Modifying problems

These problem modification routines change a problem object after it has been created.

Problem modification routines change a problem object after it has been created.

Modifications that can be made include:

• adding rows and columns to the constraint matrix,

• deleting rows and columns from the constraint matrix,

• changing the sense of the objective function,

• changing the value of coefficients in the constraint matrix,

• changing an objective or righthand side coefficient,

• changing the bounds on a variable,

• changing the sense of a constraint, and

• changing names of rows or columns.

When you are modifying multiple existing objects in the object-oriented APIs (C++, Java, C#.NET), you have a choice between modifying each object individually in a sequence of operations (such as in a loop) or making accumulated changes simultaneously in a single operation (for example, on an array of modeling objects). Generally, the single operation is more efficient than a sequence of operations.

For example, in the C++ API, it is possible to change the lower bounds on a set of variables through a sequence of invocations of the method IloNumVar::setLB; however, it is more efficient to accumulate the variables in an array (an instance of IloNumVarArray) and then to apply the method IloNumVarArray::setBounds simultaneously to all the variables in that array. The same observation about efficiency applies to objects of other classes derived from the base class IloExtractable, such classes as IloRange, IloSOS1, IloSOS2, and IloSemiContVar.