I'm not sure if you're being consistent in your

definition of P, the possible precepts, since I can't

figure out how you arrived at 15. (I agree with 39).

To make life easy, assume you have a single sensor

which can return one of three values - A, B, or C.

The possible precept combinations can therefore be

enumerated as:

A

B

C

AA

AB

AC

BA

BB

BC

...

AAA

AAB

AAC

ABA

...

I think you'll find (as the formula implies) there are

3 possible combinations after the first input is

received (t=1), there are 3^3 = 9 additonal

combinations when the second input is received (t=2),

and 3^3^3 additional combinations once the 3rd input

is received (t=3). Add them all together, and you can

see the agent would need a total of 39 table entries

to know what action to take based on it's precept

history.

Rob G.

--- kindlencotech <

fredkindl@...> wrote:

> To try and understand this equation I assumed the

> following:

> number of possible percepts P = 3

> lifetime of the agent (number of percepts it will

> receive) = 3

> I the get an answer for the summation of 39.

> However, if I manually

> determine the number of possible ways in which the

> percepts can be

> received I get 15 for the size of the lookup table.

> Can somebody tell

> me what I am not undestanding correctly?

>

> Fred Kindl

> fredkindl@...

>

>

>

>

>

>

__________________________________________________

Do You Yahoo!?

Tired of spam? Yahoo! Mail has the best spam protection around

http://mail.yahoo.com