It's going to be all over

the you know you're going through
what you're up to it really

are too good to be used to do you

want to be one but

I guess to the police to know that there's
no proof I might be your whole life

and you're the only good

news to make those prone
to news you want to hear

whatever anecdotes he might
do you must be here with

me so.

Look I'm sure you know.

We are easy and

you think you're going to be my English
recall I see numbers are childhood.

You got it and I was in your list.

We were just that you know

interacting if I was younger should
be your mommy and why the focus.

They cannot be with you but we want to
thank you Rob is the for and rightly

so cheesy as a legal mild winter night for

me because I'm out forty three
percent of the work here right.

Winter Baseball fire country out there
that you know we're on our plans for

the console.

It wasn't my chain.

I grew up in the allocation of these
parts from near the far right.

So there are major part of the Windsor on

the way the machine has its hour and

the down towers are right
will focus on the target was.

When the sound of morning is
the largest from one another.

We're also going to account for

the largest portion of which is the call
as you see in our bridges above.

This is kind of your perspective
that we really don't market and

cost about three hundred
now numbers to order issue.

And because they are no hard and
most of them didn't want to

minimize their respective major channels
A G e Bay through their contracts

I mean the buyers one three years living
without knowledge over time rival me or

humming power goes by as I write this I
will model we will illustrate to you.

The pricing and

focus of our project within that supply
chain blue dots of represents the buyers.

The yellow doesn't answer to me the men

are actually phones on how
even that is a real demand.

It's not as admired how
many hours of their life or

so wanted three years later the man

fired by the right as we are here his base
and he then was ordering the allocation.

There's ours under the wires
only crime like this.

Now show the actual numbers are as well
you know I think that's kind of resume

one of three hundred man and
how they're a million of them in a market.

It's hard for G.E. to rethink how
many hours to order from their client

in order to meet and greet him and so

we'll come up with the bone strength we
can create the capital and the captain of

the six model was interested Manderley I
feel I think I was right for

the fires that minimize So
translation over the line of the man in

the area as opposed to the next scene of
the man or woman as well in what is coming

this week of the new venture my own
doesn't work out of the man is used as.

Again what is wrong with a smile and
never on our planet

outnumbered human part of when and
where and how many times to order.

I'm out of there are three things
eight point five billion dollars

Now they want to step before
getting the money are right.

We begin with determining
our national forecast for

wind power we analyze several factors
that we believe might back not

the man included fossil fuel production
prices and the cost of winter.

Well the cost of fossil fuel has
gone up over the past decade.

There is not a significant correlation
between higher wind power again.

And these rising fuel prices.

Rather the significant
factors are as follows The S.

and P. five hundred only the index of
stocks across United States represents all

to leave the U.S. economy and general
volatility in the U.S. when market G.D.P.

which shows general row over the U.S.
economy over the past decade and also down

for general growth in the entire industry
the production tax credit that you're one

step further toward our tax credit given
to wind farms out of the tempers ten years

of insulation and wind for this and whole
wind power capacity installed for a year.

The storm here compiled over the last
decade before with your question on

those factors and him of the following
forecast model the most significant

factor we discovered that it back in
demand was the P.T.C. which the production

tax credit expiration has a strong impact
on the overall demand for wind power

on New Year's over the production
times pretty expensive to expire.

There is a spike in demand order
to have these tax credits and

the years following expiration when
there is no longer reduction tax credits

wind power becomes the next amount of
coal compared to other fuel sources and

least of all over again we
also added all the rest

of after using the years create
this demand for wind power as.

In the forecast model.

This is because there has been huge
volatility in the wind power market over

the past several years in order to counter
these various social political changes

very very difficult to want to buy we
at this are risk factors the group our

show actually man for wind towers pretty
easy got he got nine in the green

bar restaurant he doesn't tend to two
thousand and twelve show or past demand.

You may notice the usually doesn't
well yeah five the demand is

because you're an expected production tax
credit expiration at the end of the year.

Therefore one would expect by
command ordered back to the trade

in a while is all well and good is now
how many towers are of your cost us

the crux of the issue is knowing
where in the US The demands possible.

We see here the issue of very believing
state wide demand is the essence of

California your scheme doesn't
buy the two thousand five out or

has three percent of the four Wind Power
Man and you've got the six in chance of

nine percent in two thousand and seven
balls up one percent disclosure wished

on statewide against wind power is not
necessarily left of the national demand.

Therefore we implemented the following
strategy to combat this variability

amongst a wide wind power plant.

We did you first you come up with
a national point forecast as we put out

earlier we then determine
an average between demand for

forking over the past decade and
use that as a statewide mean demand and

then order to account for the variability
amongst a white man for wind power.

It is very Billie and
gender instead of casting demand scenario.

This is a very important input into
our state casting opposition all

natural that we were going to more detail
as to how our stochastic opposition model.

You know provide for the value.

Thank you David.

Now we can tell you where you were
the towers solely to match your

model takes an alcohol losing such as
their supplier location supplier pasties

towers you get your
transportation costs and

the ordering costs now are hardly
uses that information and

help with an APU will come in post
which is where we have come here to our

group of the purpose of this project is
to minimize the expected cost of G.E.

So what our model does is it minimizes
the ordering loss in the present and

is expected transportation costs.

