org.ojalgo.optimisation

Class Optimisation.Options

java.lang.Object

org.ojalgo.optimisation.Optimisation.Options

All Implemented Interfaces:

Optimisation

Enclosing interface:

Optimisation

public static final class Optimisation.Options
extends Object
implements Optimisation

Nested Class Summary

Nested classes/interfaces inherited from interface org.ojalgo.optimisation.Optimisation

Optimisation.Constraint, Optimisation.Integration<M extends Optimisation.Model,S extends Optimisation.Solver>, Optimisation.Model, Optimisation.Objective, Optimisation.Options, Optimisation.Result, Optimisation.Solver, Optimisation.State

Field Summary

Fields

Modifier and Type	Field and Description
NumberContext	feasibility Used to determine/validate feasibility.
int	iterations_abort The maximmum number of iterations allowed for the solve() command.
int	iterations_suffice Calculations will be terminated after this number of iterations if a feasible solution has been found.
BasicLogger.Printer	logger_appender If this is null nothing is printed, if it is not null then progress/debug messages are printed to that BasicLogger.Printer.
boolean	logger_detailed Detailed (debug) logging or not.
Class<? extends Optimisation.Solver>	logger_solver Which Optimisation.Solver to debug.
double	mip_defer The branch-and-bound nodes/subproblems come in pairs, and each node has a displacement (the fractional amount removed by the new bound).
double	mip_gap The MIP gap is the difference between the best integer solution found so far and a node's non-integer solution.
NumberContext	print For display only!
NumberContext	solution Describes the (required/sufficient) accuracy of the solution.
Boolean	sparse Controls if sparse/iterative solvers should be favoured over dense/direct alternatives.
long	time_abort The maximmum number of millis allowed for the solve() command.
long	time_suffice Calculations will be terminated after this amount of time if a feasible solution has been found.
boolean	validate If true models and solvers will validate data at various points.

Constructor Summary

Constructors

Constructor and Description
Options()

Method Summary

Modifier and Type	Method and Description
Optimisation.Options	abort(CalendarDateDuration duration)
void	debug(Class<? extends Optimisation.Solver> solver) Will configure detailed debug logging and validation
<T> Optional<T>	getConfigurator(Class<T> type)
void	progress(Class<? extends Optimisation.Solver> solver) Will configure high level (low volume) progress logging
void	setConfigurator(Object configurator)
Optimisation.Options	suffice(CalendarDateDuration duration)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

feasibility

public NumberContext feasibility

Used to determine/validate feasibility. Are the constraints violated or not? Are the variable values integer or not?

iterations_abort

public int iterations_abort

The maximmum number of iterations allowed for the solve() command.

iterations_suffice

public int iterations_suffice

Calculations will be terminated after this number of iterations if a feasible solution has been found. If no feasible solution has been found calculations will continue until one is found or iterations_abort is reached. This option is, probably, only of interest with the IntegerSolver.

logger_appender

public BasicLogger.Printer logger_appender

If this is null nothing is printed, if it is not null then progress/debug messages are printed to that BasicLogger.Printer.

logger_detailed

public boolean logger_detailed

Detailed (debug) logging or not.

logger_solver

public Class<? extends Optimisation.Solver> logger_solver

Which Optimisation.Solver to debug. Null means NO solvers. This setting is only relevant if logger_appender has been set.

mip_defer

public double mip_defer

The branch-and-bound nodes/subproblems come in pairs, and each node has a displacement (the fractional amount removed by the new bound). Every pair's total displacement is always exactly 1.0. From each pair the node with the smallest displacement (0.0 - 0.5) is always evaluated directly. The other node (with displacement 0.5 - 1.0) can either be forked off and evalueated in another thread, or deferred to be evaluated later in order of creation. Nodes with displacement larger than this value are deferred. If this value is set to 0.5 or less then everything is deferred, effectively making the MIP solver single threaded and deterministic. If this value is set to 1.0 or more then nothinng is deferred

mip_gap

public double mip_gap

The MIP gap is the difference between the best integer solution found so far and a node's non-integer solution. The relative MIP gap is that difference divided by the optimal value (approximated by the currently best integer solution). If the gap (absolute or relative) is smaller than this value, then the corresponding branch is terminated as it is deemed unlikely or too "expensive" to find better integer solutions there.

print

public NumberContext print

For display only!

solution

public NumberContext solution

Describes the (required/sufficient) accuracy of the solution. It is used when copying the solver's solution back to the model (converting from double to BigDecimal). Specific solvers may also use this as a stopping criteria or similar. The default essentially copies the numbers as is – corresponding to full double precision.

sparse

public Boolean sparse

Controls if sparse/iterative solvers should be favoured over dense/direct alternatives. Sparse/iterative alternatives are usually preferable with larger models, but there are also algorithmical differences that could make one alternative better than the other for a (your) specific case. There are 3 different possibilities for this option:

1. TRUE Will use the sparse linear solver and the iterative convex solver.
2. FALSE Will use the dense linear solver and the direct convex solver.
3. NULL ojAlgo will use some logic to choose for you. This is the default.

time_abort

public long time_abort

The maximmum number of millis allowed for the solve() command. Executions will be aborted regardless of if a solution has been found or not.

time_suffice

public long time_suffice

Calculations will be terminated after this amount of time if a feasible solution has been found. If no feasible solution has been found calculations will continue until one is found or time_abort is reached. This option is , probably, only of interest with the IntegerSolver.

validate

public boolean validate

If true models and solvers will validate data at various points. Validation is turned off by default. Turning it on will significantly slow down execution - even very expensive validation may be performed.

Constructor Detail

Options

public Options()

Method Detail

abort

public Optimisation.Options abort(CalendarDateDuration duration)

debug

public void debug(Class<? extends Optimisation.Solver> solver)

Will configure detailed debug logging and validation

getConfigurator

public <T> Optional<T> getConfigurator(Class<T> type)

progress

public void progress(Class<? extends Optimisation.Solver> solver)

Will configure high level (low volume) progress logging

setConfigurator

public void setConfigurator(Object configurator)

suffice

public Optimisation.Options suffice(CalendarDateDuration duration)