BoatBanter.com - View Single Post

HK

jim wrote:
John H wrote:
On Tue, 9 Sep 2008 11:17:13 -0700 (PDT), wrote:

On Sep 9, 1:55 pm, John H wrote:
One technique for multiple attribute decision making is to develop a
weighted matrix listing the criteria used for the decision compared
to the
objects being considered. We'll use three boats as an example, Boat
Alpha,
Boat Bravo, and Boat Charlie.

We must first identify the criteria which are important to us. For this
example, I've chosen: Looks, Speed, Cost, and Length

First we will determine the relative value of each criterion as
compared to
each other. I'll let Looks have a 1.0 value. Then I have to ask
myself, is
speed more or less important than looks. Let's say speed is half
again as
important as looks, then speed would get a value of 1.5. Each of the
criteria are so weighted. Cost may be twice as important as looks,
so it
would get weighted with a 2.0. Length isn't as important as looks,
so it
gets a 0.8

Now we have to weight each boat in terms of the criteria. We'll use a
number from 1 to 10 to weight each. Alpha is pretty ugly. Charlie is
gorgeous. Alpha is the fastest, but not quite as fast as I'd like.
Charlie
is a dog. Bravo is the least costly, with Alpha not far behind.
Charlie is
quite expensive. Length is easy. I want a long boat, and Bravo is the
longest. Alpha is the shortest.

Alpha Bravo Charlie
Looks 4 6 10
Speed 8 5 3
Cost 7 9 2
Length 3 8 5

Now I apply the weight of the criteria to the score of each boat,
and then
get the total for each boat.

Alpha Bravo Charlie
Looks (1.0) 4 6 10
Speed(1.5) 6 7.5 4.5
Cost (2.0) 14 18 4
Length(0.8) 2.4 6.4 4.0

Totals 26.4 37.9 22.5

And the winner is Charlie!

Of course, for an actual selection many more criteria would be
involved.
And, the technique can be much more specific by using pairwise
comparisons
for each of the criteria. I used a simple comparison of each
criterion to
speed, but it could be that comparing length to cost, length should
not be
less than half as important as cost. When done with a computer
program, the
results can be used to show 'consistency' (or lack thereof) in the
weight
of each criteria.

This technique was used, and may still be, as a way to evaluate
officers
for selection to promotion or command positions in the Army.
Lemme guess, you got promoted on looks versus speed, right?!!!

Hey, when you're investing many thousands, spending a couple hours doing
some analysis can pay off in a big way.

It does take time, but only because of the thinking involved.

Take a couple minutes to check your math.