8:28 a.m.
I have n buckets and mn entities. Each bucket must have m entities. Each
entity carries a weight. My objective is to even out the total weight per
bucket, make them close to the average.
For example, a solution like this is skewed, see their sums.
bucket 1: 1,2,4,5=12
bucket 2: 6,8,9,6=29
bucket 3: 7,7,4,7=25
The same entities may be placed into buckets like this. This is more
desirable as the sums are closer to the mean (and each other); one does not
dominate the others.
bucket 1: 1,8,7,4=20
bucket 2: 6,2,9,6=23
bucket 3: 7,7,5,4=23
*Question:*How do I write a rule for this:
for each bucket, penalize (with soft scores) a bucket whose total
entity.weight differs from the average total entity.weight of the buckets.
Will I get the same results when I penalize each bucket for differing with
*each other*?
--
View this message in context:
http://drools.46999.n3.nabble.com/How-to-even-out-n-buckets-in-Drools-Pla...
Sent from the Drools: User forum mailing list archive at Nabble.com.
9:54 a.m.
write a score rule that does an insertLogical of a class of the bucket
with the least size
when
$b : Bucket($s : size)
not Bucket($s > size)
then
insertLogical(Minimum($s))
Then write a score rule that punishes the difference squared for each bucket
when
Minimum($m : size)
$b : Bucket(size > $m, $s : size)
then
insertConstraintOccurrence(...,
($s - $m) * ($s - $m), // score punish
$b); // cause
Op 03-02-12 15:28, aitchnyu schreef:
I have n buckets and mn entities. Each bucket must have m entities.
Each
entity carries a weight. My objective is to even out the total weight per
bucket, make them close to the average.
For example, a solution like this is skewed, see their sums.
bucket 1: 1,2,4,5=12
bucket 2: 6,8,9,6=29
bucket 3: 7,7,4,7=25
The same entities may be placed into buckets like this. This is more
desirable as the sums are closer to the mean (and each other); one does not
dominate the others.
bucket 1: 1,8,7,4=20
bucket 2: 6,2,9,6=23
bucket 3: 7,7,5,4=23
*Question:*How do I write a rule for this:
for each bucket, penalize (with soft scores) a bucket whose total
entity.weight differs from the average total entity.weight of the buckets.
Will I get the same results when I penalize each bucket for differing with
*each other*?
--
View this message in context:
http://drools.46999.n3.nabble.com/How-to-even-out-n-buckets-in-Drools-Pla...
Sent from the Drools: User forum mailing list archive at Nabble.com.
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
--
With kind regards,
Geoffrey De Smet
3:54 a.m.
I got the basic idea of throwing broken constraints based on the square of
the difference.
However, the total weight is not a property of the bucket, but the *sum of
weights of entities* who happen to be inside that bucket. It is not a
property *of the bucket* as such.
This is a form of what I have in mind:
when
$smallest_bucket=Bucket(bucket for which min sum of entity.weight )
$other_bucket_sum=Number(min sum of entity.weight for each bucket !=
$smallest_bucket)
then
$insertLogical(constraint(negative soft points for difference,
responsible_object))
Could you help with formulating a rule like this?
BTW, what is the Minimum class? I saw no documentation for that, only the
min() from the 'Multi-function Accumulates'.
#########
when
$b : Bucket($s : size)
not Bucket($s > size)
then
insertLogical(Minimum($s))
Then write a score rule that punishes the difference squared for each bucket
when
Minimum($m : size)
$b : Bucket(size > $m, $s : size)
then
insertConstraintOccurrence(...,
($s - $m) * ($s - $m), // score punish
$b); // cause
#########
--
View this message in context:
http://drools.46999.n3.nabble.com/How-to-even-out-n-buckets-in-Drools-Pla...
Sent from the Drools: User forum mailing list archive at Nabble.com.
8:18 a.m.
As follow up, I included my domain model and an incomplete rule
*Problem facts*:
Entity:
weight: int
Bucket:
id: int
*Planning entity*:
EntityBucketAssignment
entity: Entity
bucket: Bucket (planning variable)
*The rule fragment is*:
when
$b1:Bucket($id:id)
$minSum:Number from accumulate(
Entity(
bucket==$b1,
$weight:wieght
),
sum($wieght)
)
not($b2:Bucket(id!=$id)
$minSum:Number from accumulate(
Entity(
bucket==$b2,
$weight:wieght
),
sum($wieght)
)
)
then
...
Could you correct this?
--
View this message in context:
http://drools.46999.n3.nabble.com/How-to-even-out-n-buckets-in-Drools-Pla...
Sent from the Drools: User forum mailing list archive at Nabble.com.
6:43 a.m.
Op 04-02-12 15:18, aitchnyu schreef:
As follow up, I included my domain model and an incomplete rule
*Problem facts*:
Entity:
weight: int
Bucket:
id: int
*Planning entity*:
EntityBucketAssignment
entity: Entity
bucket: Bucket (planning variable)
*The rule fragment is*:
when
$b1:Bucket($id:id)
$minSum:Number from accumulate(
Entity(
bucket==$b1,
$weight:wieght
),
sum($wieght)
)
then
insertLogical BucketTotal($b1, $weight)
do the rest of the rule in a new rule based on the BucketTotal facts.
See examples nurserostering's Employee*Total class or something named
like that.
not($b2:Bucket(id!=$id)
$minSum:Number from accumulate(
Entity(
bucket==$b2,
$weight:wieght
),
sum($wieght)
)
)
then
...
Could you correct this?
--
View this message in context:
http://drools.46999.n3.nabble.com/How-to-even-out-n-buckets-in-Drools-Pla...
Sent from the Drools: User forum mailing list archive at Nabble.com.
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
--
With kind regards,
Geoffrey De Smet
4678
days inactive
4680
days old
4 comments
2 participants
participants (2)
-
aitchnyu -
Geoffrey De Smet