baronset
Create or edit a BARON options structure
Syntax
options = baronset('param1',value1,'param2',value2,...)
baronset
options = baronset
options = baronset(oldopts,'param1',value1,...)
Description
The function baronset creates an options structure that you can pass to baron for customizing the solving of an optimization problem.
options = baronset('param1',value1,'param2',value2,...) creates a BARON options structure in which the specified parameters (param) have the specified values (value).
baronset with no input or output arguments displays a complete list of parameters with their valid field names.
options = baronset with no input arguments creates a default options structure.
options = baronset(oldopts,'param1',value1,...) updates the existing options structure oldopts with the new values specified.
Options
The following tables list the available options for baronset and provide a brief description of these options. For more details of the options, see the BARON manual.
Termination options
| Parameter | Datatype | Description | Default |
| EpsA | double scalar | Absolute termination tolerance | 1e-6 |
| EpsR | double scalar | Relative termination tolerance | 1e-6 |
| DeltaTerm | integer scalar | Heuristic termination | 0 |
| DeltaT | integer scalar | Heuristic termination time | -100 |
| DeltaA | double scalar | Heuristic termination absolute threshold | 1e51 |
| DeltaR | double scalar | Heuristic termination relative threshold | 1 |
| CutOff | double scalar | Ignore solutions no better than this value | 1e51 |
| AbsConFeasTol | double scalar | Absolute constraint feasibility tolerance | 1e-5 |
| RelConFeasTol | double scalar | Relative constraint feasibility tolerance | 0 |
| AbsIntFeasTol | double scalar | Absolute integer feasibility tolerance | 1e-5 |
| RelIntFeasTol | double scalar | Relative integer feasibility tolerance | 0 |
| BoxTol | double scalar | Box elimination tolerance | 1e-8 |
| FirstFeas | integer scalar | Terminate on first feasible solution | 0 |
| FirstLoc | integer scalar | Terminate on first local solution | 0 |
| MaxIter | integer scalar | Maximum branch and reduce iterations | -1 |
| MaxTime | double scalar | Maximum time solving (applied to CPU time for single-threaded runs; and to wall clock time for multi-threaded runs) | 1000 s |
| WantDual | Integer scalar | Make an effort to find a dual solution | 1 |
| NumSol | integer scalar | Number of feasible solutions to find | 1 |
| IsolTol | double scalar | Separation distance between solutions | 1e-4 |
Relaxation options
| Parameter | Datatype | Description | Default |
| NOuter1 | integer scalar | Number of outer approximators of convex univariate functions | 4 |
| NOutPerVar | integer scalar | Number of outer approximators per variable for convex multivariable functions | 4 |
| NoutIter | integer scalar | Number of rounds of cutting plane generation at LP relaxation | 4 |
| OutGrid | integer scalar | Number of grid points per variable for convex multivariate approximators | 20 |
Range reduction options
| Parameter | Datatype | Description | Default |
| TDo | integer scalar | Nonlinear-feasibility-based range reduction (bound propagation) (Off 0, On 1) | 1 |
| MDo | integer scalar | Marginals-based range reduction (Off 0, On 1) | 1 |
| LBTTDo | integer scalar | Linear-feasibility-based range reduction (bound propagation) (Off 0, On 1) | 1 |
| OBTTDo | integer scalar | Optimality-based tightening (Off 0, On 1) | 1 |
| PDo | integer scalar | Number of probing problems allowed (Off 0, Auto -2, Number of probing variables: n) | -2 |
Tree management options
| Parameter | Datatype | Description | Default |
| BrVarStra | integer scalar | Branching variable selection strategy (BARON 0, Largest Violation 1, Longest Edge 2) | 0 |
| BrPtStra | integer scalar | Branching point selection strategy (BARON 0, w 1, Bisection 2, Convex combination of w & bisection 3) | 0 |
