Re: Time Series and Appointment Management - Hospitals

Orestes Mas (conomc@tsc.upc.es)
Fri, 8 Dec 1995 17:33:03 +0100


In article <199511251132.WAA22896@mail.mel.aone.net.au> Michael Stobie <MStobie@c031.aone.net.au>
writes:

>I'm a medical administrator in a public hospital in Melbourne, Australia. I
>am developing a fuzzy logic application to manage the appointment system for
>40,000 Outpatients. The problem is trying to keep the number of
>appointments per clinic, made anywhere up to 12 months in advance, in line
>with the expected number of doctors attending clinics, sometimes not known
>until a month in advance.

>Has anyone experience with trying to produce a fuzzy solution to a similar
>problem.

>Many thanks

>Michael Stobie

They are some research in the field of time-series prediction using fuzzy
inference systems. In particular, Jang and Sung have applied the ANFIS
architecture to prediction of time-series that present a chaotic behavior.
Your problem seems to be more simple, but you can look at their solution. The
ANFIS architecture is implemented in the MATLAB Fuzzy Logic Toolbox.

The exact reference for the article is

J.-S. R. Jang & C.-T. Sun, "Predicting chaotic time series with fuzzy if-then
rules", in Proc IEEE Int. Conf on Fuzzy Syst., San Francisco, Mar. 1993.

Hope this helps

Orestes Mas
conomc@tsc.upc.es