rbatools.rba_lp API documentation

rbatools.rba_lp module

Classes

class LinearProblem (ancestors: ProblemMatrix)

Attributes

A : scipy.sparse.coo_matrix
Lefthandside of constraint Matrix (aka Constraint Matrix)
b : numpy.array
Righthandside of Constraint Matrix
row_signs : list
Type of constraints ('E' for equality 'L' for lower-or-equal inequality) –> Ax=b or Ax<=b
f : numyp.array
Objective function linear get_coefficients
LB : numpy.array
Lower bounds of decision-variables
UB : numpy.array
Upper bounds of decision-variables
row_names : list
Names of constraints
col_names : list
Names of decision-variables
rowIndicesMap : dict
Dictionary mapping constraint names to their numeric index (generated automatically)
colIndicesMap : dict
Dictionary mapping variable names to their numeric index (generated automatically)
lp_solver : str
Selected linear problem solver.
_lp_solver : rbatools.rba_lp._Solver (any descendent class) object
 

Methods

def __init__(self, *matrix, lp_solver='cplex')

Initiates LP object. If provided with an rbatools.rba_problem_matrix.ProblemMatrix object as argument, assumes its attributes. If not provided with an ProblemMatrix object, initiates as empty LP object. Parameters

matrix : ProblemMatrix
Optional
lp_solver : str
Specifies which LP-solver should be used ('cplex' or 'swiglpk') Default: 'cplex'
def add_matrix(self, matrix)

Merges the Problem with the one provided as input-argument to this method.

Matrix elements unique to input-matrix are added. Elements occuring in both matrices are overwritten with the value from new matrix. Generates CPLEX problem from merged problem.

Parameters

matrix : ProblemMatrix
The matrix with elements to be added
def build_lp(self)

Constructs an LP-object, based on solver attribute

Returns

bool : Boolean, indicating whether LP-object could be constructed
 
def get_constraint_types(self, constraints=[])

Returns constraint types for specified rows. (E: = ; L: <= ; G:>=)

Parameters

constraints : list of row names.

Returns

dict
Dictionary with constraint-IDs as keys and type identification-characters as values.
def get_lb(self, variables)

Returns lower bounds for specified columns.

Parameters

variables : list of column names.

Returns

dict
Dictionary with variable-IDs as keys and lower bounds as values.
def get_objective(self, variables=[])

Returns objective coefficients for specified columns.

Parameters

variables : list of column names.

Returns

dict
Dictionary with variable-IDs as keys and objective coefficients as values.
def get_problem_coefficients(self, inputTuples=[])

Returns problem coefficints of constraint matrix.

Parameters

inputTuples : list of coordinate tuples of coefficients (row,col).

Returns

dict
Dictionary with index tuples as keys and matrix coefficients as values.
def get_right_hand_side(self, constraints=[])

Returns RHS values for specified rows.

Parameters

constraints : list of row names.

Returns

dict
Dictionary with constraint-IDs as keys and RHS-values as values.
def get_ub(self, variables)

Returns upper bounds for specified columns.

Parameters

variables : list of column names.

Returns

dict
Dictionary with variable-IDs as keys and upper bounds as values.
def initiate_lp_solver(self)
def return_dual_values(self)

Returns (dual) solution vector.

Returns

dict : Problem constraints as keys and dual values (shadow prices) as values.

def return_objective_value(self)

Returns objective value, after solving the problem.

Returns

float : objective value
 
def return_primal_values(self)

Returns (primal)solution vector.

Returns

dict : Problem variables as keys and solution values as values.

def return_solution_status(self)

Returns solution status, after solving the problem.

Returns

str : solution status
optimal, feasible or infeasible solution stati returned by solvers: optimal: swiglpk -> GLP_OPT cplex -> 1 feasible: swiglpk -> GLP_FEAS feasible_only_before_unscaling: cplex -> 5
def set_constraint_types(self, inputDict)

Sets the type of specified constraints. (E: = ; L: <= ; G:>=)

Parameters

inputDict : dict
Constraint IDs as keys and constraint type as values
def set_lb(self, inputDict)

Sets variable lower bounds to specified values.

Parameters

inputDict : dict
Variable IDs as keys and lower bounds as values
def set_objective(self, inputDict)

Sets coefficients in objective function to specified values.

Parameters

inputDict : dict
Variable IDs as keys and objective coefficient as values
def set_problem_coefficients(self, input)

Sets coefficients in constraint matrix.

Parameters

inputDict : dict
Dictionary with index tuples as keys and matrix coefficients as values.
def set_right_hand_side(self, inputDict)

Sets the rhs-value of specified constraints.

Parameters

inputDict : dict
Constraint IDs as keys and rhs-values as values.
def set_ub(self, inputDict)

Sets variable upper bounds to specified values.

Parameters

inputDict : dict
Variable IDs as keys and upper bounds as values
def solve_lp(self)

solves linear problem

def unscale_lp(self)

Unscale LP

def update_matrix(self, matrix, Ainds=None, Binds=None, CTinds=None, LBinds=None, UBinds=None, ModifiedProblem=True)

Overwrites coefficients with new values from argument 'matrix'.

Parameters

matrix : ProblemMatrix
The matrix with elements to be added
Ainds : list of tuples
List of index-pair tuples [(row_coeff1,col_coeff1),(row_coeff2,col_coeff2),…], specifying which elements are updated. Default is None (then all elements of argument matrix (present in both matrices) are taken to update)
Binds : list
List of constraint-IDs for indices of the RHS, specifying which RHS Values are updated. Default is None (then all constraints (present in both matrices) are taken to update)
CTinds : list
List of constraint-IDs, specifying which row_signs are updated. Default is None (then all constraint types (present in both matrices) are taken to update)
LBinds : list
List of variable-IDs, specifying which lower-bounds values are updated. Default is None (then all variables (present in both matrices) are taken to update)
UBinds : list
List of variable-IDs, specifying which upper-bounds values are updated. Default is None (then all variables (present in both matrices) are taken to update)
ModifiedProblem : boolean
Whether problem has been modifed. If false own lp components are fully replaced by input, if rows and col names are identical .

Inherited members