[Help] Decision Support Systems - Utility Functions

[Shovel]

Right, yes this is after some coursework help. My Decision Support Systems module at uni has set me some coursework, for which the questions are a step up from those we were lectured on. This is fine by me, but there has been the slight neglect on the part of the lecturer to recommend where we go about learning the extra we need. He recommended only one book for the module and that doesn't cover it.

Google isn't helping, so I've come here. Ha!
It seems to be very much stats-maths, so here goes:

The question centres around a situation where you are offered a 50-50 chance of either winning £400,000 or nothing, £200,000 or nothing or losing £100,000 or nothing.
It concerns 'how much would you sell that chance for (or in the case of the latter, 'how much would you pay to avoid the risk').

It offers two utility values - an upper pound-reference of U(£400,000) set to 160 , and a lower pound-reference amount U(-£100,000) set to -40.

Now, the process of working from those two initial value sets and making up 3 more intermediate sets is fine, I can do it. However what I'm unsure of is whether the values (160 and -40) have to be normalised in any way first. The original example in my lecture notes uses the values 1 and 0 for upper and lower respectively, and no prior note mentions converting from unnormalised values. However, I have no idea whether he picked 1 and 0 because they're correct, or because it's convenient and easier as a first time example.

As I say, the text book just ignores this whole sodding chapter so... well... kinda stuck

Question is - do I have to normalise U(£x) values, and if so: how?

Thanks very much

 

[Lazarus]

42.


:s

 

[Shovel]

Ta Laz... ahem

 

[dysfunction]

I remember these types of questions.

 

[TdC]

I feel for you Ben, I really do

Now's the time to drop out, tie Jo to the back of your Harley and ride off into the sunset to become a youth-hostel owner in Thailand matey!

 

[TriGGer]

The underlying question here tho is ......










"You want fries with that ? "

 

[fatbusinessman]

Now's the time to drop out, tie Jo to the back of your Harley and ride off into the sunset to become a youth-hostel owner in Thailand matey!

*mental image*

 

[Shovel]

Do they have good Internet connections in Thailand?


Right, that question is unfinished - solely on this normalisation issue - though I'm increasingly convinced that it is supposed to be normalised. I'll grab someone else on the course this afternoon and pick their brains on the subject though.


Hopefully a slightly easier Excel question (same coursework), but one that's no more obvious to me at this point...

The Regression Tool. I eventually found it, used it, all is good. However, I've been told to 'model the relationship between the Time and the Batch number as a multiplicative function, using the Regression Tool of Excel.'

Since there's no check box in Excel that says 'Multiplicative', I'm wondering what exactly it is. I spent an hour going through my A-Level stats notes to brush up on this business but this never come up -- or at the very least I never come across this terminology before. What is a Multiplicative function? Is is something specific and clever or is it something vague and non-descript that only serves to confuse me when all I'm meant to do is multiply them together or something?

Contrary to what posting 2 questions to you lot in as many days would imply, the coursework is actually going ok

 

[dave]

glad I don't do DSS