Re: [rules-users] firing rules without modifying activations
Wow! This is quite an idea! Actually I AM using insertLogical() already but I never thought about using them for this particular problem. I actually think that your Idea is quite brilliant as it not only solves my initial problem but also introduces the possibility to split "detaction" and "action". As you already stated, it is absolutely true, that we will probably have a list of problem indications (detections) and a separate list of problem solutions (action) which can have more than just ono-to-one links. Thanks for the fast response and this great idea! I'm currently in the final stages of my PhD thesis, but once I'm done, I'll gladly put some information about our project on the webpage you mentioned. -- View this message in context: http://www.nabble.com/firing-rules-without-modifying-activations-tp19623537p19629221.html Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
[rules-users] firing rules without modifying activations
Hi,
I'm doing an expert system for a highly complex high energy physics detector
(CERN->ATLAS->Pixel Detector->Detector Control System).
In principle the detector is a complex machine with a rather hierarchical
structure.
The idea is to use a Drools based expert system to inform the operator of
the Detector about what to do in case of a problem. The system itself is
controlled by something that we call a "Finite State Machine" which is
reducing the complex operation of the detector to a number of simple
button-actions.
As the Finite State Machine does not suggest, what to do, if something goes
wrong, the expert system should do that.
In order to do so, the expert system shall display "solution advices" to the
Detector operator. As there can be many unrelated problems at the same time,
we use a tree-structure to show, where in the Detector the problem happened.
If a rule gets activated by a fact (in my case a process parameter of the
detector, e.g. a temperature), the corresponding node in the tree is being
highlighted by a dedicated icon. Once the facts change and the activation is
cancelled, we remove the highlighting icon, so that the tree only shows the
current activations (we don't need logging here, as this is done elsewhere).
This works absolutely fine for us, BUT:
We can't FIRE the activations, as this would immediately remove them from
the agenda, which in turn would remove them from our tree => no operator
would see them.
At first I thought, that NOT-FIRING those activations wouldn't hurt, as we
just wrote the "solution" on the RHS of the rule and parsed this RHS by
ourselves. This way, we displayed the solution belonging to a given
activation as long as it was still on the agenda.
You can see, that the main problem of this is that one CANNOT do things
like:
PrintSolution("Reason for problem is a over-temperature of: " + $temp + "°C
on Module: " + $moduleName);
In our case we could still parse the code, but would be unable to access the
local variables $temp and $moduleName (example) without actually firing the
rule.
Firing the rule on the other hand would cancel it's activation, rendering
our display tree useless.
I think the thing that would be most usefull for us would be a way to fire a
rule "silently", without notifying the agenda about it (so that the
activation would NOT be removed).
I'm not sure whether I was able to explain the problem sufficiently, but
maybe someone has an idea how to handle the situation.
Thanks very much!
--
View this message in context:
http://www.nabble.com/firing-rules-without-modifying-activations-tp19623537p19623537.html
Sent from the drools - user mailing list archive at Nabble.com.
___
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] activationCancelled() not being executed?
I don't think that I can add something of value to the discussion from this point on as my expertise concerning Drools is quite limited. I followed Edson's advice and used ruleflow-groups instead of agenda filters. This works like a charm and enables me to have my rules react as I want. >From my limited point of view, it would probably make sense to stick to what might be a new users expectation of an agenda filter: Drools Manual, 2.5.7.3. Agenda Filters wrote: > > "Filters are optional implementations of a the filter interface which are > used to allow/or deny an activation from FIRING." > After I read this section of the manual, I was under the impression, that the ONLY thing an agenda filter would do, was to prevent a rule from FIRING. This would be the case if the "fire an "activation cancelled/skipped" for the rules filtered out" fix would be implemented, as the rules filtered out would then be treated as any other rule, apart from the fact that they do not fire. The other suggested solution might have it's own benefits, that I can't judge at the moment. If the "not remove filtered out activations from the agenda" fix should be implemented, I think that the manual should be updated in order to inform people about the fact, that it is not only the FIRING that is prevented by the agenda filter. Edson Tirelli-3 wrote: > > " > As it is now, a call to fireAllRules() will always remove all the > activations from the agenda, but will simply skip over the rules that do > not > match the filter. Since I don't think the semantics for the agenda filters > were ever spec'd, we need to do it in order to decide if the "fix" for > this > is to fire an "activation cancelled/skipped" for the rules filtered out or > if the "fix" is to not remove filtered out activations from the agenda. > " > -- View this message in context: http://www.nabble.com/activationCancelled%28%29-not-being-executed--tp12387991p19609068.html Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] activationCancelled() not being executed?
