Modelling and Analytics for Supply Chain Management

Professor Anupam Ghosh

Welcome to “Modelling and Analytics for Supply Chain Management”! We are into Week 3,
Module 3 of the program and our topic is „transportation modelling for supply chain
analytics‟. Now, we had dealt with two types of transportation modelling up to till now. What
are they? One is cost model, means we had some suppliers or warehouses and we had some
markets or factories; means from the supply point to the markets they have to, the products
have to reach from the supply point to the market or from the supply point to the factory or
from warehouse to the market the products have to reach.

So, basically it has a origin and a destination. And the cost from each origin to each
destination was given. Our objective was to minimize the total cost of transportation. What
problem happened in this process? The problem happened in this process was there are
multiple routes that are possible, multiple options that are possible. Some options were taken;
some options were not taken because my objective was to minimize cost. So, I took only
those options where my cost was, cost per unit of transportation multiplied by number of
units that was less.

But in that process what happened? Many routes were not used, basically were shut down.
But it might happen that the routes that my mathematical model recommended me, one of
these routes might not function after some time, may be some natural catastrophe, that route
is closed; may be the suppliers who are supplying at that particular node, those suppliers
become non-functional.

So, there are many reasons why, depending on only one route or two routes is not
recommended for an organization. So the organization might tell you that look, the minimum
cost model is fine but then we are willing to incur some extra amount, we are willing to incur
some extra amount but then I want more number of routes, more number of possible routes so
tomorrow if one route is shut down I can use another route.

So, in that case we had used the Min-max and Max-min problem. In that my cost increased a
bit but in that process 2-3 more routes were opened up. Now another benefit for this is,
sometimes some routes that are opened up, there may, a new markets may grow in those
routes also. So, it should not be taken as a cost model only. It should also be taken as an
opportunity model.

The next problem that we ended up with telling you was that like, there is a supplier at one
end of the city and there is a factory at the other end of the city and the factory operates in a
just-in-time environment; so supplier at one end, factory at the other end and the factory
operates in a just-in-time environment.

In a every hour the factory needs, let us say 1600 kgs of a product or tonnes of a product, let
us say tonnes, metric tonnes; 1600 tonnes of a product the factory needs every hour and the
factory operates in a just-in-time system. So ideally the travel time from these two end points,
that is supplier end point till the market end point; supplier end point till the market end point,
this travel time should be 1 hour or in another way in 1 hour they should be able to transport
1600 kgs of material.

So, this travel time should be 1 hour or in 1 hour they should be able to transport 1600 kgs.
So, what is the maximum possible vehicle movement that is the other way to say is – what is
the maximum possible vehicle movement in 1 hour time from origin to destination; that we
will have to find out? The maximum possible vehicle movement is 16 trucks. Each truck is
carrying 1000 kg of material so we are getting 16000 kg of raw material.

If 16 trucks can travel in 1 hour from origin to destination each truck carrying 100 units of
product so 1600 units of the product required in the factory so my demand is met. If however
16 trucks cannot travel in 1 hour from the origin to the destination point then I will have to
think of alternate modes of transport also because my factory is running in JIT and my
existing road systems or existing vehicular planning is not able to satisfy that 1600 kg
demand in 1 hour then I will have to plan a different transportation system.

So our job today is to find out what is the maximum number of vehicles that can ply from one
end of the city to another, within a given span of time. Now you will say that what is so
rocket science in that? There is a long road. Measure the length of the road. Measure the
speed of the vehicles and measure the carrying capacity. Measure the length of the truck.
Measure the gap between two vehicles. So basically passenger car unit, and then calculate.
But life is not that easy.
What happens is if it is a big truck it might not be able to go on a small route, on a small road,
narrow road. If it is a small vehicle it might be able to go anywhere. But small vehicle has
less carrying capacity. In some cases there are one-way restrictions on the roads. So you can
go but you cannot come back on the same road. So you will have to take a detour. Portions of
the road, you have to take a detour so that you can reach the destination.

In some cases during particular hours of the day, right particularly from morning till late
night big vehicles are not allowed. So this maximum flow, calculating the maximum flow is
not as simple as we might be thinking of it. It has so many other complications also. So today
we will learn how to find out the maximum number of vehicles that can pass through within a
given amount of time.