So our model looks to this
situation in two stages

ordering stage new transportation
the model makes the decision

of how many towers to order and
where to water them from in the present

using these present on track
off work off and hedges against

those demands in areas which we generated
previously which take place in the future.

Now you may want to do is actually
going to implement these models.

So we work closely with cheese end users

to develop a tool to solve
this optimization problem.

He is data is loaded
into Microsoft Excel and

then that data is sent to intruders
we've developed which runs.

L B S all which is an open source
software becomes a no cost to G.D.P.

The data is then I'll put no useful
back into the pool in the summary and

then the story of along with
the original data that created here.

This is our actual tool
that we've given to G..

So you can see a user would simply come in
late this time period in question you're

interested in in this case from first
quarter fourth quarter two thousand day

by clicking the read data well.

The data is loaded into the model and then
the user is presented with three options.

We've talked about the optimize You
can hear an awful long term off

the user going off the road and also
just to show their suppliers ready for

the moment money of their
work the system for the live.

Vision as well as regional analysis
issue where it would be useful

to add a new manufacturing facility in
this case by clicking the outline of fun.

You'll see a summary of the results for
your article through all three and that

information will be written for live you
feel free used in the marks of yourself

and now real will go into the case studies
that we perform follow their own law

it is illustrated however long this five

of the first case study takes
into account deterministic and

uses as an actual demand to do it.

I didn't get into the optimization and
the column in the middle shows

actual patient in two thousand and
eight in terms of little towers or

pieces and this whole shows given
gene actually demands a dozen eggs.

What solution a model would provide and
some of the interesting

differences are with poor examples
of biology Gene ordered six hundred

towers from the supplier two thousand
a Wal-Mart only order twenty and

that can be explained by the fact
that much less than predicted demand

in that region and to because it was just
like she had to ship out those towers

to another region with higher demand and
doing so in her millions of dollars

but someone knows the demand
in this case it was this and

summary of the potential savings that
Gene could have if they had you are out.

The Schmo will demand already you know

and this could be considered
the upper bound of potential savings

the demand is going to be
known which is unrealistic.

Of course and all of this is accounted for

by transportation because that
is where the cost variability of

the second case study only
the first it takes the uncertainty

into account due to the fact that
the contracts of the suppliers of maybe.

One to three years before the demand
actually curse and in this

case the demand for is the demand of the
areas that we generate using the state.

Fortunately for them and with very good
again as you discussed in the form get.

And each run of the market is
on the basis of hundreds of

statements never once again.

G.'s actual condition of the reference and
this is the solution that our market would

provide and once again you can see
that the supplier G M O has much less

power than the actual However in
this case the end of it is due

to the mathematics from the forecasting
model and not from actual demand.

So it makes sense that from these my
scenarios much more of the towers

are awarded than we knew that this is
a summary of the savings as you would have

if they could use our optimization
with a store cast to demand and

it is important to know that this
is the main to bloom and either for

G.E. you want to use or still have to do
math to generate a particular policy.

This is a more realistic estimate of

savings of five million job I
don't really think Mariel's And

once again this is contributing to buy
most of the trust with the case and

because it was a supplier and
consumers are much.

Your suppliers work in the mall and
the way you conduct in this is all

the using the existing all suppliers and
the one I want to take his or

her out of the model to see the end
happen on the objective function or

the total cost and obviously the supplier
that have had the highest reason for

cause may think now is
the most crucial supply and

you because you know that supplier K.
is by far the most and

once again supplier G.E. which we've
discussed before the demand is a lot

of this gets to those who are houses but
in terms of the region.

We take each of the fifty two
regions are more mobile and

consider which will be the best place
to install the use of wireless and

this is done while giving other costs
like boring cost control of cost and

we just consider the best
place to store and

use wired in terms of transportation
because the Oklahoma and

Texas are very highly ranked which
makes sense because there will wind up

locations and unfortunately in Georgia
as well when we're doing the loop

and the loop so
we are going to buy the easy way for

us by way of the red car they will be out
front with your policy tools out there and

we're just explain your jihad
use them all in school.

It's definitely a big illustrate eight
point five billion dollars We also know

that G.E. with the original Big Wheels you
integrate the first senior transition.

Congress is currently in transportation
resource an apartment and in the first

year stores an apartment does not have
access to transportation confit but

IMO you really get that and we also have
G.E. with nominally right more than one.

And marginal and
we also like easily spaniel.

Currently I want to
point off the low point.

I mean what I'm also going to you
Bill is in my book and me my favorite

among can also account for the supplier to
locations as well as be integrated with

the current database where you
update the data in the future.

I soak meat in there.

The moment.

Others are there certain
of later the machine and

also to me you can't use
your neck like a neck and

lose our presentation be in question which

I do not taken to calling you from cause
of our or there is no reason I see you.

But for a what's wrong.

So if you have this information you need
to use a very I believe the highest way or

do you know what is the leading
cause you call if you know

the surface of the ordering cos
the total whirring cost at for

example first supplier in China will look
at their ports locations the locations

where the hours get shipped to
sew up the total power cost for

that supplier China for example in
a in a port look ation is already

includes both the power cost and
the transportation cost to get it there.

So it is included in the purchase and

you all were

changing their number to

the patients zero from those
locations which are extremely

well

with their own all you know they would.

Almost right now on fire or

you're starting out you know you should

tell your wife we're

very good we're more work

to erode.