After Edson helped me sort some things out, it came clear, that the actual problem is the use of agenda filters in my test. If one uses them, the "activation cancelled" events will not reach the listeners (I don't know whether the activations are cancelled or not). Disabling the agenda filters, the activation cancelled events will be propagated to the listeners. According to Edsons suggestion I filed a JIRA bug report that can be found at JBRULES-1775. I will try to avoid the problem by using agenda groups instead. After finding out, what caused the problem, it was possible to reduce the test code by quite a bit: http://www.nabble.com/file/p19598702/smaller_test.zip smaller_test.zip -- View this message in context: http://www.nabble.com/activationCancelled%28%29-not-being-executed--tp12387991p19598702.html Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] activationCancelled() not being executed?
Hi! It's been quite a while and we tried several other approaches but in the end it all comes back to the old problem: a) fact creates activation b) fact changes to a value that does NOT fulfill the LHS of the formerly activated rule c) activation does NOT get cancelled d) fact changes again to a value that DOES fulfill the LHS of the still activated rule e) fact creates ANOTHER activation (!) I tested this behaviour with the test code that can be found in the test.zip archive a few mails earlier in this thread. IT DOES NOT DEPEND on whether one uses property change support or update() . The only activation cancellations I'm able to get correspond to the fact, that update first retracts the fact from the working memory and the reinserts it afterwards (something the property change obviously does not). But even with the retraction-cancellations of the update(), the problem remains the same: I have ONE fact that changes over time. Let's say all 5 seconds. If that leads to a new activation every 5 seconds, I will quickly end up with an enormous amount of activations for basically the same thing. I would expect that if one gets and activation A for a value of -15°C (example from test.zip) and the temperature value changes to -16°C creating another activation B, that activation A would be cancelled (as its facts are no longer existing). I really need a way to limit the number of activations caused by the same fact to ONE. Sorry to bother you again with this. -- View this message in context: http://www.nabble.com/activationCancelled%28%29-not-being-executed--tp12387991p19597177.html Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
[rules-users] Initiating global in rule
Hi!
I know that this topic has been partially covered in this:
http://www.nabble.com/Initialize-Global-to13388685.html#a13388685
thread.
Nevertheless I think that the geiven answers are only partially satisfying.
Basically the main problem is still unsolved:
One has a global variable inside the DRL and wants to initialize it in a
rule (most likely the initial rule). Althoug several posts in the above
thread point out, that this should be easily possible doing this:
global String test
rule "Initial Rule"
when
then
> session.setGlobal("test", "Hello World");
> System.out.println(test);
will lead to infinite recursions of this initial rule without the code ever
reaching the println()!
I tried several variations of the above. At least of of the ones that have
been used in the aforementioned thread: using no-loop true, using
initialization from Java application using a function call, testing for null
pointer with eval(test == null) and so on.
I'm using Drools 4.0.3 and I have the strong impression that this is a bug,
as all the code examples mentioned in the other thread DO NOT SOLVE the
problem. The recursion still exists.
Thanks for any help in advance!
--
View this message in context:
http://www.nabble.com/Initiating-global-in-rule-tp17081556p17081556.html
Sent from the drools - user mailing list archive at Nabble.com.