(Refer Slide Time: 07:40)

There is a JIT movement as we are mentioning from a supplier plant to the mother plant, the
required quantity per hour by the mother plant from the supplier plant is x. The distance
between the supplier plant and mother plant is covered by a network of roads. Total vehicle
handling capacity of all the roads in the network put together per hour is y. So basically
required quantity is x. Distance between the supplier plant and mother plant is covered. Total
vehicle handling capacity is y.
(Refer Slide Time: 08:13)

Within the network there are different categories of roads, heavy vehicles, medium, some less
commercial vehicles, et cetera. As such the vehicle handling capacity of the roads handling
different types of vehicles also differ; this all that we have mentioned already.

(Refer Slide Time: 08:28)

The objective is to find out the maximum number of vehicles that can move from the supply
point to the factory in an hour using the network of roads. If vehicles movement per hour into
load per vehicle less than equal to requirement of materials per hour, then alternate transport
modes have to be thought of.
(Refer Slide Time: 08:47)

Let us look at the mathematical problem and we have to find out the maximum number of
vehicles from origin road crossing 1 to destination road crossing 7. So origin is 1, road
crossing 1, destination road is road crossing 7 and you will have to find out the maximum
number of vehicles that can flow from the supplier to the manufacturer in an hour.

(Refer Slide Time: 09:12)

This is the problem. What it says is, road crossing 1 to road crossing 2, 5 vehicles can move
in an hour. Road crossing 1 to road crossing 3, 6 vehicles can move in an hour. Road crossing
1 to road crossing 4, 5 vehicles can move in an hour. In this way all the crossings, how many
vehicles can move from one point to another is given.
So basically what are crossings? These are basically traffic signals. From this traffic signal to
this traffic signal in 1 hour 5 vehicles can move. From this traffic signal to this traffic signal
in 1 hour, how many vehicles can move? 6 vehicles! From this traffic signal to this traffic
signal how many vehicles can move in 1 hour? 5. So this and these are my road crossings;
crossing 1, crossing 2, crossing 3.

So, this is what is meant by this 5, 6, 5, 2, 3, 2, etc. So road crossing 1, 2, 3 this is the vehicle
movements. So what it is saying is, from crossing 1 to crossing 2 in 1 hour, 5 vehicles can
move. From crossing 1 to crossing 3, may be in 1 hour, 6 vehicles are moving; then 5
vehicles are moving from the next signal point to the other. So this is basically the vehicle
movement from one point to other, right?

(Refer Slide Time: 10:57)

So, what to do? First step is to draw the diagram.

First step is to draw the diagram.

(Refer Slide Time: 11:13)

If you see, if you can map this, if you can map this and if you see the notations what we are
saying is let X1 vehicles, let this 5 be, 5 vehicles can move, let this distance be X1.
(Refer Slide Time: 11:38)

And this is the diagram for this. This is crossing 1, crossing 2, crossing 3, crossing 4, 5, 6 and
7. So, from crossing 1 to crossing 2, 5 cars can move, 5 trucks can move in 1 hour; crossing 2
to crossing 3, 2 trucks can move; 6, 5, 3, 4, etc. So, whatever was been given to you has been
converted into a network diagram. No rocket science, it is simple drawing. It is simply
drawing the problem. It is a simple drawing of the problem, simple drawing of the problem.

Now what we were saying is, we this was original drawing. We are saying we will draw
another line that links up crossing 7 to crossing 1. This was the original drawing; this was the
original drawing. We are linking up from crossing 7 to crossing 1, another imaginary road
that has no signals, no crossings in between, a direct road from 7 to 1. What is the idea? The
idea is, all the vehicles that are moving from point 1 to point 7 through so many traffic
signals, all the vehicles that are moving from point 1 to point 7 with so many traffic signals
and reaching point 7 has to come back.

So basically at 7, how many vehicles are waiting or garaged? 0. So, whatever is coming in to
7 has to move out. Similarly, at 1 also if you can look at it that way, whatever is coming back
to 1 has to go back to 7 because these are trucks which will take the full load to 7, come back
empty and again take the full load back. So, basically there is no storing at any of these
nodes. So, the vehicles that have gone, the vehicles that have gone are coming back.

Similarly, whatever is coming back has to go again. So, basically this 7 to 1 is the maximum
number of vehicles that can move in this route. So, if you can draw this 7 to 1, whatever
value of x comes, whatever value of x comes, this is the maximum number of vehicles that
can ply in this route; because 7 does not store anything. It is like morning all the vehicles are
moving from 1 to 7.

