SPEAKER 0

And so I just wanna touch base with the integer programming so that, uh, you guys
can have a a feeling of because obviously we are not going to, uh, solve integer
programming models. But I, I just wanna give a, uh difference. OK, mention the
difference between linear programming and integer programming. So inte programming
as the name indicates OK, the variables, the decision variables, as I mentioned
last week, would take discrete integer values. The example of integer programming
problems would be if you are solving for number of persons, OK, if you are solving
for machines like shovels and trucks so we cannot divide persons into halves, we
cannot divide machines into halves Uh, similarly mining blocks If we if we don't
allow mining partial block mining OK, mining blocks would be integer just like we
implemented in our manual project. We we assume that each block is an integer
value. Right? We took We took the whole block. OK, we did not allow parcel block
mining. So you guys have already implemented when you do ultimate limit analysis
manually, OK, in in fact, we implemented integer uh integer solution, OK, because
each block, um, the value for each block was either equal to one or zero a block.
Is either mine or not mine. OK, so mining blocks take in T values as well. Um, an
integer programming problem could be all integer. Could be mixed integer and could
be 01. Uh, all integer means, uh, we have this problem both X one and X two. They
take integer values. So all the variables within this problem they take integer
values. It could be mixed integer as well. Some of the variables. They take real
numbers. OK, they can take any value between zero and infinity. The other one is
restricted to be an integer. OK, so this becomes an a mixed integer, uh, programme
or we could have 01 type problem. OK, so both are integers, but not only they are
integers, but they are binary. They can only take zero or one value. So So these
are the nature of our types of in these programming models. I'm going to, uh, show
three examples. OK, which which are kind of relevant to, uh uh to mining, uh, as
well. Um, So I'm going to actually give you examples with respect to mining OK, uh,
capital budgeting or project Selection, assignment, problem and size selection
problem. We can solve these problems through an integer programme. So capital
budgeting or project selection So as a mining company OK, As a mining company, we
have option to invest in six different projects. OK, but we have limited budget
limited capital available. So our capital is is is the resource that is available
and we wanna spend or invest that capital. Uh such that we are maximising value. We
are maximising return. We want to maximise return on our investment right? And each
one of these projects six projects requires a certain capital investment OK, and if
we invest into that project it will give us some return as well. So the return from
that project is known The capital requirement for a given project is known OK, so
given this information, we wanna write a mathematical programme which is going to
decide for us which project should be Invest in food ok out of these six options.
So whenever we define a mathematical model, what is the first thing we need to do?
We define the CN variable. So what is the CN variable in this case. So we have six
options. What are the decision? What is the decision? Variable With respect to
project one. So we either we either invest into a project or or we don't. Right? We
cannot have invest, right? We we can only either we invest or we don't invest.
Right? So it means the nature of the CM variable is binary, right? So that's why X
one is equal to one. If, uh, Project one is is, uh, selected or it will be zero. If
it is not selected, it will be zero. So the value for X one which represents
Project One, will be equal to one if Project One is selected otherwise its value
will be zero. OK, so similarly, we will have 66 variables, right? OK, so this is X
one. Similarly, we will have variable for each one of these each one of these
projects. So that's why I'm saying X I Because I is equal to 126 Ok, and then what
is the objective function? So we want to maximise return Maximise Z is equal to
objective function coefficient multiply by decision variable What is objective
function coefficient for project one. Do we see that? The return from Project One V
One. Right. So it means V one X one plus V two X two all the way up to V six X six.
Do we agree? OK, subject to there is only one constraint in this problem capital
budgeting. So there is limited capital available. So the maximum capital available
is, let's say C is the total capital available? OK, and we need to invest in these
six projects. So it means investment in two projects has to be less than or equal
to what is available capital. Do we agree so that if available capital is C capital
C, our investment into these projects has to be less than or equal to what is
available. So what is the nature of this inequality less than or greater than less
than so subject two. Less than or equal to this is the total capital available. And
if you invest into any of these projects C one X one plus C two X two all the way
up to C six X six. Do we agree now? If now when we solve this problem, the solution
will say X one equal to one X two equal to zero similar. Something like that. So if
X one is one, what will happen? Both of these will multiply, right? And we will be
able to calculate return. OK, Uh, if that's two is zero, what will happen? This
will diminish. OK, so same thing, same thing for constraint. OK, so this is very
unique kind of mathematical model with binary variables. OK, they can only take
variable and they can only take value equal to one or zero. OK, now these are other
possibilities. Let's try to understand them as well. Ok, um with respect to so So
this is one of the constraints. OK, but these are, let's say, additional
constraints as well. Ok, either Project One or Project Two either or so if you are
investing into Project One, we shouldn't be investing into project too, because we
are going to use. So so the team, The team that we need for Project one and Project
two it is same team, right? So if we are, if we are investing into Project One, we
cannot invest into Project two because our team is busy doing Project one. They
won't be able to do Project two so we can only invest. So this is an additional
constraint. We can either invest into Project One or project. How do I write that?