___
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
[rules-users] Type-Casting in when-part of rule
Hi! I'm using pattern matching on Java objects in the when part of my rules. One example might be: CommonDataObject($name : name matches ".*STATUS_D.._B..\.fsm.*", type == Type.STRING, $stringValue : stringValue != "OK" && stringValue != "UNKNOWN") So you can see, that my Java object "CommonDataObject" has at least 3 members: name, type and stringValue. This works all fine. Unfortunately my object could also hold a Float or an Integer value. The current implementation uses additional "floatValue" and "integerValue" members in order to access them from the rules. I would prefer very much to remove my "stringValue", "floatValue" and "integerValue" members at all. As I already have a "type" member (which can be Type.STRING, Type.FLOAT or Type.INTEGER), I would like to just store one member called "value" which would then just be an Object. As the rules have access to the type-member, they should then be able to cast the value-member to the correct type: CommonDataObject($name : name matches ".*STATUS_D.._B..\.fsm.*", type == Type.STRING, $value : (String)value != "OK" && (String)value != "UNKNOWN") Unfortunately it seems, that a direct type-cast is not possible in the when part of a rule. Maybe someone knows a way to realize this functionality... Thanks in advance! -- View this message in context: http://www.nabble.com/Type-Casting-in-when-part-of-rule-tp16628142p16628142.html Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
[rules-users] Using the audit viewer in own panels
Hi! I really like the audit viewer that you get if you use the drools perspective in eclipse. It would be great if one could use the audit viewer and the possibility to open old log files within a java user application. Is there a possibility to integrate / use the audit viewer in my own java UI application? If so, how would I do that (as I don't find any fitting classes in the API-doc)? Thanks! -- View this message in context: http://www.nabble.com/Using-the-audit-viewer-in-own-panels-tp14250616p14250616.html Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
re[rules-users] trieving the RHS of a rule at run-time
Hi! I want to display all my current activations in a JTable object. On selection of one of the hereby created rows (each representing an activated rule by its name), I want to display additional info on this rule. It is no problem to retrieve the rule from the RuleBase, as I know the Package and Rule names. Unfortunately the toString()-output of the associated Consequence is not really what I want. What I really want is the RHS of the rule in analogy to the Rule.getLHS() method. Unfortunately there is no such method for the RHS. Is there a work-around for this? I would really like to avoid parsing the rule file just to get the RHS of a rule. Thanks! -- View this message in context: http://www.nabble.com/retrieving-the-RHS-of-a-rule-at-run-time-tf4653148.html#a13294400 Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] activationCancelled() not being executed?
Hi, I tested the now really self contained code on a colleagues machine. He isn't using Eclipse which has the nice side-effect of proving that it's not something with my IDE settings... We were able to reproduce the bug with the 6 files included in the attached .zip. http://www.nabble.com/file/p13253739/test.ZIP test.ZIP Short description of the files: Test.java: The main class including the main. CommonDataObject: My POJO used for storing the test fact. It automatically notifies the rule engine in case of a value change by the use of the firePropertyChange() method. Types.java: A very simple enumeration object used in CommonDataObject. XPS_AgendaFilter: Very simple agenda filter. Rules.drl: My two very simple rules. The agenda filter used in XPS_AgendaFilter ensures that only the first rule is fired. Rules.dsl: An even more simple dsl file. Temperatures are good if -10.0 < T < 40.0. What the test app does: It creates a temperature fact and sets its values to -15.0, 15.0, -25.0 and 25.0 with a little delay of 2 seconds in between. An activation is created for the first value. Nothing happens after the second value (15.0) comes in. -25.0 creates another activation, which is again not canceled by the 4th value modification. I might remember you that you'll need to change that one line in the code of Test.java that specifies the location of the rules. Thank you very much, Fermion -- View this message in context: http://www.nabble.com/activationCancelled%28%29-not-being-executed--tf4348054.html#a13253739 Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] activationCancelled() not being executed?
