Linear relaxations
Some parts of an optimization problem may be best expressed and solved in the declarative (solver-independent) manner of Mathematical Programming (MP). Other parts benefit from constraint propagation and search directions provided by a CP solver. So why not trying to get the best of both worlds by combining the MP and CP approaches?
Xpress Kalis allows you to benefit from CP and MIP hybridization in a clean and easy way by means of linear relaxations of the CP model. Xpress Kalis can build automatically several linear (or mixed integer linear) relaxations of the CP problem and use Xpress Optimizer to solve them.
This process can be fully configured ranging from a complete user definition of the relaxation and choice of when to launch the relaxation to a fully automatic relaxation.
Xpress Kalis uses a double modeling approach by exchanging automatically and in bidirectional way, informations such as objective bounds, infeasibility, optimal relaxed solutions and reduced costs between the cp solver and the linear relaxations solver(s) during the search for a solution.
Note: the linear relaxation functionality of Xpress Kalis can only be used if Xpress Optimizer is installed and licensed.
You will find below the list of relaxation related functions and procedures defined within Xpress Kalis:
|
Add a linear relaxation solver to the linear relaxation solver list
|
|
|
Clear the linear relaxation solver list
|
|
|
Returns an automatic relaxation of the cp problem
|
|
|
Remove a linear relaxation solver from the linear relaxation solver list
|
|
|
Pretty printing of a linear relaxation
|
|
|
Export the linear relaxation in LP format
|
|
|
Fix the continuous variables to their optimal value in the relaxation solver passed in argument
|
|
|
Generate and add cuts to the relaxation passed in parameters
|
|
|
Get an indicator variable for a given variable and a value.
|
|
|
Get the linear relaxation for a constraint
|
|
|
Returns a linear relaxation solver from a linear relaxation, an objective variables and some configuration parameters
|
|
|
Get a reduced cost value from a linear relaxation solver
|
|
|
Returns the optimal relaxed value for a variable in a relaxation
|
|
|
Get a largest reduced cost variable selector from a linear relaxation solver
|
|
|
Get a nearest relaxed value selector from a linear relaxation solver
|
|
|
Launch LP/MIP solver without CP branching.
|
|
|
Set integrality flag for a variable in a linear relaxation
|
|
|
Parameter setting for a linear relaxation solver.
|
|
|
Set the verbose level for a specific linear relaxation solver
|
© 2001-2021 Fair Isaac Corporation. All rights reserved. This documentation is the property of Fair Isaac Corporation (“FICO”). Receipt or possession of this documentation does not convey rights to disclose, reproduce, make derivative works, use, or allow others to use it except solely for internal evaluation purposes to determine whether to purchase a license to the software described in this documentation, or as otherwise set forth in a written software license agreement between you and FICO (or a FICO affiliate). Use of this documentation and the software described in it must conform strictly to the foregoing permitted uses, and no other use is permitted.