naginterfaces.library.opt.handle_solve_dfno_rcomm¶
- naginterfaces.library.opt.handle_solve_dfno_rcomm(handle, irevcm, x, f, io_manager=None)[source]¶
handle_solve_dfno_rcomm
is a reverse communication Derivative-free Optimization (DFO) solver from the NAG optimization modelling suite (DFNO) for small to medium-scale nonlinear problems with bound constraints.Note: this function uses optional algorithmic parameters, see also:
handle_opt_set()
,handle_opt_get()
.For full information please refer to the NAG Library document for e04je
https://support.nag.com/numeric/nl/nagdoc_30.2/flhtml/e04/e04jef.html
- Parameters
- handleHandle
The handle to the problem. It needs to be initialized (e.g., by
handle_init()
) and to hold a problem formulation compatible withhandle_solve_dfno_rcomm
. It must not be changed between calls to the NAG optimization modelling suite.- irevcmint
Does not need to be set on the first call of
handle_solve_dfno_rcomm
. On subsequent calls, must be set to a positive integer if all the required function evaluations have been correctly provided in . Otherwise, if a problem occurred during a monitoring step or while providing objective values, it is possible to set it to a negative value:If function evaluations were required, the solver will attempt a rescue procedure and request an alternative point. If no function were required (monitoring step), the solver will stop with = 20.
The solver will cleanly exit and return the best available point as a well as the solve statistics.
- xfloat, array-like, shape
The first column contains , the initial estimates of the variables .
- ffloat, array-like, shape
Does not need to be set on the first call to
handle_solve_dfno_rcomm
.If after the last call of
handle_solve_dfno_rcomm
, the first elements must contain the function values of the requested points.- io_managerFileObjManager, optional
Manager for I/O in this routine.
- Returns
- irevcmint
Indicates what action is to be performed before the next call to
handle_solve_dfno_rcomm
.Final exit of the solver.
objective evaluations are required.
Monitoring step, no evaluation is required, and contain the best evaluation of the objective yet.
- nevalint
Indicates the number of objective evaluations required for the next call of
handle_solve_dfno_rcomm
in . The coordinates of the points to evaluate are provided in the first columns of .- xfloat, ndarray, shape
If or , the first column contains the best computed estimate of the solution.
If , the first columns contain the coordinates of the points to evaluate.
- ffloat, ndarray, shape
If or , the first element contains the function value of the best computed point.
- rinfofloat, ndarray, shape
Optimal objective value and various indicators at monitoring steps or at the end of the final iteration. The measures are given in the table below:
Objective function value .
, the current lower bound of the trust region.
, the current size of the trust region.
The number of interpolation points used by the solver.
–
Reserved for future use.
- statsfloat, ndarray, shape
Solver statistics at monitoring steps or at the end of the final iteration as given in the table below:
Number of calls to the objective function.
Total time spent in the solver (including time spent evaluating the objective).
Total time spent evaluating the objective function.
Number of steps.
–
Reserved for future use.
- Other Parameters
- ‘Defaults’valueless
This special keyword may be used to reset all options to their default values. Any value given with this keyword will be ignored.
- ‘DFNO Detect Unbounded’str
Default
The solver can try to detect whether the problem is unbounded. This option allows you to turn on or off the unboundedness detection heuristic.
Constraint: or .
- ‘DFNO Objective Limit’float
Default
This option sets an additional convergence criterion. The solver will stop if it finds a point for which the function value is lower than this parameter.
- ‘DFO Maximum Slow Steps’int
Default
If , this argument defines the maximum number of consecutive slow iterations allowed. Set to deactivate the slow iteration detection. The algorithm can stop in two situations:
and with = 50,
with = 24.
Constraint: .
- ‘DFO Max Objective Calls’int
Default
A limit on the number of objective function evaluations the solver is allowed to compute. If the limit is reached, the solver stops with = 21.
Constraint: .
- ‘DFO Max Soft Restarts’int
Default
The maximum total number of soft restarts that can be performed if the objective function is declared as noisy ().
Constraint: .
- ‘DFO Max Unsucc Soft Restarts’int
Default
The maximum number of consecutive unsuccessful soft restarts that can be performed if the objective function is declared as noisy ().
Constraint: .
- ‘DFO Monitor Frequency’int
Default
If , the solver will stop at the end of every th step for monitoring purposes.
handle_solve_dfno_rcomm
needs to be called again to continue the optimization.Constraint: .
- ‘DFO Noise Level’float
Default
Indicates the noise level expected when evaluating the objective function. If .
Constraint: .
- ‘DFO Noisy Problem’str
Default
Indicates if the function evaluations provided to the solver are noisy. If , some algorithmic features will be activated:
The trust region update becomes slower to reflect the decreased confidence in the objective values.
Soft restarts of the algorithm can be performed to ensure the algorithm did not get stuck because of the noise (see ‘DFO Max Soft Restarts’, ‘DFO Max Unsucc Soft Restarts’, ‘DFO Number Soft Restarts Pts’ to control the restart characteristics).
