Re: [rules-users] drools planner planning entity
If you mean more then 1 planning entity class: Yes, you can, but the construction heuristics in 5.4.0.Final cannot handle more than 1 class currently. 5.5.0.Beta1's "unified selectors" will fix that. If you mean more then 1 planning entity instance: Of course, real problems can have 50 000 entities and more. Op 18-06-12 21:36, domingo schreef: > Can I have more than one planning entity, let say 4, is it recommended or > poor design? > If any performance related warnings? > > thanks in advance > > - > with kind regards, > > -- > View this message in context: > http://drools.46999.n3.nabble.com/drools-planner-planning-entity-tp4018034.html > Sent from the Drools: User forum mailing list archive at Nabble.com. > ___ > rules-users mailing list > rules-users@lists.jboss.org > https://lists.jboss.org/mailman/listinfo/rules-users > -- With kind regards, Geoffrey De Smet ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] drools planner constrains
Read this section in the manual http://docs.jboss.org/drools/release/5.4.0.Final/drools-planner-docs/html_single/index.html#d0e150 Op 18-06-12 20:20, domingo schreef: > Hi, > In drools planner > > what exactly breaking a constraint means? > > thanks lot, > > - > with kind regards, > > -- > View this message in context: > http://drools.46999.n3.nabble.com/drools-planner-constrains-tp4018033.html > Sent from the Drools: User forum mailing list archive at Nabble.com. > ___ > rules-users mailing list > rules-users@lists.jboss.org > https://lists.jboss.org/mailman/listinfo/rules-users > -- With kind regards, Geoffrey De Smet ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] drools planner CloudProcessDifficultyComparator & CloudComputerStrengthComparator
They are used by some of the construction heuristics (Best Fit *, * Fit Decreasing). Read the manual section 4.3.3.2. Planning entity difficulty http://docs.jboss.org/drools/release/5.4.0.Final/drools-planner-docs/html_single/index.html#d0e2001 Op 18-06-12 22:01, domingo schreef: > Hi, > > In the cloudbalance example...I noticed two comparator classes... > CloudProcessDifficultyComparator (this is for planning entity class > comparator) and CloudComputerStrengthComparator (this is for planning > variable comparator class), > > please help me to understand more about these planning entity difficulty and > planning variable strength comparator classes. > why do we need this and why not use one class level comparator? > > thanks, > > > - > with kind regards, > > -- > View this message in context: > http://drools.46999.n3.nabble.com/drools-planner-tp4018035.html > Sent from the Drools: User forum mailing list archive at Nabble.com. > ___ > rules-users mailing list > rules-users@lists.jboss.org > https://lists.jboss.org/mailman/listinfo/rules-users > -- With kind regards, Geoffrey De Smet ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Hibernate Configuration with Drools 5.4
*Hi Vincent ! 50k records is very less for the projects scaling over 1 lac(100k) records and when I started working on Drools, I was assured with performance and ease in developing business rule. But if xls format(DTs) not even supporting 26k records , how can I implement it for client requirement. Initially after Heap size, I was successful in running 26k records. After few trials ,it stopped supporting 26k and we have to scale down to 3.5k. That is why I was asking to tell me some alternate solution to rectify this problem because if it can run for 26k initially, then it should continue running with the same or more number of records. So kindly help me in solving this issue which me and my team struggling for a 2 weeks and with your help I am sure we can solve this problem which would be help not only for me but also for this WHOLE COMMUNITY. Thanks !!!* -- View this message in context: http://drools.46999.n3.nabble.com/Hibernate-Configuration-with-Drools-5-4-tp4017981p4018040.html Sent from the Drools: User forum 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] drools-planner-distribution5.4 final unzipping error....
On Fri, Jun 15, 2012 at 11:46 PM, domingo wrote: > Hi, > I am try to unzip the drools-planner-distribution-5.4.0.final.zip, while > unzipping I got the following warnig messages... > DeciderScoreComparatorfact...is password protested.Please enter the > passwordin the box.. > File 'initializer' is password protected.Please enter the passwordin the > box.. > File 'Factory' is password protected.Please enter the passwordin the box.. > File 'MachineReassignmentSolution' is password protected.Please enter the > passwordin the box.. > File 'MrOriginalMacineSolutionIn..' is password protected.Please enter the > passwordin the box.. > File 'EmployeeConsecutiveAssignment' is password protected.Please enter the > passwordin the box.. > ... > ..etc > I suspect some of the files are password protected...how to get rid of > this? > > thanks > > - > with kind regards, > > -- > View this message in context: > http://drools.46999.n3.nabble.com/drools-planner-distribution5-4-final-unzipping-error-tp4017999.html > Sent from the Drools: User forum mailing list archive at Nabble.com. > ___ > rules-users mailing list > rules-users@lists.jboss.org > https://lists.jboss.org/mailman/listinfo/rules-users > -- Thanks, Mani ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Guvnor jackrabbit session issue?
