Changing option values

You can use execute blocks to change CPLEX parameters, CP parameters, and OPL settings within an OPL model.

The code examples are available at the following location:

<Install_dir>\opl\examples\opl

where <Install_dir> is your installation directory.

Changing CPLEX parameters

Any CPLEX parameter can be set from a script statement in an execute block. In case of conflict, if a different value has been set from the IDE for the same parameter, the value set in the script statement takes precedence.

The following code extract shows how to switch off CPLEX presolve and enable simplex logging in the model.

Changing CPLEX parameters via scripting (product.mod)

execute CPX_PARAM {
  cplex.preind = 0;   
  cplex.simdisplay = 2;   
}

In the following code extract, the script named PARAMS sets a time limit on each call to the optimizer.

Preprocessing script statement setting a parameter  (transp4.mod)

execute PARAMS {
  cplex.tilim = 100;
}

You can find the product.mod and transp4.mod models at the following location respectively:

<Install_dir>\opl\examples\opl\production

<Install_dir>\opl\examples\opl\transp

where <Install_dir> is your installation directory.

CPLEX solution status

This table lists the status values for solutions to LP, QP, or MIP problems. The status value is returned by IloCplex.status or IloCplex.getCplexStatus.

Code number Solution status
1 Optimal solution is available.
2 Problem has an unbounded ray.
3 Problem has been proven infeasible.
4 Problem has been proven either infeasible or unbounded.
5 Optimal solution is available, but with infeasibilities after unscaling.
6 Solution is available, but not proved optimal, due to numeric difficulties during optimization.
10 Stopped due to limit on number of iterations.
11 Stopped due to a time limit.
12 Stopped due to an objective limit.
13 Stopped due to a request from the user.
14 This status occurs only with the parameter feasoptmode set to 0 on a continuous problem. A relaxation was successfully found and a feasible solution for the problem. (if relaxed according to that relaxation) was installed. The relaxation is minimal.
15 This status occurs only with the parameter feasoptmode set to 1 on a continuous problem. A relaxation was successfully found and a feasible solution for the problem. (if relaxed according to that relaxation) was installed. The relaxation is optimal.
16 This status occurs only with the parameter feasoptmode set to 2 on a continuous problem. A relaxation was successfully found and a feasible solution for the problem. (if relaxed according to that relaxation) was installed. The relaxation is minimal.
17 This status occurs only with the parameter feasoptmode set to 3 on a continuous problem. A relaxation was successfully found and a feasible solution for the problem. (if relaxed according to that relaxation) was installed. The relaxation is optimal.
18 This status occurs only with the parameter feasoptmode set to 4 on a continuous problem. A relaxation was successfully found and a feasible solution for the problem. (if relaxed according to that relaxation) was installed. The relaxation is minimal.
19 This status occurs only with the parameter feasoptmode set to 5 on a continuous problem. A relaxation was successfully found and a feasible solution for the problem. (if relaxed according to that relaxation) was installed. The relaxation is optimal.
20 Model has an unbounded optimal face.
21 Stopped due to a limit on the primal objective.
22 Stopped due to a limit on the dual objective.
23 The problem under consideration was found to be feasible after phase 1 of FeasOpt. A feasible solution is available.
24 Solution satisfies first-order optimality conditions for a solution returned by the barrier optimizer.
25 Stopped due to a deterministic time limit.
30 The problem appears to be feasible; no conflict is available.
31 The conflict refiner found a minimal conflict.
32 The conflict refiner concluded contradictory feasibility for the same set of constraints due to numeric problems. A conflict is available, but it is not minimal.
33 The conflict refiner terminated because of a time limit. A conflict is available, but it is not minimal.
34 The conflict refiner terminated because of an iteration limit. A conflict is available, but it is not minimal.
35 The conflict refiner terminated because of a node limit. A conflict is available, but it is not minimal.
36 The conflict refiner terminated because of an objective limit. A conflict is available, but it is not minimal.
37 The conflict refiner terminated because of a memory limit. A conflict is available, but it is not minimal.
38 The conflict refiner terminated because a user terminated the application. A conflict is available, but it is not minimal.
39 The conflict refiner terminated because of a deterministic time limit. A conflict is available, but it is not minimal.
101 Optimal integer solution has been found.
102 Optimal solution with the tolerance defined by epgap or epagap has been found.
103 Solution is integer infeasible
104 The limit on mixed integer solutions has been reached.
105 Node limit has been exceeded but integer solution exists.
106 Node limit has been reached; no integer solution.
107 Time limit exceeded, but integer solution exists.
108 Time limit exceeded; no integer solution.
109 Terminated because of an error, but integer solution exists.
110 Terminated because of an error; no integer solution.
111 Limit on tree memory has been reached, but an integer solution exists.
112 Limit on tree memory has been reached; no integer solution.
113 Stopped, but an integer solution exists.
114 Stopped; no integer solution.
115 Problem is optimal with unscaled infeasibilities.
116 Out of memory, no tree available, integer solution exists.
117 Out of memory, no tree available, no integer solution.
118 Problem has an unbounded ray.
119 Problem has been proved either infeasible or unbounded.
120 This status occurs only with the parameter feasoptmode set to 0 on a mixed integer problem. A relaxation was successfully found and a feasible solution for the problem (if relaxed according to that relaxation) was installed. The relaxation is minimal.
121 This status occurs only with the parameter feasoptmode set to 1 on a mixed integer problem. A relaxation was successfully found and a feasible solution for the problem (if relaxed according to that relaxation) was installed. The relaxation is optimal.
122 This status occurs only with the parameter feasoptmode set to 2 on a mixed integer problem. A relaxation was successfully found and a feasible solution for the problem (if relaxed according to that relaxation) was installed. The relaxation is minimal.
123 This status occurs only with the parameter feasoptmode set to 3 on a mixed integer problem. A relaxation was successfully found and a feasible solution for the problem (if relaxed according to that relaxation) was installed. The relaxation is optimal.
124 This status occurs only with the parameter feasoptmode set to 4 on a mixed integer problem. A relaxation was successfully found and a feasible solution for the problem (if relaxed according to that relaxation) was installed. The relaxation is minimal.
125 This status occurs only with the parameter feasoptmode set to 5 on a mixed integer problem. A relaxation was successfully found and a feasible solution for the problem (if relaxed according to that relaxation) was installed. The relaxation is optimal.
126