Let us take the public transportation system. Let us take the rail system. What happens? All
the vehicles go and at night, at another garage it rests. Then from the morning the first bus
takes, starts moving at 4 am or 5 am and then again comes here. Then again they go. Again
they come here. So, basically what you want to say is at night there is some resting. Here
what you are saying is there is 0 movement, 0 resting.

Whatever goes comes out immediately. So this arc, this 7 to 1, this 7 to 1 basically takes care
of all the vehicle movement that has happened through so many traffic signals. Now, if this
route, so all the vehicles movements, so this route has to be maximized. So maximize this 7
to 1 route vehicle flow. Maximize the vehicle flow from 7 to 1. Maximize the vehicle flow
from 7 to 1. This is our problem, nothing else. This is our problem.

(Refer Slide Time: 16:43)

This is what we mentioned. We introduce a dummy road. The diagram is given here in very
small figure here. We introduced a dummy road from node 7 to node 1. Since, no cars are
created at node 7, the entire number of cars flowing out of node 7 must consist of entire
number of cars that flowed into node 7, whatever cars are moving out must be equal to the
cars that have come into node 7.
Similarly, since no cars are consumed at node 1 the entire number of cars that enter node 1
through the node number 7 to 1 must consist of cars that originally flowed out of node 1. So,
whatever car has come in is equal to the cars that had already moved out. So this now has to
be maximized. So, we will see this is basically X15.

(Refer Slide Time: 18:00)

If you take this as X1, yeah; if you take this as X1, sorry, just take X3, what are these? The
maximum number of vehicles that should ply, X4, X5 you can write them on your own; X1,
X2, X4, I am just writing down whatever I have written down here, so up to X14 and this
route is X15 and our objective is to maximize this X15, our objective is to maximize this
X15, so this is my model.
(Refer Slide Time: 19:06)

And so this is the equation; max: X15. Now, here there is a question. What are the subjects,
subject to, rather the constraints; X1 plus X2 plus X3 minus X15 is equal to 0. What does that
mean? Remember X1, X2, X3 minus X15 is equal to 0. X1, X2, X3 minus X15 is equal to 0.
Let us go back.

(Refer Slide Time: 19:39)

What was the diagram? X1, X2, X3, X15, remember? What is happening here at node 1?
What are we saying? Whatever vehicles are coming in are going out. It is like a traffic signal
here at 2. This is at 1. At 2 what is happening? At node 2 or signal point 2 there is traffic
signal. Whatever vehicles are standing, that is coming in at the traffic signal; after the signal
becomes green, after the signal becomes green are moving out, agreed?

So, at node 1 what is happening? How many vehicles are moving out? So what routes? X1
vehicles are moving along this route. X2 vehicles are moving along this route and X3
vehicles are moving along this route and what is coming in? X15 vehicles are coming in.
What is coming in? X15 vehicles are coming in. So, X1 plus X2 plus X3 is equal to X15,
agreed? This is at node 1. X1 plus X2 plus X3 is equal to X15.

Take X15 another side, on the other side. X1 plus X2 plus X3 minus X15 is equal to 0. We
need not do equal to 0. We have done it just for modelling purpose, no rocket science. So at
each of the nodes whatever is coming in has to go out. Let us take X4. What is coming in? X3
is coming in. Look at the arrow. This was my X9. X9 is coming, what is going out? X10. So
at node 4 what is happening? X3 plus X9 is coming in, should equal to X10; X3, X9, X10.
So, at every node this will happen. Now let us go back.

(Refer Slide Time: 22:12)

Yep, so you see this is what we did. Our objective was to maximize X15, that is node 7 to 1
and X1 plus X2 plus X3 is equal to X15; minus X15 we have done to put... X4, X5 is equal to
X1 plus X6. That is two things coming in, two things going out. X10, X3 plus X7 is equal to
X10.

Let us go to node 4. At node 4 what has happened? Remember? X3, no there was X9 sorry,
some X3, X9 whatever, X3, X9 we have mentioned, this is 7; X3, X7 and what is going out?
X10. So, now what is the maximum capacity at node 1; that is X1? Maximum number of
vehicles that can go is 5. How many vehicles are we planning? It should be less than equal to
5. Let us go back.

(Refer Slide Time: 23:13)

If this is not understandable, let us go back and see, yep. X1 right, oh sorry, X1 right. What is
the value of X1? We do not know, right? What is the maximum handling capacity of this road
in 1 hour, handling capacity? 5. So what should X1 be, less than equal to 5? What is this one?
This was X3, remember? This was X2, right? X3, what is the handling capacity of this
particular road? 5.