In addition, if , the solver will trigger a soft restart if all the function values are within the noise level.
- ‘DFO Number Interp Points’int
Default
The maximum number of interpolation points in (9) used to build the linear models of the residuals. If , the number of points is chosen to be where is the number of non-fixed variables.
Constraint: .
Consistency constraint, the solver stops with = 6 if not met:
.
- ‘DFO Number Soft Restarts Pts’int
Default
The number of interpolation points that are replaced during a soft restart.
Constraint: .
- ‘DFO Print Frequency’int
Default
If , the solver prints the iteration log to the appropriate units at the end of every th step.
Constraint: .
- ‘DFO Random Seed’int
Default
The random seed used to generate the random points used to build the initial model. If , the random seed will be based on values taken from the real-time clock, potentially resulting in the solver taking a different path each time it is run. Set it to a positive value to get fully reproducible runs.
Constraint: .
- ‘DFO Starting Trust Region’float
Default
, the initial trust region radius. This argument should be set to about one tenth of the greatest expected overall change to a variable: the initial quadratic model will be constructed by taking steps from the initial of length along each coordinate direction. The default value assumes that the variables have an order of magnitude .
Constraint: .
Consistency constraints, the solver stops with = 5 if not met:
.
.
- ‘DFO Trust Region Slow Tol’float
Default
The minimal acceptable trust region radius for the solution to be declared as acceptable. The solver stops if:
and .
Constraint: .
Consistency constraint, the solver stops with = 5 if not met:
.
- ‘DFO Trust Region Tolerance’float
Default
, the requested trust region radius. The algorithm declares convergence when the trust region radius reaches this limit. It should indicate the absolute accuracy that is required in the final values of the variables.
Constraint: .
Consistency constraint, the solver stops with = 5 if not met:
.
- ‘Infinite Bound Size’float
Default
This defines the ‘infinite’ bound in the definition of the problem constraints. Any upper bound greater than or equal to will be regarded as (and similarly any lower bound less than or equal to will be regarded as ). Note that a modification of this option does not influence constraints which have already been defined; only the constraints formulated after the change will be affected.
Constraint: .
- ‘Monitoring File’int
Default
If , the unit number for the secondary (monitoring) output. If , no secondary output is provided. The information output to this unit is controlled by ‘Monitoring Level’.
Constraint: .
- ‘Monitoring Level’int
Default
This argument sets the amount of information detail that will be printed by the solver to the secondary output. The meaning of the levels is the same as with ‘Print Level’.
Constraint: .
- ‘Print File’int
Default
If , the unit number for the primary output of the solver. If , the primary output is completely turned off independently of other settings. The default value is the advisory message unit number at the time of the options initialization, e.g., at the initialization of the handle. The information output to this unit is controlled by ‘Print Level’.
Constraint: .
- ‘Print Level’int
Default
This argument defines how detailed information should be printed by the solver to the primary and secondary output.
Output
No output from the solver.
The Header and Summary.
, , ,
Additionally, the Iteration log.
Constraint: .
- ‘Print Options’str
Default
If , a listing of options will be printed to the primary output and is always printed to the secondary output.
Constraint: or .
- ‘Print Solution’str
Default
If , the solution will be printed to the primary and secondary output.
Constraint: or .
- ‘Task’str
Default
This argument specifies the required direction of the optimization. If , the objective function (if set) is ignored and the algorithm stops as soon as a feasible point is found with respect to the given tolerance.
Constraint: , or .
- ‘Stats Time’str
Default
This argument turns on timings of various parts of the algorithm to give a better overview of where most of the time is spent. This might be helpful for a choice of different solving approaches. It is possible to choose between CPU and wall clock time. Choice ‘YES’ is equivalent to ‘WALL CLOCK’.
Constraint: , , or .
- ‘Time Limit’float
Default
A limit to the number of seconds that the solver can use to solve one problem. If during the convergence check this limit is exceeded, the solver will terminate with = 23.
Constraint: .
- Raises
- NagValueError
- (errno )
has not been initialized.
- (errno )
does not belong to the NAG optimization modelling suite, has not been initialized properly or is corrupted.
- (errno )
has not been initialized properly or is corrupted.
- (errno )
This solver does not support the model defined in the handle.
- (errno )
The problem is already being solved.
- (errno )
On entry, .
Constraint: .
- (errno )
On entry, , expected .
Constraint: must match the current number of variables of the model in the .
- (errno )
The information supplied does not match with that previously stored.
On entry, must match that given during the first call of the function, i.e., .
- (errno )
Inconsistent options ‘DFO Trust Region Tolerance’ and ‘DFO Starting Trust Region’ .
Constraint: .
Use
handle_opt_set()
to set compatible option values.- (errno )
Inconsistent options ‘DFO Trust Region Tolerance’ and ‘DFO Trust Region Slow Tol’ .
Constraint: .
Use
handle_opt_set()
to set compatible option values.- (errno )
Option ‘DFO Starting Trust Region’ , , and .
Constraint: if in coordinate , then .