Oh, yes! Of course, stupid enough! That's the problem if one does not actually test something that is meant to be self contained on another machine... I'll first test it on the machine of a colleague of mine and then I'll post ALL the necessary files. Promised. Until then: of course it was not possible for you to see this (as the file was missing) but actually I'm (at least trying) to use automatically updating objects. Instead of calling an update() method, I referred to the manual and used a firePropertyChange(). But as you will see once I managed to get you all the files (shame on me), this is not the problem, as data IS updated... I'll post again, once it really is "self contained". Thanks for the fast help,... again! Fermion -- View this message in context: http://www.nabble.com/activationCancelled%28%29-not-being-executed--tf4348054.html#a13252596 Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] activationCancelled() not being executed?
Hi! Self-contained test is available. See last post. Thanks, Fermion Edson Tirelli-3 wrote: > >It is really difficult to answer that since your understanding seems > correct. Is it possible for you to provide a self contained test showing > the > problem? > >[]s >Edson > > -- View this message in context: http://www.nabble.com/activationCancelled%28%29-not-being-executed--tf4348054.html#a13249572 Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] activationCancelled() not being executed?
Hi again! Finally I realized that it would be best to do what I was asked for: supply some self-contained test code. Although it was a rather painful process, it was worth it, as the strange behavior is still present. I'm far from saying that this is a bug. On the contrary, I'm quit sure it's just a very stupid thing in my code. But at least it should be easier for someone else to have a look now... Just download the 3 following files and copy them into the same directory. As I was to lazy to include something more clever, I hardcoded this path, so that you'll need to modify the line String absoluteFilePath = "C:/Dokumente und Einstellungen/Henss/workspace/DeploymentTest/"; to point to your folder. If you use Eclipse and the Drools plugin, you should also be able to browse this directory for the created log files (of course after running the small app at least once) using the "Audit View". I hope this helps to narrow it down a bit. Thanks, fermion http://www.nabble.com/file/p13214347/Test.java Test.java http://www.nabble.com/file/p13214347/Rules.drl Rules.drl http://www.nabble.com/file/p13214347/Rules.dsl Rules.dsl -- View this message in context: http://www.nabble.com/activationCancelled%28%29-not-being-executed--tf4348054.html#a13214347 Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] activationCancelled() not being executed?
I did some major code-cleaning in the meantime and I feel a bit more comfortable now... Unfortunately this still didn't help me with my two problems: a) I still have the problem that my activations don't get canceled. b) I don't think that I'll be able to distill the code in order to get a self-contained project out of it. So instead I had a closer look at the "Manners" example and I found only two obvious differences from my code: 1. Manners uses a "StatefulSession" object, whereas I used a "WorkingMemory" instance. I changed my code to use a StatefulSession as well, but the strange behavior didn't disappear. 2. Manners uses the standard ClassLoader while I'm using an URLClassLoader in order to access my rule files (I plan to deploy my application as a single executable .jar file (created with "Fat Jar" eclipse plugin), but I want to have my rule files in a separate folder). Changing this didn't help either. So this basically means, that there is another difference that I've missed so far. What I HAVE seen is, that the fact IDs of my rules change (I think this is a good thing): If I insert a fact with a temperature value of (e.g.) "-15 °C", a temperature-out-of-range rule gets activates (not fired). In my little table an activation is added caused by the fact with the fact ID 21:34. Now I change the temperature of this fact back to something "in range" (e.g.) "20°C". Instead of the activation being canceled and therefore disappearing from my table, it just stays there. Now I change the temperature back to something that is out of range: "-30 °C". A NEW activation is inserted into my table with a fact ID of 21:38. At the same time, the fact ID of the first (still persistent) activation changes to 21:38 as well! I hope that this means that drools realizes that the fact value of the first value has changed. But if this is true, then drools also noticed that the value changed before, without cancelling the activation. To my knowledge this also proves that my implementation of the fact beans hashCode() and equals() methods is working correctly. There has to be an error in my reasoning. Thanks again, fermion -- View this message in context: http://www.nabble.com/activationCancelled%28%29-not-being-executed--tf4348054.html#a13211834 Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] activationCancelled() not being executed?