On 2012/6/16 1:15, 0beron wrote: > I am having the same issue and I am using "guvnor-5.3.0.Final-jboss-as-7.0". > Could you please let me know any way to stop this. It keeps making my log > file (super) bloated. Hi, the issue reported by the previous email had been fixed in Guvnor 5.3 final. If you still see this problem, could you please provide following info so that we can take a deeper look. 1. Guvnor version 2. Application server type (JBOSS AS 7, AS 6, Tomcat?) 3. Jdk version 4. detailed stack trace. Thanks, Jervis > Thanks in advance! > > -- > View this message in context: > http://drools.46999.n3.nabble.com/Guvnor-jackrabbit-session-issue-tp3868067p4017998.html > Sent from the Drools: User forum 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 mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Cannot get Custom Forms to work in Guvnor
What is the url you are using for the custom form? Did you try to see what is going on using something like firebug of chrome developer tools? Is the URL even invoked? Best Regards, Esteban Aliverti - Developer @ http://www.plugtree.com - Blog @ http://ilesteban.wordpress.com On Tue, Jun 19, 2012 at 12:21 AM, sams wrote: > I cannot get custom forms to work in guided editor. I can create a working > set, add the field and value, and define the custom form OK. When I get to > the editing experience and click on the 'pencil' I get a popup the size > that > I declared bit with no content. It makes no difference what URL I enter. > > Can someone please confirm that this feature indeed works in 5.4.0 on > Tomcat? > > Thanks, > Sam > > > -- > View this message in context: > http://drools.46999.n3.nabble.com/Cannot-get-Custom-Forms-to-work-in-Guvnor-tp4018036.html > Sent from the Drools: User forum 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 mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
[rules-users] Cannot get Custom Forms to work in Guvnor
I cannot get custom forms to work in guided editor. I can create a working set, add the field and value, and define the custom form OK. When I get to the editing experience and click on the 'pencil' I get a popup the size that I declared bit with no content. It makes no difference what URL I enter. Can someone please confirm that this feature indeed works in 5.4.0 on Tomcat? Thanks, Sam -- View this message in context: http://drools.46999.n3.nabble.com/Cannot-get-Custom-Forms-to-work-in-Guvnor-tp4018036.html Sent from the Drools: User forum 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] drools planner
Hi, In the cloudbalance example...I noticed two comparator classes... CloudProcessDifficultyComparator (this is for planning entity class comparator) and CloudComputerStrengthComparator (this is for planning variable comparator class), please help me to understand more about these planning entity difficulty and planning variable strength comparator classes. why do we need this and why not use one class level comparator? thanks, - with kind regards, -- View this message in context: http://drools.46999.n3.nabble.com/drools-planner-tp4018035.html Sent from the Drools: User forum 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] drools planner planning entity
Can I have more than one planning entity, let say 4, is it recommended or poor design? If any performance related warnings? thanks in advance - with kind regards, -- View this message in context: http://drools.46999.n3.nabble.com/drools-planner-planning-entity-tp4018034.html Sent from the Drools: User forum 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] drools planner constrains
Hi, In drools planner what exactly breaking a constraint means? thanks lot, - with kind regards, -- View this message in context: http://drools.46999.n3.nabble.com/drools-planner-constrains-tp4018033.html Sent from the Drools: User forum 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] drools-planner-distribution5.4 final unzipping error....