| BrVarPr* | integer vector | Branching variable priority (Unset NaN, Do not Branch 0, Higher number higher priority) | [] |
| NodeSel | integer scalar | Specifies the node selection rule for exploring the search tree (BARON 0, Best bound 1, LIFO 2, Minimum infeasibilities 3) | 0 |
Local search options
| Parameter | Datatype | Description | Default |
| DoLocal | integer scalar | Local search option for upper bounding (No local search 0, BARON rule 1) | 1 |
| NumLoc | integer scalar | Number of local searches done in preprocessing (Off 0, Until termination -1, BARON rule -2) | -2 |
Infeasibility diagnosis options
| Parameter | Datatype | Description | Default |
| CompIIS | integer scalar | Search for an IIS (No search 0, heuristic 1, deletion filtering 2, addition filtering 2, addition-deletion filtering 4, depth-first search, 5) | 0 |
| IISorder | integer scalar | Order in which constraints are considered for an IIS (Auto -1, problem order 1, ascending order of degree 2, descending order of degree 3, random 4) | -1 |
| IISint | integer scalar | Consider general integers as part of IIS (no 0, yes 1) | 0 |
Output options
| Parameter | Datatype | Description | Default |
| PrFreq | integer scalar | Log output frequency in number of nodes | 1e6 |
| PrTimeFreq | integer scalar | Log output frequency in number of seconds | 30 s |
| PrLevel | integer scalar | Option to control log output (Off 0, On 1) | 1 |
| LocRes | integer scalar | Option to control local search log (Off 0, On 1) | 0 |
Subsolver options
| Parameter | Datatype | Description | Default |
| LPSol | integer scalar | Specifies the LP solver to be used (Auto -1, CPLEX 3, CLP/CBC 8) | -1 |
| CplexLibName | string | Location of the CPLEX callable libraries on the user's computer. | 'libcplex.so' (Linux) 'cplex1200.dll' (Windows) 'libcplex.dylib' (OSX) |
| LPAlg | integer scalar | LP algorithm (Auto 0, Primal Simplex 1, Dual Simplex 2, Barrier 3) | 0 |
| NLPSol | integer scalar | NLP solver (None 0, Dynamic selection -1, BARON 1, IPOPT 9) | -1 |
| AllowIPOPT | integer scalar | Allow the use of IPOPT in dynamic local solver selection (Off 0, On 1) | 1 |
| AllowFilterSD | integer scalar | Allow the use of FilterSD in dynamic NLP solver selection (Off 0, On 1) | 1 |
| AllowFilterSQP | integer scalar | Allow the use of FilterSQP in dynamic NLP solver selection (Off 0, On 1) | 1 |
Other options
| Parameter | Datatype | Description | Default |
| ProbName* | string | Specify problem name (Default: 'BARON_Problem', Custom: any string) | 'BARON_Problem' |
| LicName | string | Specify name of BARON license file (Default: 'baronlice.txt', Custom: any string; must include path unless in execute directory) | 'baronlice.txt' |
| sense* | string | Objective sense (Minimization 'min', Maximization 'max') | 'min' |
| eqtype* | integer vector | Specify constraint equation types [a vector with length(constraints) elements] (Default 0, Convex 1, Relaxation Only 2) | [] |
| threads | integer scalar | Number of threads allowed in MIP subproblem solution (One 1, n n) | 1 |
| filekp* | integer scalar | Save BARON intermediate and output files (.bar, .dat) in current directory (Off 0, On 1) | 0 |
| optsfile* | string | Append (and overwrite) BARON options to this structure from an options file[1] (Off [], On 'filename') | [] |
| tracefile* | string | Generate post-solve tracefile in current directory (Off [], On 'filename') | [] |
| barscratch* | string | Temporary file directory for BARON intermediate files (Default [], User specified 'path') - Note must be a full path | [] |
| chkfun* | integer scalar | Check generated BARON equation(s) versus original MATLAB function(s) (Off 0, On 1) | 1 |
| custom* | string | Specify custom options that may be made available to certain users (Default [], User specified such as {'intopt1: 1', 'realopt2: 3.5'}) | [] |
*These are custom MATLAB/BARON interface options.
[1]An options file is a text file of the form <option name> <option value> with one option per line.