diagnosis network functions in GeNie / SMILE

The engine.
Post Reply
Kai
Posts: 4
Joined: Thu May 22, 2008 6:05 pm

diagnosis network functions in GeNie / SMILE

Post by Kai »

Hi,

I’m trying to use the diagnosis network functions in GeNie / SMILE.
So I first mad a model in Genie with 4 Aberservation nodes, 1 Auxilary node and 4 Target nodes. Everything worked fine in the GeNIE Testing diagnosis Windows. After that, I tried to use SMILE for .NET with the same Network. I was able to read all kind of information about a nodes and also to instantiate an Observation, but I never get a FaultCount > 0 from the DiagNetwork atteched to my Network.

So I also builded the same Car example network as it is descriped in you doc. But I still have the same problem. And I tried your EngineFuelSystem.xdsl Example. Still FaultCount, UnperformedTestCount and the length of GetUnperformedObservations() are 0.

I use SMILE VC 9.0 RTM [12.5 MB] (March 25, 2008) and SMILE.NET Binaries [597 kB] (March 25, 2008).

I attached an image of my sample network.

Greetings,
Kai
Attachments
SampleNetwork.jpg
SampleNetwork.jpg (73.07 KiB) Viewed 8433 times
shooltz[BayesFusion]
Site Admin
Posts: 1417
Joined: Mon Nov 26, 2007 5:51 pm

Re: diagnosis network functions in GeNie / SMILE

Post by shooltz[BayesFusion] »

Can you post the actual network (not just the screenshot) here?
Kai
Posts: 4
Joined: Thu May 22, 2008 6:05 pm

Post by Kai »

Here it is.
Attachments
kai.xdsl
Network
(5.37 KiB) Downloaded 587 times
shooltz[BayesFusion]
Site Admin
Posts: 1417
Joined: Mon Nov 26, 2007 5:51 pm

Post by shooltz[BayesFusion] »

Can you remove the 'mandatory' attribute from Error_105 and check what happens then?
Kai
Posts: 4
Joined: Thu May 22, 2008 6:05 pm

Post by Kai »

It's working now! Thanks. Should it not work as well to have the mandatory attribut and than use diagNet.InstantiateObservation(NodeID, "ShowsError");

I Like to have a few obligatory observations?

Grettings, thx
Kai
shooltz[BayesFusion]
Site Admin
Posts: 1417
Joined: Mon Nov 26, 2007 5:51 pm

Post by shooltz[BayesFusion] »

You should set some evidence on observation nodes marked as mandatory before performing the diagnosis (with DSL_nodeValue::SetEvidence in C++ or with Network.SetEvidence in .NET).
Kai
Posts: 4
Joined: Thu May 22, 2008 6:05 pm

Post by Kai »

thank you a lot for the quick help. everything works fine for me now.

Greetings,
Kai
Post Reply