Originally posted by: BugHunter

After solving LP I called cplex.Solution.dual and got the message:
??? Reference to non-existent field 'dual'. at the same time the interactive cplex gives me the dual solution.
What is wrong with cplex.Solution.dual ?

thanks
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

Which API do you use? Python?
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

If you use Python try function

cplex.solution.get_dual_values

to read the dual solution values.
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: BugHunter

I am using MATLAB API Version 7.10.0.499(R2010a)
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: BugHunter

Nobody knows what the problem? There is no method for the class that supposed to be and nobody comes across to this problem?
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: John Cui

Could you please try below code and tell me the cplex.Solution?

>> cplex=Cplex()
     
ans=
     
   Cplex handle
     
   Properties:
      Model: [1x1 struct]
      Param: [1x1 struct]
      DisplayFunc: @disp
     
   Methods, Events, Superclasses
>> cplex.addCols(1)
     
ans=
     
   Cplex handle
     
   Properties:
      Model: [1x1 struct]
      Param: [1x1 struct]
      DisplayFunc: @disp
     
   Methods, Events, Superclasses
>> cplex.addRows(1,1,1)
     
ans=
     
   Cplex handle
     
   Properties:
      Model: [1x1 struct]
      Param: [1x1 struct]
      DisplayFunc: @disp
     
   Methods, Events, Superclasses
>> cplex.solve
 
ans = 
 
          status: 1
    statusstring: 'optimal'
            time: 0
          objval: 1
               x: 1
          method: 2
            dual: 1
              ax: 1
     reducedcost: 0
           basis: [1x1 struct]
 
>> cplex.Solution.dual
 
ans =
 
     1

You can see we have cplex.Solution.dual
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: BugHunter

John, thanks a lot. I got the problem: when you set the 'ctype' parameter, even it is 'C' - for continuous vars, it takes the problem as an MIP and doesn't calculate the duals (even I set the lpmethod to the primal cplex
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: John Cui

If you want convert your mip model to lp, please remove the ctype field by:

cplex.Model = rmfield(cplex.Model, 'ctype');

then solve the model again.
#CPLEXOptimizers
#DecisionOptimization