We're working on multi-threading, both:
- in Drools (based on the Phreak algorithm): fine grained, for score
calculation
- and in OptaPlanner ( https://issues.jboss.org/browse/PLANNER-76 ): course
grained, for batches of moves to be evaluated
--
View this message in context:
Duplicates
http://stackoverflow.com/questions/23213141/how-to-make-optaplanner-a-rest-webservice
--
View this message in context:
http://drools.46999.n3.nabble.com/How-to-make-optaplanner-a-REST-webservice-tp4029318p4029357.html
Sent from the Drools: User forum mailing list archive at
Sounds like a typical OptaPlanner problem.
Domain model:
- CareTaker
- OldPerson
- Day
- Visit(OldPerson, Day)
- @PlanningEntity VisitAssignment (Visit, @PlanningVariable CareTaker)
Using the OptaPlanner employee rostering example:
- CareTaker == Employee
- OldPerson == ShiftType
- Day ==
Btw, this post wasn't synced to the drools mailing list or gmane (see the
yellow bar inside your post on nabble), so your posts don't show up on my
default radar (as I use gmane).
--
View this message in context:
Use Immovable filter (see docs chapter about continuous planning) to make
sure the already planning entities don't move. See course scheduling and
nurse rostering video's for 2 different approaches.
You 'll want to load recent historical data because they affect constraints
on the newly placed
Mats Norén wrote
Hi,
I've got a really simple problem that I can't seem to fix. :-)
I'm trying to use a simplified model for a room planning problem.
The planning entity is a Topic and the planning variables are Room and
TimeSlot.
A Person is tied to a Topic as a moderator.
Each Person can
erere121 wrote
Hello every one ,I study the drool.plannar recently and both tried the
planner 5.5 and planner 6.0(optaplanner) to solve my problem. I found
something puzzles:
1 Is there any big difference between planner 5.5 and planner 6.0. I run
the same process in both planner 5.5 and
pppoole wrote
I have a planning entity which has a ResourceType on it and I am
attempting to schedule resources on the entity. The Resource is the
Planning variable and the Resource instance has a ResourceType on it as
well. I want to filter to make sure that a ResourceType matches between
Mister Nono wrote
Hi,
I use Drools Planner and an object does not accepted by a hard rule.
And then, I obtain à negative hard result and I must present to user why
the result is not correct.
How to know the object broken by the hard rule when the result is negative
?
Thank.
Bye,
mvermand wrote
drools planner to forcefully reject any solution that breaks this one
rule,
continue searching for other solutions and guarantee that the best
solution will not propose a combination that fails this rule?
Use terminationscoreAttained...//. See manual section termination
for
Nevermind my last post, my mind wasn't fully awake anymore :)
You can set your own RuleBase through the config API, look for the section
that explains that in the Planner reference manual. Do ctrl-f on setRuleBase
--
View this message in context:
The nurse rostering example is an implemtation of INRC2010: that competition
defined the problem.
Look in the planner manual for the link to inrc2010 website and look for a
pdf there:
iirc it contains a full spec of the input format
--
View this message in context:
Paul T wrote
Exception in thread main java.lang.IllegalArgumentException: The
valueSelectorConfig (ValueSelectorConfig(null)) has no configured
planningVariableName (null) for planningEntityClass (class
com.domain.Orders) and because there are multiple in the
planningVariableNameSet
skatta1986 wrote
Hi,
Could any one tell me if there is any encryption in Drools.
If you're referring to the bouncycastle dependencies in maven - that's do to
a rogue dependency that drags them in because that rogue forgot to tag
bouncycastle as optional - they shouldn't be there. Do mvn
Jon L wrote
Is there a schema for the drools planner configuration file?
I'd like to validate a configuration file at the beginning of my
application. It would also help in knowing the possible values for some
of the configuration entities.
Thanks,
Jonathan
Currently no xsd schema file
Not sure if it's possible like this or if it needs a hack/pull request.
In the latter case, see these files:
https://github.com/droolsjbpm/drools-planner/blob/master/drools-planner-core/src/main/java/org/drools/planner/config/score/director/ScoreDirectorFactoryConfig.java#L235
--
View this
Skip the tests with -DskipTests
See README on github. If jenkins doesnt report the failure, make a jira for
the locale specific failing test
venkatrangan59 wrote
Hi,
I attempted to build the latest drools distribution using Maven. The
drools-core built successfully, but drools-compiler failed
softConstraintsBroken : calculate the total length of the tour, right ?