this can be avoided using proper unzip tools. If any have this issue use winrar/winzip or any unzip tool. I used windows xp to unzip the file, that's why i got that error. - with kind regards, -- View this message in context: http://drools.46999.n3.nabble.com/drools-planner-distribution5-4-final-unzipping-error-tp4017999p4018032.html Sent from the Drools: User forum 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] Hibernate Configuration with Drools 5.4
ok sorry ... may be one idea then: drools is using POI as java layer to read xls sheets, and it is very very very memory consuming ... especially with .xlsx format. May be that drools parser don't even reach a point where it can check the row count ... hope this helps... sorry again . - Mail original - De: "Wolfgang Laun" À: "Rules Users List" Envoyé: Lundi 18 Juin 2012 19:40:25 Objet: Re: [rules-users] Hibernate Configuration with Drools 5.4 Relax, Vincent, it was sent due to my explicit request, and it was a maller size version. I've taken this off-list to get a first-hand impression of why these DTs are considered as a viable solution, and how it fails, etc.. And I think that the OPs Q is legitimate; techniques propagated in Drools Expert without any caveats (e.g., don't do it for more than N rows) should fail gracefully, with an error message such as "after umpteen zillion rules: out of memory" and not just turn belly-up. After having received some feedback from the OP about alternative approaches I'll report back to the list. -W On 18 June 2012 19:26, Vincent LEGENDRE < vincent.legen...@eurodecision.com > wrote: Did you really send a 50k rows table to a personal email ??? I can't believe it ... I think you should : - read the mailing list rules - read the previous Wolfgang post telling that you should consider not using such big tables (consider a rule using your 50k rows as parameters) - ask questions that are related to rules, and not java heap size limits ... De: "zeeshan" < zeeshan.spr...@gmail.com > À: rules-users@lists.jboss.org Envoyé: Lundi 18 Juin 2012 08:19:27 Objet: Re: [rules-users] Hibernate Configuration with Drools 5.4 Hi Laune ! I have mailed you my spreadsheet at ur gmail id. Thanks ! -- View this message in context: http://drools.46999.n3.nabble.com/Hibernate-Configuration-with-Drools-5-4-tp4017981p4018017.html Sent from the Drools: User forum 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 mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Hibernate Configuration with Drools 5.4
Relax, Vincent, it was sent due to my explicit request, and it was a maller size version. I've taken this off-list to get a first-hand impression of why these DTs are considered as a viable solution, and how it fails, etc.. And I think that the OPs Q is legitimate; techniques propagated in Drools Expert without any caveats (e.g., don't do it for more than N rows) should fail gracefully, with an error message such as "after umpteen zillion rules: out of memory" and not just turn belly-up. After having received some feedback from the OP about alternative approaches I'll report back to the list. -W On 18 June 2012 19:26, Vincent LEGENDRE wrote: > Did you really send a 50k rows table to a personal email ??? I can't > believe it ... > I think you should : >- read the mailing list rules >- read the previous Wolfgang post telling that you should consider not > using such big tables (consider a rule using your 50k rows as parameters) >- ask questions that are related to rules, and not java heap size > limits ... > > -- > *De: *"zeeshan" > *À: *rules-users@lists.jboss.org > *Envoyé: *Lundi 18 Juin 2012 08:19:27 > *Objet: *Re: [rules-users] Hibernate Configuration with Drools 5.4 > > > Hi Laune ! > > I have mailed you my spreadsheet at ur gmail id. > > Thanks ! > > -- > View this message in context: > http://drools.46999.n3.nabble.com/Hibernate-Configuration-with-Drools-5-4-tp4017981p4018017.html > Sent from the Drools: User forum 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 mailing list > rules-users@lists.jboss.org > https://lists.jboss.org/mailman/listinfo/rules-users > > ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Hibernate Configuration with Drools 5.4
Did you really send a 50k rows table to a personal email ??? I can't believe it ... I think you should : - read the mailing list rules - read the previous Wolfgang post telling that you should consider not using such big tables (consider a rule using your 50k rows as parameters) - ask questions that are related to rules, and not java heap size limits ... - Mail original - De: "zeeshan" À: rules-users@lists.jboss.org Envoyé: Lundi 18 Juin 2012 08:19:27 Objet: Re: [rules-users] Hibernate Configuration with Drools 5.4 Hi Laune ! I have mailed you my spreadsheet at ur gmail id. Thanks ! -- View this message in context: http://drools.46999.n3.nabble.com/Hibernate-Configuration-with-Drools-5-4-tp4017981p4018017.html Sent from the Drools: User forum 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 mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] How to write a Heuristic Knowledge in Drools Expert?
Hi Sotty.. It's ok.. Here you can find something about my project that I run into.. BioAgents http://books.google.com.br/books?id=ohX4jvy96aoC&printsec=frontcover&hl=pt-BR#v=onepage&q=BioAgents&f=false First it started using JESS.. after that Drools and Reinforcement Learning.. But the method applied to the project has not been well accepted.. This project was started for Master Thesis. There are two master thesis in respect to the BioAgents.. When I started my undergraduate research. I needed to run into this project.. That the programming wasn't modular.. The multiagent system doesn't work wery well.. The rules adopted can't change, if it change the system needed to change too.. And I needed to do all the project again alone.. The prototype now work very well.. But with rule sources limited for each biologist's viewpoint. All I need to do.. is to have independence of viewpoint.. and suggest the best solution.. -- View this message in context: http://drools.46999.n3.nabble.com/How-to-write-a-Heuristic-Knowledge-in-Drools-Expert-tp4018012p4018028.html Sent from the Drools: User forum 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] How to write a Heuristic Knowledge in Drools Expert?
