[rules-users] Selective TSP?

[rengana]

Hi, 

i'm evaluating Drools Planner as a solver for a "normal" OP problem but i
can't find any example of drools usage on problems with partial set as a
feasible solution, all drools examples and use cases seems to always try to
plan every object even if that makes the solution unfeasible. 

I was able to model a OP problem using a modified version of the vehicle
routing example using two vehicles, one for containing the customers visited
(real solution) and a second vehicle for containing customers not visited.
It works OK but this approach seems very dirty... 

Doesn't Drools Planner support a partial set solution that is feasible but
doesn't have to contain all problem facts? Maybe a parameter to allow
partial set solutions, and if so, be able to define the minimal problem
facts needed to consider a posible solution.

I was expecting that at least in the contruction phase the planner found
some feasible solutions, since each step it is putting one element in the
solution and then evaluating score, but found that the best solution passed
to local seach isn't the best feasible solution of the contruction phase but
always a unfeasible solution because somehow it's forcing to contains all
problem facts.

Is this correct?


Best regards, 
Rui



--
View this message in context: http://drools.46999.n3.nabble.com/Selective-TSP-tp4022903.html
Sent from the Drools: User forum mailing list archive at Nabble.com.