Yes, but they use the occurrences injected by distanceToPreviousAppearance
and distanceFromLastVisitToDomicile.
distanceToPreviousAppearance adds the distance from a Visit to the
previous visit or Domicile (= starting
See
http://drools.46999.n3.nabble.com/Planner-Planning-Entity-difficulty-comparator-Vs-difficulty-weight-factory-tp4019858.html
http://drools.46999.n3.nabble.com/Planner-Planning-Entity-difficulty-comparator-Vs-difficulty-weight-factory-tp4019858.html
--
View this message in context:
See
http://drools.46999.n3.nabble.com/planning-difficulty-problem-td4019857.html
http://drools.46999.n3.nabble.com/planning-difficulty-problem-td4019857.html
--
View this message in context:
spinjala wrote
Once the solver finds that a workorder (WO1) couldn't be assigned to any
resource in the list of resources (due to the constraints), it simply
picks up the 1st resource in the list with highest hard score and assigns
it (even though this assignment breaks all constraints). My
typo:
Planner perceives an planning variable with the planning value *null* as
being uninitialized, not as unassigned. This is due to
https://issues.jboss.org/browse/JBRULES-3317
--
View this message in context:
spinjala wrote
I am new to Drools planner. I have one issue and wanted to understand how
to resolve it best.
I am trying to use planner to optimize the scheduling of work orders to
resources. I have certain resources that need to be assigned to a work
order in a way that maximizes the
Snoopy wrote
Hello,
where can i find an example for your Drools Planner bin packing
function, please ?
Regards,
Snoop
Cloud balancing is a example of non-related 3D bin packing.
Non-related means that the dimensions don't interact to form a surface or
volume.
There's no example for
Take a look at the VRP example too: it's basically the TSP example, but with
multiple vehicles.
1) It depends, but I would add a hard constraint for this. Once 5.5.0.Final
is out, you might use move Filtering instead, so it doesn't even evaluate
those solutions (but that also blocks it from
Garf wrote
I wanted to get a better understanding of this problem.
I'm getting a score Corruption error -- but only when environmentMode is
set to TRACE.
When it's set to REPRODUCIBLE, I get no error.
Jon
Score corruption detection greatly reduces performance, so it's only enabled
Yes, you're right: it's because of performance reasons during incremental
score calculation.
If a CloudComputer changes during planning too, it should be set as dirty in
Drools by Planner calling (workingMemory.modify(computer)). If the whole
computer is made dirty, that has a serious effect on
Garf wrote
Ah, I see.
It would be handy to update the docs to clarify such (is there a community
wiki around?)
Not sure. Normally, I'd like to document as much as possible, but I 'd
rather leave this one undefined for now and see how people solve, instead of
pushing people into a certain
Antlr is special because version 3 depende on version 2. For more info look
for the jira in our issue tracker. It works though.
--
View this message in context:
http://drools.46999.n3.nabble.com/Guvnor-5-3-Dependency-with-antlr-3-3-POM-XML-issue-tp3968780p4017611.html
Sent from the Drools: User
((DroolsScoreDirector) stepScope.getScoreDirector()).getWorkingMemory()
michael.zeising wrote
Hi there,
I'm writing my own /BestSolutionRecaller/ for collecting every solution
that is evaluated. I'm overriding
so I get the new solution by calling
Now, I'd like to find the
Thanks for sharing! Nice work :)
Some ideas:
- The score calculation count is low. See the chapter about incremental
score calc in manual.
- from planning entity should work as specified in the manual. Raise a jira
if it doesnt. I dont see why you d need it in this use case though.
Geoffrey
YuraK wrote
I'm new in Drools Planner. I'm investigating the Hospital bed planning
(PAS - Patient admission scheduling) problem,
I'm interesting about an existence of feasible solution for the current
problem.
The idea is follow:
1) Doctor/User is sending a request (booking bed for
Drools Planner has 7 real examples, which use large datasets and are solved
near optimal in reasonable time. For example, the nurse rostering example
have search spaces over 10^2000. Research competions show that
metaheuristics are the only algorithems that can handle it in robuust way.
If I understand your question correctly:
So basically your rules are now of the form:
rule a
salience 4
when
travel trough heathrow // score
then
try not to travel to heathrow // search
end
Since Planner searches for you, it would be something like this:
rule a
when
travel trough
multi level score definition is easy to implement, see manual on how to
define a custom score definition.
as for rule generation at runtime, there is an internal api for that that
edson wrote. start a new question with a good subject for that. But i am not
sure how to go about reading the
Untill 5.4.0.beta2, the examples GUI reused the Solvers WorkingMemory, which
is bad. In CR1 this is no longer the case (see github master), but not
perfect yet. For 5.5 I ll improve and document it
https://issues.jboss.org/browse/JBRULES-3328
In all 3 cases the gui can get the constraint
There is a lot of demand for this feature and its not a big change, so I am
going to prioritize it. Keep an eye on the git repo in the next few
weekends. I have to finish a vrp improvement first (chaining swap), but that
shouldn't take more then 1 weekend.
--
View this message in context:
hartnellg wrote
Thanks for the suggestions for reading from Excel files.
I've used poi previously, and it will be worth having a look at jexcelapi.
As far as the text files in the input directory, it's about how best to
create new ones.
For example where I have different shifts, unwanted
barnesjd wrote
Hi everyone. I'm new to Drools (using 5.2). I'm using it in Grails
(2.0.1) with the drools-gorm plugin (0.5.6) (warning: Grails is also new
for me). I've noticed that whenever I do something that is perhaps
syntactically invalid, my rules simply don't fire. Given that I'm
Anton Hughes wrote
The working directory should be set to the directory that
contains the data directory. This is different in a git clone
(drools-planner/drools-planner-examples) and the release zip.