Hi, as you might have noticed, I'm trying to avoid supplying a self-contained test package. The problem with that is, that my application code has grown quite a bit and it will not be very easy to extract the code. Nevertheless I will provide such a test package, should I not succeed in finding the error myself very soon. As I have next to no experience in Java, I just want to rule out that it is no stupid Java-beginner mistake. Actually I'm quite sure that it is; otherwise the bug would most likely be known to someone in the community. Therefore I will first clean up my code a bit more. Thanks for the help anyway and I'll report back... fermion -- View this message in context: http://www.nabble.com/activationCancelled%28%29-not-being-executed--tf4348054.html#a13162482 Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] activationCancelled() not being executed?
It's been a while since I had time to investigate this problem. I followed your advice and hooked up the DebugAgendaEventListener to my workingMemory instance. This clearly shows, that my activations are created but never canceled. So at least it is consistent with what I'm observing in my application. But before I start even considering it being a bug (I'm quite sure it is not), maybe it's better to understand what is happening in my code: I think that I nailed my problem down to an issue with my fact IDs: If I have, for example, a real world value (lets say a temperature) that is connected to a java bean, my understanding is, that the fact ID of this value inside the workingMemory should not change, even if I change the real world value (temperature rises)(and thus the java beans value). In fact the change is propagated to the working memory (as I call "firePropertyChange" inside my bean), but a NEW rule is activated by the changed fact, having a DIFFERENT fact ID! I'm quite sure that this is the reason for my problem with the missing cancellations. I'd assume that for a rule to be canceled the fact with the CORRECT fact ID would have to be changed. Now, in my application, it seems that every modification creates a NEW fact inside the working memory (which is of course not, what I want). I had the hope, that calling the firePropertyChange from inside the beans "setters" would notify the working memory automatically about the correct fact-ID. I checked my "equals()" and "hashCode()" methods but the hash code is always constant and equals is correspondingly true. I also checked that I don't accidentally re-assert the fact on modification. So what I don't understand is WHY the working memory thinks that the fact ID has changed... See you, fermion -- View this message in context: http://www.nabble.com/activationCancelled%28%29-not-being-executed--tf4348054.html#a13158335 Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Basic questions about rule deployment
Thanks for the warning! I tested this with a path that contains spaces. For some reason I don't have any problems with the corresponding URLs. Nevertheless I tried to convert those URLs into URIs and THEN I see the problem of the spaces not being accepted. As the URLClassLoader takes URLs, this is not really an issue for me so far. Still I don't really understand what is going wrong, as I had the impression that URLs are a SUBSET of URIs. But this relationship seems not to be strict... The only thing I can say is that my app works with spaces inside the path without the need of the described URI-workaround. -- View this message in context: http://www.nabble.com/Basic-questions-about-rule-deployment-tf4601653.html#a13155173 Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Basic questions about rule deployment
Thank you very much! I followed your advice and created URLs for my .drl and .dsl files. Then I created a new URLClassLoader instance with these two URLs. Finally I assigned this class loader to my packager builder using: myPackageBuilder.getPackageBuilderConfiguration().setClassLoader(myURLClassLoader); Absolute path names seem to be Ok, as long as one uses "file:" as a prefix to the URL string. I plan to use a config file to specify the locations of my rule files (dsl and drl). The location of the config file itself will then be defined by a command line argument. Thanks again! -- View this message in context: http://www.nabble.com/Basic-questions-about-rule-deployment-tf4601653.html#a13152365 Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
[rules-users] Basic questions about rule deployment
