaded by 100000840454458 from CourseHero.com on 03-07-2022 18:50:48 GMT -06:00

https://www.coursehero.com/file/45107139/IE-2100-Case-Study-Phase-1docx/
 The objective value is improved by about 20% on the current base of Rs. 51333625, obtained by
transporting in a different pattern than the current one. It is worth noting that this result is
based on the “10%” extra charge rule and all the assumptions with respect to transportation
cost in Presumption 3. In reality, the actual optimal value may vary based on whether we adopt
a “5%” rule or a “15%” or an“X%”rule, if one of them gives us a better approximation for the
real transportation cost between the 3 transshipment points and other locations. Also, the
assumptions in Presumption 3 can be loosen to achieve a better objective value if extra
information is provided.

Q1: Identify the challenges faced by FCI in developing a monthly food grain movement plan.
A1: The top challenge is the balance between the aggregate plan and each variety-based plan.
One can develop a successful aggregate plan but this optimal plan could be infeasible when
variety-based constraints are added. Inversely, if one has achieved all the variety-based
optimality, then the aggregate constraint is also satisfied, but the total cost could be higher than
only considering the aggregate plan. However, in this case, due to the fact that the cost is
indifferentiate with respect to rice varieties, so we conjecture that as long as the aggregate
constraints are satisfied, then all the variety requirement can be satisfied automatically since it’s
not only balanced in an aggregated level, but also in the variety level. The shortage of one
variety in on location could be supplied by another variety without causing infeasibility and
deterioration in the objective value. So, the above result is also feasible and optimal for all the
subproblems.

Q2: Formulate the FCI movement plan as a transportation plan: identify source, demand
locations, quantities, distances, origin, i.e., destination transportation cost, etc. (Assume that
Karimnagar would be supplied by road from Krishna.)

This study source was downloaded by 100000840454458 from CourseHero.com on 03-07-2022 18:50:48 GMT -06:00

https://www.coursehero.com/file/45107139/IE-2100-Case-Study-Phase-1docx/
 A2:

Since demand at Karimnagar is satisfied by Krishna by road, we can simply use the data
provided to formulate a balanced transportation problem. Adding the optimal value with the
cost by road from Presumption 2, we have the minimum cost of Rs. 43353713, which is about
7% larger than the global optimal solution.

Q3: What are the implications of three different rice varieties in the formulation in (2):
A3: As already answered in Q1, due to the special (somewhat weird, since there is no cost
difference in transporting different varieties) structure of this case, we can solve in terms of
aggregate level to simultaneously satisfy the variety requirements.

This study source was downloaded by 100000840454458 from CourseHero.com on 03-07-2022 18:50:48 GMT -06:00

https://www.coursehero.com/file/45107139/IE-2100-Case-Study-Phase-1docx/
Powered by TCPDF (www.tcpdf.org)