If you 're using the distribution zip, the runExamples.sh or bat should just
work. If
Maybe jexcelapi or poi can help you read the excel file and you can convert
it into the domain objects directly, no need to go throug xml.
Alternatively, take a look at the txt files in the dir input and the import
button in the example app.
--
View this message in context:
Great work Reinis!
--
View this message in context:
http://drools.46999.n3.nabble.com/rules-users-Drools-Planner-Hard-constraint-appears-to-be-ignored-tp3777209p3812319.html
Sent from the Drools: User forum mailing list archive at Nabble.com.
___
ndipiazza wrote
Related to posts:
http://drools.46999.n3.nabble.com/Drools-5-2-performance-building-KnowledgeBase-tc3239746.html#a3284343
We don't have an example yet, but it's basic seam3 security with picketlink.
Check those docs
--
View this message in context:
http://drools.46999.n3.nabble.com/rules-users-Guvnor-5-4-0-beta-and-Authentication-tp3751650p3752944.html
Sent from the Drools: User forum mailing list archive at
johan.kumps wrote
Hi,
Could you point me to a document providing some more information on this.
Using the Drools Planner documentation alone I'm not able to find a design
for my problem.
Kind regards,
This problem, the most optimal service composition achieving a certain
consumer
I've heard a similar use case before and someone on irc.codehaus.org #drools
implemented one with Planner.
Some of the improvements I am working on for vehicle routing might be
applicable in this use case too.
--
View this message in context:
ssandban wrote
Are there and hardware or memory requirements for using planner in solaris
environment
I like to use a max memory of 512M, but most examples (which have real-sized
data, see reference manual for numbers) run fine with far less memory (the
default of 64M is fine IIRC).
In the manual, read delta score calculation. That gets broken somehow.
Drools 5.1 had a bug that broke that, fixed since 5.2.
you're probably changing a fact without doing wm.update
--
View this message in context:
In the examples, for the lower right button constraints i hack it, but
don't do that
Just create a new rulebase from the same score drls and throw it in there,
then extract the facts filtered on constraintoccurence
Note that you canvwrite your own cons occ, domain specific and use that in
Try mySolution.getScore().getHardScore() == 0
--
View this message in context:
http://drools.46999.n3.nabble.com/drools-planner-5-3-0-Final-Sorting-successfully-planned-entities-tp3471628p3472238.html
Sent from the Drools: User forum mailing list archive at Nabble.com.
See manual, section on setRulebase() instead of scoredrl
Feel free to open a jira if bits can improve.
--
View this message in context:
http://drools.46999.n3.nabble.com/How-can-Drools-Planner-consume-Guvnor-generated-drl-file-tp3467722p3469265.html
Sent from the Drools: User forum mailing list
KD wrote:
Hi,
Reading the manual I came across Creating a starting Solution. Is there
a way to not provide this? For a combinatorial linear model heavy with
hard constraints and with resonably large number of binary variables,
implementing a method that will generate this intial solution
KD wrote:
I went through some presentations on bin packing/cloud balance problems.
I am interested in knowing about computational results relating to these
problems, specially related to the size of the instances tackled and
computational times. So how well do the methods scale with problem
Take a look at the manual on how to get started:
http://docs.jboss.org/drools/release/5.2.0.CR1/drools-planner-docs/html_single/index.html#d0e137
To get started in Eclipse, either just download the zip and add the jar
binaries to your classpath.
(Or better: use maven (with optionally
Do I need Guvnor also for defining rules?
No, that's an optional improvement you can try to use later. Start with
using plain drl files like in the examples.
The integration between Guvnor and Planner is still very raw, but it's
possible via *SolverConfig.setRuleBase().
--
View this message in
For the drools-camel-server war there are not specific wars yet (unlike for
guvnor).
The trick is to remove some of the libs from WEB-INF/lib inside the war.
See the github link on this page to get an idea of what libs should be
removed:
Nothing up to date AFAIK. I assigned the jira to me a while back.
If you write it and make a pull request on github I ll sponsor it and see if
I can merge it into master.
--
View this message in context:
Dont just schedule adds: also schedule reserve adds in a queue.
--
View this message in context:
http://drools-java-rules-engine.46999.n3.nabble.com/Subject-Drools-Planner-Design-Suggestions-tp2169722p2196930.html
Sent from the Drools - User mailing list archive at Nabble.com.
Gabor's guestimate sounds like a good backup plan implementation.
Like you cant predict which nurses fall sick, you can presume (by historic
data) the average and max sick leaves over 3 seqential timeslots. And
scheduling your nurses tightly around that, minizes the irreplaceable nurses
who
I agree on the normal constraints with vincent. The 100% sold is like nurses
falling sick.
You cannot plan it in advance. There are 2 things todo to handle it.
- Include a backup plan in your solution with a softer score: add A on time
X, add B if A sells out. Like scheduling reserve nurses.
-
60 matches
Mail list logo