Hi, I've finished my first small application using JBoss Rules. My current deployment "strategy" is that I want to have a single executable .jar file containing my application and an external (not included in the .jar) folder that contains the rules that should be used. Unfortunately, this seems to be harder, than I thought, as my application refuses to use .drl or .dsl files that are outside of the package. I'm quite sure that it's just my misunderstanding of the manual that leads to the problem, as this seems to be one of the most basic deployment approaches, as stated by the manual: " Chapter 7. Deployment and Testing 7.1. Deployment options Once you have rules integrated in your application (or ideally before) you will need to plan how to deploy rules along with your application. Typically rules are used to allow changes to application business logic without re-deploying the whole application. This means that the rules must be provided to the application as data, not as part of the application (eg embedded in the classpath). " Unfortunately I have quite some problems understanding the rest of the chapter. So I just used and modified on of the given examples. CODE/* //read in the source Reader source = new InputStreamReader(JBossRules.class.getResourceAsStream(KNOWLEDGE_FILE)); //read in the DSL Reader dsl = new InputStreamReader(JBossRules.class.getResourceAsStream(DOMAINSPECIFICLANGUAGE_FILE)); //Use package builder to build up a rule package. //An alternative lower level class called "DrlParser" can also be used... PackageBuilder builder = new PackageBuilder(); //Use the following instead of above if you are using a DSL: builder.addPackageFromDrl(source, dsl); */CODE If I use KNOWLEDGE_FILE = "/Rules_withDsl.drl"; DOMAINSPECIFICLANGUAGE_FILE = "/Rules_withDsl.dsl"; it all works wonderfully... but then the rules are within my .jar :-( So I tried to change this into a folder outside the package: KNOWLEDGE_FILE = "C:/test/Rules_withDsl.drl"; DOMAINSPECIFICLANGUAGE_FILE = "C:/test/Rules_withDsl.dsl"; Now, it doesn't work any more, because the "getResourceAsStream" returns null pointers... Any idea? Thanks -- View this message in context: http://www.nabble.com/Basic-questions-about-rule-deployment-tf4601653.html#a13138467 Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
[rules-users] activationCancelled() not being executed?
Hi! I'm using a JTable to display rules that have been activated. My goal is to have the table display a set of activations ordered by their salience. The user should then choose which one of them to fire (if you want to know why, be invited to read the background). Of course this makes only sense as long as the activations are valid. Rules whose activation has been cancelled should still be displayed, but in a different style (like greyed out). In order to accomplish this, I want to use an implementation of the AgendaEventListener-interface, implementing the activationCancelled()-method. The Listener itself works, as I use the activationCreated()-method to add the rule as a row to my table (which works fine). Unfortunately the activationCancelled()-method doesn't seem to be executed at all! I assume that I have a fundamental misunderstanding of what should trigger an activationCancelled() event? Background: (just if you're curious...) I'm working as a PhD student in physics at the CERN international high-energy physics laboratory, for the ATLAS experiment. As a member of the Detector Control System (DCS) group, I'm going to write an expert system in order to assist the shift crew with the detector operation. The operation of the detector is a difficult task, because more than 7 parameters (like voltages, currents and temperatures) have to be controlled. A Finite State Machine (FSM) computes a defined and limited number of states from the given parameters (like ON, OFF, STANDBY,...). I'm going to synchronize the expert system application to ERROR states of the FSM, in order to extract the relevant parameters from the system (as it would be impossible to feed all parameters into the XPS all the time). This is handled by quite some control rules that decide which data has to be requested, released and so on. Once all necessary data has been received, this might lead to activations of the (more interesting) user rules. Those rules have an explanatory character like "The temperature of this XYZ is too high, so please try ABC." For a given set of facts I assume more then one rule to be activated. Whilst the rule engine should take care of the execution (firing) of all control related rules, the execution of the user rules should be under direct user control. Displaying all activations in a GUI will allow the user to pick one (try it) and rate it (the proposed solution solved the problem "yes/no"). Changing the rules salience according to the answer, will allow the system to dynamically reflect the current system status. (Rules that are related to broken connections / wrongly connected cables are likely to be found at the beginning of the experiment but hopefully only on rare occasion later on.) In order to ensure that no information is lost during the absence of DCS experts (which will not always be around), a log file should be written for each major FSM-ERROR. At the moment the WorkingMemoryFileLogger seems to be predestined for this job. This way I hope to collect large amounts of ERROR data during the initial phase of the experiment that helps to extract rules for numerous error conditions. This will become especially handy, as such errors have a tendency to come back, even if the original detector experts are gone for years... -- View this message in context: http://www.nabble.com/activationCancelled%28%29-not-being-executed--tf4348054.html#a12387991 Sent from the drools - user mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