So how many vehicles can move at the max? Less than equal to 5, max is 5. How many
vehicles should move, less than equal to 5? It cannot be greater than equal to. So, this is, so
these are the other constraints. Let us take this one, X6 to X7, what was the variable? X14, so
X14. What is the maximum possible? 7 units; so how many vehicles will move? Less than
equal to 7, X14; so X14 less than... So, in this way, all the constraints are to be mapped.
(Refer Slide Time: 24:36)

See last one, X14 is less than equal to 7. So, once you do that and once you use this particular
software or Excel or whatever, once you use any mathematical optimization software you
will get X15, you will get X15 sorry, you will get X15 is equal to 7. You will get X15 is
equal to 7.

(Refer Slide Time: 25:14)

This is my, X15 is equal to, sorry X15 will be equal to 14, solution. So, what does it mean?
14 equivalent size vehicles can pass origin to destination in 1 hour. If the load required to be
sent to the destination is more than 14 vehicles then an alternate mode of transport is to be
thought of. How did we...? So, 14 equivalent size vehicles can pass from the origin to the
destination in 1 hour. If the load required to be sent to the destination is more than 1, then an
alternate mode of transport is to be thought of. That is how basically you arrange your
vehicles, clear? Now you can, I will show you.

(Refer Slide Time: 26:08)

This is the result sheet from LIPS. X15 is equal to 14. That means this is the maximum
number of vehicles that can move. And it is saying 3 vehicles should move in X1. What was
the maximum capacity of X1, remember? X1 was 5. 5 vehicles can move. It is saying that 3
vehicles should move from signal 1 to signal 2. Maximum capacity is 5. It is saying 3
vehicles should move. In this way the number of vehicles that should move through these
signal points are all given to you. So this is the way by which you can solve this problem.
(Refer Slide Time: 27:02)

Now see this is the total number of vehicles that can move. You can use similar methods for;
say you are having lot of small vehicles. Then it can enter the city during the daytime also.
Large vehicles are only allowed after evening. So assume that you are having only small
vehicles. So you can use this type of modelling to understand the product mix that means
how many small vehicles you need and how many big vehicles you need. That also you can
model.

For example, let us take an example. What is the maximum value that we have got for X15,
that is 14? That means we need 14 vehicles, or rather 14, in terms of city planning, 14
passenger car units, you need 14 passenger car units. And assume 1 passenger car unit
consumes a space that is required by Tata Ace which is a small vehicle, and Tata Ace can
carry, assume 100 kgs of material. So how many kgs can be transferred in 1 hour? 1400 kgs!
If you are required to transport 1600 kgs so you have to look for alternate modes of transport.

Now assume your factory is in such a location that no alternate mode is available. So what
you can do? You know 1400 can be transported, another 200 is remaining, right. So, what
you do, and each one has a 100 kg capacity. So you need 2 more, but where will you fit 2
more? Not possible. So, instead have 2 larger vehicles of 200 kgs capacity each. So, then that
problem is taken care of. So, not only maximum flow but it can also, this model can be used
for managing the size of the vehicles that you are having in your portfolio, which size
vehicles to use.
Now see another issue; then you will tell me that, “Sir, why shall I use 14 vehicles, 14 small
vehicles. I will use 7 big vehicles and carry 200 kgs each.” Then what will happen? Your
products will reach early and there will be storing problem. Where will you store? Because
you have a JIT system; so you have a storing problem. And then what will happen? You will
ask the vehicles to wait till 8 am or 9 am, till your production schedule starts. So, now what is
happening? Your vehicles are waiting so vehicle waiting time is there. That is a cost. So your
transport cost is increasing.

So when you are operating in a JIT system you will have to be very-very careful about how
you are synchronizing your transportation schedules, vehicle schedules, delivery schedules
and that schedule is somewhere which now you are understanding, that schedule is
somewhere linked with the network of roads that your vehicles are taking, the width of the
road, whether national highway or state highway or local road which in an essence is
basically the vehicle handling capacity of that road per hour.

So, JIT system, in a nutshell what we want to say is JIT system is linked to the type of
vehicles that you will hire, your vehicle scheduling system as also your distance and the
vehicle handling capacity of the roads. So JIT system is to be linked to route planning and
vehicle handling capacity also. So, this is the maximum flow problem. In the next module
what we will do is we will take care of the minimum time or the minimum distance that needs
to be traveled from place A to place B given that multiple routes are available, that is the
shortest possible route. This we will take up in the next class. Thank you!