How do I mathematically write that this English into mathematics? I can invest
either in Project One, which is represented by X one or Project two, which is
represented by X two. How do I write that constraint? Because this is an additional
constraint, right? How do I write that? Huh? Yeah. The answer is there. The answer
is there. X one plus X two less than or equal to one. Let's evaluate this. X two X
one plus X two less than equal to one. If X one is one with this constraint in the
model wood X two will ever become one. No, because if it becomes one, it will be
violating the constraint. Right? So it will always be equal to zero. The other way
around is true as well. If this is one, this will be zero. How about equal to one?
Why not equal to one? Because he couldn't decide not to go basketball, which one?
Less than equal to equal to equal means. You have to invest in one of them. Do you
guys see this? Do you guys see this? OK, equal to means we have to invest in one of
them. Ok, One of them has to be equal to one. So that's another scenario, right?
OK, uh but less than equal to inequality gives us an additional flexibility because
both of them could be equal to zero as well. That gives us an additional
flexibility. X one and X two. Both of them could be zero. So this is, uh do you
guys see this? Ok, uh, when we are designing or, uh, designing the structure of our
problem, these are the things that, uh we may use. If Project one is selected, then
Project two should be selected. Otherwise none of these should be selected. If you
are selecting Project One, we should be selecting Project two as well. So this is a
condition. So X one minus X two equal to zero represents this. This scenario, if X
one is one X two has to be equal to one because otherwise we are violating the
constraint. If X two is one X one has to be equal to one otherwise we are violating
the constraint. OK? Do you guys see this? This one is, uh this one is ultimate
limit problem. The third one. Let's read through that. If Project X one is
selected, then project X two should be selected. However, if X two is selected, the
next 1 may or may not be selected. This is a limit problem. How is it ultimately
limiter problem? Hm? If we need to mine black four. What do we need to do? Mind
block 12 and three if you need to mind Block one. Do we need to mind Block four?
No. Do you guys see this? So how do we write this equation? X one minus. So in
order to implement ultimate limit problem mathematically. OK, in order to write
president's requirement precedence. This is pre constraint. OK, so in our do mine
um, Block four, we have to mind block one in order to mine block four. We have to
mine block two in order to mine block four, we have to mine block three. So if you
have millions of blogs in the blog model right, so we will have millions of
constraints. X four minus X one less than or equal to zero. So if X four is one OK,
In order to satisfy this constraint, X one will have to be equal to one but other
way around. If X one is one, X four doesn't have to be equal to one. Did you guys
see this? Can I simplify this constraint? Is there any way to simplify this So
rather than three constraint Because the more constraints you have, the bigger your
problem is I'm I'm trying to address computational complexity. Remember, we
discussed computational complexity.
Lecture number one. OK, mind planning problems are computationally complex. That
is uniqueness of mind planning problems. So this is one of the examples. So the
more constraints are going to add to computational complexity. How do I address
that? So there are ways to address computational complexity. Can I simplify these?
These three equ How Mario? Any suggestions? Add them. How do I add them? X four,
uh, three or three out of four. Minus less than a week or two. Yeah, so I can
convert three into one Constraints because it serves the same purpose. Did you guys
see this? OK, and then you can look at the last one as well. Ok, so capital
budgeting and then with capital budgeting we also looked at modelling strategies.
We, uh I call it modelling strategies. So through these examples, we learned
modelling strategies How to write constraints. OK, with integer programming. OK,
with 01 programming assignment problem assigned drivers to trucks. Ok, so each
driver has a special, uh, pro provide special efficiency if he or she is driving a
a particular truck. So we know their efficiency with respect to each truck. OK? And
we wanna assign them such that we are generating value out of this. OK, uh,
maximising value maximising efficiency overall efficiency of the operation. OK, so
we need to assign drivers to trucks or trucks to drivers. Beside that, we are
maximising efficiency. OK, so this is the assignment problem and each driver can be
assigned to one truck. So that's why uh, X 11 is equal to one If driver one is the
is signed two drug one, otherwise it is equal to zero. So if you have two trucks
and two drivers, let's say in assignment problem, the number of machines and number
of persons has to be same. So we have two trucks, two drivers licence. So for that
scenario, you How many variables do we have? Four variables. Ok, similarly, we have
X 12 equal to one If kur one is assigned to otherwise it is equal to zero and then
we have X 21 equal to one. Yes, sir Driver to is is designed to track one Ok,
otherwise zero and then final X 22 So this is very typical problem signing uh
machines to, uh machines to personnel. OK, so drivers to trucks in this case So we
wanna maximise efficiency so maximum maximise Z is equal to We know the efficiency
of each driver with the each truck which is represented by E So E 11 X 11 OK plus E
+12 X +12 plus E +21 X +21 plus E +22 X two To define the objective function. OK,
so two of these variables will be active. Two of these variables will be inactive,
subject to we have only two types of constraints. We have to assign drivers to one
truck and trucks assigned to one driver. Ok, uh so the constraints, Sir, if you If
