Originally posted by: SystemAdmin

[o_O said:]

Hi All,

It's my first time to use this language for programming. The purpose of the following codes is to block material flow with high cost penalty (store = 0 and ship = 1).

dvar boolean store[products][distributionCenters];                       
dvar boolean ship[orders][distributionCenters];
// 0:Stored in plant warehouse, 1 Otherwise     
// 0:Shipped from plant warehouse, 1 Otherwise                                                                                                     
dvar float+ inv[distributionCenters][products][timePeriods];     
dvar float+ production[products][timePeriods];         

// Objective Function Terms
// Shuttle Costs
dexpr float ShuttleCosts=
       sum(d in demands, t in timePeriods, pr in products, dc1 in distributionCenters, dc2 in distributionCenters, ol in orderlines, s in ShuttlCosts: d.order==ol.order && d.product==ol.product && s.name==dc1 && s.period==t && d.period==t)
       (d.demand*(2*store[ol.product][dc1]-ship[ol.order][dc2])*s.cost + 0.01/(store[ol.product][dc1]-ship[ol.order][dc2] + 1.01));   

The error message:

The algorithm cannot extract expression: 0.01/(store[ol.product][dc1]+ship[ol.order][dc2]*(-1) + 1.01).


Please help.


#DecisionOptimization
#OPLusingCPOptimizer

Originally posted by: SystemAdmin

[Didier Vidal said:]

Do you want to solve this problem with CPLEX(Math Programming) or with CP Optimizer(Constraint Programming) ?

If you want to use CPLEX, you can't divide by decision variables (this term is illegal for CPLEX: 0.01/(store[ol.product][dc1]-ship[ol.order][dc2] + 1.01) )

If you want to use CP, then there is an other restriction: only discrete decision variables are supported (all dvar must be either int or boolean. They can't be float). In addition, you should start your model with 'using CP;'

Hope this helps.

Didier.
#DecisionOptimization
#OPLusingCPOptimizer