This status occurs only after a call to feasOpt, when the algorithm terminates prematurely, for example after reaching a limit.

This status means that a relaxed solution is available and can be queried.

127 The problem under consideration was found to be feasible after phase 1 of FeasOpt. A feasible solution is available. This status is also used in the status field of solution and mipstart files for solutions from the solution pool.
128 This status occurs only after a call to the method populate on a MIP problem. The limit on mixed integer solutions generated by populate, as specified by the parameter populatelim, has been reached.
129 This status occurs only after a call to the method populate on a MIP problem. Populate has completed the enumeration of all solutions it could enumerate.
130 This status occurs only after a call to the method populate on a MIP problem. Populate has completed the enumeration of all solutions it could enumerate whose objective value fit the tolerance specified by the parameters solnpoolagap and solnpoolgap.
131 Deterministic time limit exceeded, but integer solution exists.
132 Deterministic time limit exceeded; no integer solution.
133 Could not bound convex relaxation of nonconvex (MI)QP.

For more information

  • Details of the CPLEX solution statuses can be found in Solution Status Symbols in the CPLEX Callable Library (C API).
  • See the description of class IloCplex in the IBM ILOG Script Reference Manual.
  • You can also find the complete reference documentation of CPLEX parameters in the CPLEX documentation, Parameters of CPLEX.

Changing CP parameters

You can set any constraint programming parameter from a script statement in an execute block. In case of conflict, if a different value has been set from the IDE for the same parameter, the value set in the script statement takes precedence.

The following code extract extends the logPeriod parameter to 10000, sets the search type to DepthFirst and the time limit to 600.

Changing CP parameters via scripting (timetabling.mod)

   var p = cp.param;
   p.logPeriod = 10000;
   p.searchType = "DepthFirst";
   p.timeLimit = 600;

The timetabling example is available at the following location:

<Install_dir>\opl\examples\opl\timetabling

where <Install_dir> is your installation directory.

For more information

See the description of class IloCP in the IBM ILOG Script Reference Manual. You can also find the complete reference documentation of CP parameters in the CP Optimizer documentation.

In the steel mill example, the solution is found very quickly. However, if you want to illustrate the engine log, you can decrease the periodicity (that is, the number of branches between which a line of log is printed). To do so, write:

execute {
        cp.param.Workers = 1;
        cp.param.LogPeriod = 50;
}

The general syntax to change engine parameters is:

 execute {cp.param.param_name = param_value
}

Changing OPL settings

You can set certain OPL settings from a script statement in an execute block. Not all OPL parameters can be set by scripting; you can change only the parameters that are listed as properties of IloOplSettings in the IBM ILOG Script Reference Manual. In case of conflict, if a different value has been set from the IDE for the same parameter, the value set in the script statement takes precedence. For an example, see Executing preprocessing scripts in Using IBM ILOG Script for OPL.

For more information

See the description of class IloOplSettings in the IBM ILOG Script Reference Manual.