Load data to the problem object (extending class
'>optObj). Use this method to generate problem
objects.
# S4 method for optObj_clpAPI
loadLPprob(lp,
nCols, nRows, mat, ub, lb, obj, rlb, rtype,
lpdir = "max", rub = NULL, ctype = NULL,
cnames = NULL, rnames = NULL, pname = NULL,
defLowerBnd = SYBIL_SETTINGS("MAXIMUM") * -1,
defUpperBnd = SYBIL_SETTINGS("MAXIMUM")
)# S4 method for optObj_cplexAPI
loadLPprob(lp,
nCols, nRows, mat, ub, lb, obj, rlb, rtype,
lpdir = "max", rub = NULL, ctype = NULL,
cnames = NULL, rnames = NULL, pname = NULL)
# S4 method for optObj_glpkAPI
loadLPprob(lp,
nCols, nRows, mat, ub, lb, obj, rlb, rtype,
lpdir = "max", rub = NULL, ctype = NULL,
cnames = NULL, rnames = NULL, pname = NULL)
# S4 method for optObj_lpSolveAPI
loadLPprob(lp,
nCols, nRows, mat, ub, lb, obj, rlb, rtype,
lpdir = "max", rub = NULL, ctype = NULL,
cnames = NULL, rnames = NULL, pname = NULL)
Number of columns (variables) of the constraint matrix.
Number of rows (constraints) of the constraint matrix.
An object of class Matrix. The constraint matrix of the
problem object. The number of columns in mat must be nCols
and the number of rows in mat must be nRows.
A numeric vector of length nCols giving the upper bounds of the
variables of the problem object.
A numeric vector of length nCols giving the lower bounds of the
variables of the problem object.
A numeric vector of length nCols giving the objective coefficients
of the variables of the problem object.
A numeric vector of length nRows giving the right hand side of the
problem object. If argument rub is not NULL, rlb
contains the lower bounds of the constraints of the problem object.
See Details.
A character vector of length nRows giving the constraint type:
"F": |
free constraint (GLPK only) | \(-\infty < x < \infty\) |
"L": |
constraint with lower bound | \(\mathrm{lb} \leq x < \infty\) |
"U": |
constraint with upper bound | \(-\infty < x \leq \mathrm{ub}\) |
"D": |
double-bounded (ranged) constraint | \(\mathrm{lb} \leq x \leq \mathrm{ub}\) |
"E": |
fixed (equality) constraint | \(\mathrm{lb} = x = \mathrm{ub}\) |
If rtype[i] is not one of "F", "L", "U",
"D" or "E", the value of rtype[i] will be set to
"E". See Details.
Single character string containing the direction of optimization.
Can be set to "min" or "max".
Default: "max".
A numeric vector of length nRows giving the right hand side of the
problem object. If not NULL, it contains the upper bounds of the
constraints of the problem object. See Details.
Default: NULL.
A character vector of length nCols giving the variable type. If set
to NULL, no specific variable type is set, which usually means, all
variables are treated as continuous variables.
Default: NULL.
"C": |
continuous variable |
"B": |
| binary variable |
"I": |
integer variable |
"S": |
semi-continuous variable |
"N": |
| semi-integer variable |
"C": |
continuous variable |
Values "S" and "N" are not available for every solver
software. Check documentation of the solver software if semi-continuous and
semi-integer variables are supported.
If ctype[j] is not "C", "B", "I", "S",
or "N", the value of ctype[j] will be set to "C".
A character vector of length nCols containing symbolic names for
the variable of the problem object.
Default: NULL.
A character vector of length nRows containing symbolic names for
the constraints of the problem object.
Default: NULL.
A single character string containing a name for the problem object.
Default: NULL.
signature(lp = "optObj_clpAPI")method to use with package clpAPI.
signature(lp = "optObj_cplexAPI")method to use with package cplexAPI.
signature(lp = "optObj_glpkAPI")method to use with package glpkAPI.
signature(lp = "optObj_lpSolveAPI")method to use with package lpSolveAPI.
Method loadLPprob can be used any time after a problem object is
initialized by initProb.
In order so set constraints, usually only parameter rlb is required
and parameter rub can be left at NULL (which is the default).
If rub is not NULL, rlb and rub must have the same
length. Parameter rub is required, if a particular constraint is a
ranged or double bounded constraint. The general idea is, for any constraint
i, the value in rlb[i] gives the lower bound and the value in
rub[i] gives the upper bound. If the constraints of the optimization
problem do only have one bound (type "L", "U" and "E"),
all bounds can be set via rlb and rub is not required. If any
constraint is of type "D" (a double-bounded or ranged constraint)
additionally rub is required. It is of course also possible to use
rlb strictly for all lower bounds and rub for all upper bounds.
Again, if both rlb and rub are given (not NULL), they
must have the same length. For equality constraints (type "E"), allways
the value in rlb is used.
For the '>optObj_cplexAPI method: CPLEX uses so called
ranged constraints for double bounded constraints. The values in rlb
and rub will be transformed into range values for ranged constraints.
The range for a ranged constraint \(i\) is given as
abs(rub[i] - rlb[i]), so that the valid interval is denoted as
[rlb[i], rlb[i] + range].
For the '>optObj_glpkAPI method: if cnames or
rnames is not NULL, an index will be created.
For the '>optObj_clpAPI method: if cnames is
not NULL, rnames must be also not NULL and vice versa.
For the '>optObj_lpSolveAPI method: if cnames is
not NULL, rnames must be also not NULL and vice versa.
Round brackets ("(" and ")") will be replaced by underscores
"_".
Superclass '>optObj and constructor function
optObj.