you look at this constraint here X 11 plus X 12 is equal to one. What does this
mean? Driver one assigned to truck one driver one assigned to truck 21 of that
where these two variables is equal to one Because both of them cannot be equal to
one driver. One can be assigned either to track one or two Track two. Same thing
here. OK? X 11 plus X 21 equal to one. Uh, same thing here. Driver two can be
assigned either to truck one or two truck two and then truck one can be assigned to
one of the drivers. So these two constraints So Track One, which is second
subscript here. OK, Track one is either assigned to driver one or Track one is
either or it is assigned to driver two. Same thing here. So these are the two types
of constraints in assignment problem site selection. So we are a mining company,
OK? And we have, um, four mine sites. OK, and we are thinking of, um providing a
crushing facility. OK, we need a crushing facility for these four mine sites. OK,
and we want to minimise the number of crushing facilities we wanna provide as less
as possible so that these four mine sites should share these crushing facilities
OK, because ideally, we would have a crushing facility at each mine site. Right? So
if you have crushing facility on each mine site, then obviously each each one will
have their own. So but But we want to minimise that we want to minimise that
investment, OK, So that they are able to share uh among the mind sites so minimise
the number of crushing facilities. So obviously, if if if a mine site doesn't have
crushing facility, what do we need? We will have trucks moving from one mine site
to another mine site. OK, so they will deliver or to the crushing facility. OK, so
it means, uh uh based on the travel distance as well as the degree of difficulty
with respect to each heart load, Um, we know travel time from 11 side to another.
So this data is D one. OK, so you can see from mine side one to mine side one.
Travel time is zero. OK, because it is at the same location but from my side one to
my side. Two we will be travelling 16 minutes. OK, 20 minutes. 25 minutes. Uh, same
thing for others. But what we as a company? Our policy is that we are not going to
travel more than 20 minutes. We will not travel more than 20 minutes. We need to
build crashing. We need to provide rushing facility such that we are minimising
cost. Let's say, uh, providing this infrastructure. But OK, we are going to travel
only 20 minutes up to 20 minutes. So based on that, um the the the the the
variables would be X one equal to one K for yes, sir, Provided at my side one.
Otherwise it will be equal to zero, right? So if X one is with respect to mine side
one, it means we have a up to four of these variables, right? Because we have four
mine sites and what is the objective function? We want to minimise the number of
crushing facilities. Right? So what is the objective function? X one plus, sir X
two plus X three plus X four. Ok, and then there will be some constraints subject
to now with respect to mindsight one, there will be a constraint right now. We need
to. Our objective is to go less than or equal to 20 minutes. So if so, obviously
for mine. Site one. The most suitable location is mine. Site one. OK, but it can
go. We can go up to mine. Site two. We can go up to mine. Site three. Can we go to
mine? Site four from mine Site one if if you are with. Are we within this policy
requirement? No. So it means for mine. Site one. We can have a crushing facility at
mine. Site one or my side? Two or my side three. Right? How do we say this?
Mathematically, huh? Yeah. X one plus X one plus X two Plaza X three less than or
equal to or rather greater than or equal to. Because if we say less than or equal
to it, we it could become zero. There will be no crushing facility for my side. One
so greater than or equal to one. Yeah, if there is any other suggestion. Yes, let's
explore that. That is that was either or that was with respect to that Uh that
example can be used for, uh, project team building. OK, so let's say, um uh, we are
building, um, a team of people. OK, and we know that this person and this person
cannot work together, so we we won't We won't have them in one team. Right? OK, so
if there is a conflict between two persons, we can't have them in in one team.
Right? Or two persons work well together. So we will. We will have them in the
team, right? So so it can be used. That example can be used in this way as well.
Uh, project, um uh, project selection. But in this site selection. Uh, we can see
that for mine. Site one. We can have a facility at mine. Site one or mine side two
or mine side 31 of them. OK, At least one of them should have a crushing facility
for my side. One same thing will happen for, uh, two. So we cannot have anything at
mine. Site one for mine. Site two. We can have it at at mine. Site one. we cannot,
uh, we can have it at mine site three, but we cannot have it. Mine site four. So
that's why with respect to mine site two, we will have it at mine site two or mine
site three greater than or equal to one and the other two constraints are available
here. So if we solve this problem, we will be able to allocate crushing facility to
mine sites minimising the provision of crushing facilities. So site selection
problem can be solved by a binary integer programme as well. So it's a kind of an
an awareness of, uh, in teacher programming, especially binary Uh uh bi with bin
binary variables. And in mining, uh, we use in teacher programming quite a lot, OK,
in my planning, especially just like I gave example ultimate limit problem. We have
to have an integer variable which represents each block. If you are doing
production scheduling OK, we will have to have integer pro, uh, variables. OK,
because the block is either mild in period one or period two or period three over
period 15. OK, so depending on how long, uh, what is the what is the stain or the
or the planning horizon? OK, we will have variables for each one of those blocks,
so that's why kind of an awareness. So any questions, comments on integer
programming