Predictive models, such as neural networks, support vector machines, cluster models, decision trees etc can be used in two ways. One is classification: you take some input data/facts and the model will estimate if this particular set has a given characteristic or not. E.g. given a set of genes, you might want to see if the individual is prone to suffer from a disease. You might be more interested in regression/prediction: given a set of input data/facts, the model will estimate a dependent variable. E.g. given a sequence of genes, the model will try to predict the next in line. Models can give probabilities, too. The model would have to be trained with an external data mining tool, but then could be imported by drools to provide a runtime execution environment. Even then, before thinking of the particular implementation, I'd try to model the problem. Interviewing the domain experts and casting their words into an organized representation will not be easy, I'm afraid. We can share thoughts sometimes, if you want/can. Davide -- View this message in context: http://drools.46999.n3.nabble.com/How-to-write-a-Heuristic-Knowledge-in-Drools-Expert-tp4018012p4018027.html Sent from the Drools: User forum 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] How to write a Heuristic Knowledge in Drools Expert?
Even more i read things seems more complex that are.. but in the same way seems not.. I partially read about PMML and the following model seems a good solution: http://www.ibm.com/developerworks/industry/library/ind-PMML1/image001.gif If I have facts mined expecting some value to show how precise is the alignment.. It's good idea to normalize them and predict the most significant fact. If I know it well drools could be the Decision processing module.. But I need to learn about it first.. I doesn't have the knowledge.. Thanks so much.. Daniel Souza -- View this message in context: http://drools.46999.n3.nabble.com/How-to-write-a-Heuristic-Knowledge-in-Drools-Expert-tp4018012p4018026.html Sent from the Drools: User forum 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] How to write a Heuristic Knowledge in Drools Expert?
Hi Geoffrey.. Thanks for reply.. I'm brainstorming the biologist's Knowledge domain.. maybe heuristic approach could be the solution or I can do it in a simple way.. But for now,. I have just rules with different ways to use them.. specific for each biologist. The Idea is to mix them to have an unique solution.. Bye.. Daniel Souza.. -- View this message in context: http://drools.46999.n3.nabble.com/How-to-write-a-Heuristic-Knowledge-in-Drools-Expert-tp4018012p4018025.html Sent from the Drools: User forum 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] How to write a Heuristic Knowledge in Drools Expert?
Hi Daniel, if you have predictive models you might be interested in the PMML integration we're working on.. but it's indeed necessary that you define the nature of your problem and your requirements first :) Let us know if you need more assistance Best Davide -- View this message in context: http://drools.46999.n3.nabble.com/How-to-write-a-Heuristic-Knowledge-in-Drools-Expert-tp4018012p4018024.html Sent from the Drools: User forum 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] How to write a Heuristic Knowledge in Drools Expert?
Hi sotty.. I know how A* algorithm works, but it's only a theoretical knowledge.. It's good to know that I can express Heuristic Knowledge using drools rules.. I just need to adjust well the problem domain to know what I really need to do.. Thanks for showing how to use an heuristic behavior inside the rules.. I'm reading a lot.. And my problem is that, for an specific situation in my problem, the biologists express the problem in different viewpoints and I trying to understand well about these possible conclusions.. whose facts have values with data mining to expect how precise is the alignment between the sequence submitted and the query matched. But there are constraints that I need to know first.. Bye Daniel Souza -- View this message in context: http://drools.46999.n3.nabble.com/How-to-write-a-Heuristic-Knowledge-in-Drools-Expert-tp4018012p4018023.html Sent from the Drools: User forum 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] How to write a Heuristic Knowledge in Drools Expert?
Thanks for reply.. Laune There are things in the book that would answer some questions.. Well, it's good to know that I can express a Heuristic Knowledge in Drools Expert. And I have drools planning to use metaheuristic.. But, I need to model my problem well first to know what approach I will follow to express well my rules. With a lot of interviews in how to express the biologist's expertise about manual annotation of genetic sequences.. I noticed that there are different viewpoint in how to express the manual annotation. The task of the manual annotation problem was divided in smaller problems using multiagent system approach. I tried to write the problem here.. but It will be too large.. and I erased part of what I wrote.. :) Bye.. Daniel Souza -- View this message in context: http://drools.46999.n3.nabble.com/How-to-write-a-Heuristic-Knowledge-in-Drools-Expert-tp4018012p4018022.html Sent from the Drools: User forum 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] How to write a Heuristic Knowledge in Drools Expert?