Use
handle_opt_set()
to set compatible option values.- (errno )
There were unequal bounds and the option ‘DFO Number Interp Points’ .
Constraint: .
Use
handle_opt_set()
to set compatible option values.- (errno )
The number of initial interpolation points is different from the total set by ‘DFO Number Interp Points’.
Growing the interpolation set is not supported for this solver.
- Warns
- NagAlgorithmicWarning
- (errno )
The problem was solved to an acceptable level after consecutive slow iterations.
Use the option ‘DFO Maximum Slow Steps’ to modify the maximum number of slow steps accepted.
- NagAlgorithmicMajorWarning
- (errno )
Rescue failed: the trust region could not be reduced further after some function evaluation could not be provided. Check the specification of your objective and whether it needs rescaling. Try a different initial .
- (errno )
Some initial interpolation points could not be provided. Rescue cannot be attempted at this stage.
Check the specification of your objective and whether it needs rescaling. Try a different initial .
- (errno )
The predicted reduction in a trust region step was non-positive. Check the specification of your objective and whether it needs rescaling. Try a different initial .
- (errno )
The solver failed at the model building phase and the maximum number of restarts was reached. Check the specification of your objective and whether it needs rescaling. Try a different initial .
- (errno )
Maximum number of function evaluations exceeded.
- (errno )
The solver terminated after the maximum time allowed was exceeded.
- (errno )
No progress, the solver was stopped after consecutive slow steps.
Use the option ‘DFO Maximum Slow Steps’ to modify the maximum number of slow steps accepted.
- (errno )
The problem seems to be unbounded.
The unboundedness detection heuristic can be turned off with the option ‘DFNO Detect Unbounded’.
- NagCallbackTerminateWarning
- (errno )
User requested termination during a monitoring step.
- (errno )
User requested termination during an objective evaluation step.
- Notes
handle_solve_dfno_rcomm
uses reverse communication for function evaluations and monitoring steps. Every time the solver requires an evaluation of the objective function, it pauses its progress, exits and waits for the function to be called again with the objective value provided in the argument .handle_solve_dfno_rcomm
is aimed at minimizing a nonlinear objective function subject to bound constraints:Here is a smooth nonlinear function and and are -dimensional vectors defining bounds on the variables.
handle_solve_dfno_rcomm
serves as a solver for compatible problems stored as a handle. The handle points to an internal data structure which defines the problem and serves as a means of communication for functions in the NAG optimization modelling suite. To define a compatible problem handle, you must callhandle_init()
followed byhandle_set_nlnobj()
to initialize it and optionally callhandle_set_simplebounds()
to define bounds on the variables. Ifhandle_set_simplebounds()
is not called, all the variables will be considered free by the solver. It should be noted thathandle_solve_dfno_rcomm
always assumes that the gradient of the objective is dense, therefore, defining a sparse structure for the residuals in the call tohandle_set_nlnobj()
will have no effect. See the E04 Introduction for more details about the NAG optimization modelling suite.The solver allows fixing variables with the definition of the bounds. However, the following constraint must be met in order to be able to call the solver:
for all non-fixed variable , the value of must be at least twice the starting trust region radius (see the consistency constraint of the option ‘DFO Starting Trust Region’).
The solver is based on a derivative-free trust region framework. This type of method is well suited for small to medium-scale problems (around 100 variables) for which the derivatives are unavailable or not easy to compute, and/or for which the function evaluations are expensive or noisy. For a detailed description of the algorithm see Algorithmic Details.
The algorithm behaviour and solver strategy can be modified by various options (see Other Parameters) which can be set by
handle_opt_set()
andhandle_opt_set_file()
at any time between the initialization of the handle byhandle_init()
and a call to the solver. The options’ names specific for this solver start either with the prefix DFO (Derivative-free Optimization) or DFNO (Derivative-free Nonlinear Optimization). The default values for these options are chosen to work well in the general case, but it is recommended you tune them to your particular problem. In particular, if the objective function is known to be noisy, it is highly recommended to set the option ‘DFO Noisy Problem’ to ‘YES’. Once the solver has finished, options may be modified for the next solve. The solver may be called repeatedly with various starting points and/or options.The underlying algorithm implemented for
handle_solve_dfno_rcomm
is the same as the one used byhandle_solve_dfno()
.handle_solve_dfno_rcomm
serves as a reverse communication interface to the derivative-free solver for nonlinear objective functions.
- References
Cartis, C, Fiala, J, Marteau, B and Roberts, L, 2018, Improving the Flexibility and Robustness of Model-Based Derivative-Free Optimization Solvers, Technical Report, University of Oxford
Conn, A R, Scheinberg, K and Vicente, L N, 2009, Introduction to Derivative-Free Optimization, vol. 8 of MPS-SIAM Series on Optimization, MPS/SIAM, Philadelphia
Powell, M J D, 2009, The BOBYQA algorithm for bound constrained optimization without derivatives, Report DAMTP 2009/NA06, University of Cambridge, https://www.damtp.cam.ac.uk/user/na/NA_papers/NA2009_06.pdf