If your problem is A* search efficient, you should use A* search. It's better and simpler than any metaheuristic. Problem is, most real-life problems are NP complete and A* search doesn't work efficiently on those. Some examples: - Quickest route between 2 cities: Use A* search - Quickest route to visit 4+ cities (TSP): Use metaheuristic (or LP) - Bin packing: Use metaheuristic The easiest way to make this decision, is to take a look if you find your problem on this list: http://en.wikipedia.org/wiki/List_of_NP-complete_problems (which is a long list of Karp's list: http://en.wikipedia.org/wiki/Karp%27s_21_NP-complete_problems ) If your problem or a subproblem of it is on that list, use Planner, because A* search will take a few billion years too long. Otherwise go for A* search. Op 17-06-12 21:57, Davide Sottara schreef: > Are you familiar with the A* algorithm? A heuristic function is typically > used in (state-)space search problems: at a given point, you may have > alternative "paths" you can choose and the heuristic will help you > estimate/predict how "successful" exploring that branch would be by giving a > score. The function usually involves reviewing your path so far and trying > to look ahead before making another step. > More generally, heuristic decision making is a process where you choose an > action (the "path") according to your current knowledge (your state) and > some prediction criteria which is not necessarily the best, but the most > promising. > > A heuristic search rule could look like this: > > rule "Example" > when >// you have a current state, given by one or more patterns >$here : CurrentState( $history : path ) >// then you query for / generate a list of possible follow-up alternatives >?possiblePath( $here, $nextFromHere ) >// optionally, you filter those which make sense qualitatively >Filter( ... ) from $nextFromHere // notice: $nextFromHere might > not be in the WM >// optionally, and/or you apply some heuristic scoring function, filtering > quantitatively >// notice that it might use the history, the current position or the next > tentative step >$score : Double( this> THRESHOLD ) from heur( $path, $here, $nextFromHere > ) > >// any additional logic, e.g. accumulate scores? > > then >// make your choice >insert ( new Step( $nextFromHere ) ); >// the step will later modify the currentState > end > > Notice that the following MUCH simplified version - a rule - is doing the > same thing... Simply, there is only one promising and meaningful path - e.g. > administering an aspiring - which is automatically chosen given the current > state (of Fever). > > rule "a more traditional rule" > when >$path : MedicalHistory( ... ) >$here : FeverDiagnosis( temp< 38 ) >// ?possiblePath( $here, $nextFromHere ) --> let's say aspirin, strong > antibiotics and bone marrow transplant >// but you can already figure out the scores a realistic function would > give to the alternatives... >// (ok, allergies from $path would make it different, but ignore...) >// $score : Double( this> THRESHOLD ) from heur( $path, $here, > $nextFromHere ) > then >// so you hardcode your choice! >insert( new AspirinAdministration() ); > end > > > Drools Planner is slightly different: you might indeed use it to define a > plan - i.e. a sequence of actions to reach a goal - ***provided*** that you > can reliably estimate the consequences of any action (move) you allow. > Planner, however, already implements its own search strategies. What you > provide there is a "score" for the current state (aka candidate "solution"), > expressed in terms of hard and soft constraints, that planner algorithms > will consume. You still have to provide the alternative moves and their > effects, but the main difference between the scoring (the constraints) and > heuristic function is that the former usually just look at the current > state. So, in planner 1) you tentatively make a step and evaluate the > results, 2) then backtrack and repeat until you have seen all possible > paths, 3) you choose the best one according to the strategy (it's called > meta-heuristic) > > With a heuristic function, you evaluate your possible future outcome (not > just one step ahead) before you do make the step, so that's "predictive". > > Now, coming back to your question : how do you implement a heuristic > function? > The answer is: first you have to define it, given your problem :) > Only then you can think of an implementation, but functions(), queries and > rules with optional accumulates seem good candidates! > > Best > Davide > > -- > View this message in context: > http://drools.46999.n3.nabble.com/How-to-write-a-Heuristic-Knowledge-in-Drools-Expert-tp4018012p4018015.html > Sent from the Drools: User forum mailing list archive at Nabble.com. > ___ > rules-users mailing list > rules-users
