Skip to contents

Create constraint object to add custom constraints to the model.

Source: R/class-constraint.R
newConstraint.Rd

The function creates a new constraint object that can be used to add custom constraints to the model.

Usage

newConstraint(
  name,
  desc = "",
  ...,
  eq = "==",
  for.each = NULL,
  rhs = data.frame(),
  defVal = NULL,
  interpolation = "inter",
  replace_zerros = 1e-20
)

newConstraintS(
  name,
  type,
  eq = "==",
  rhs = 0,
  for.sum = list(),
  for.each = list(),
  defVal = 0,
  rule = NULL,
  comm = NULL,
  cout = TRUE,
  cinp = TRUE,
  aout = TRUE,
  ainp = TRUE
)

Arguments

name

character. Name of the constraint object, used in sets.

desc

character. Description of the constraint.

...

named or unnamed list(s) of left-hand side (LHS) linear terms (summands) to define the constraint. Every summand is defined as a list with the following elements:

  • variable - name of the variable in the summand.

  • mult - multiplier for the variable in the summand.

  • for.sum - list of sets for which the summand is defined. The summands can be passed as named or unnamed lists. They will be added to the lhs slot of the constraint object as linear terms of mulipliers and variables.

eq

Type of the relation ('==' default, '<=', '>=')

for.each

list or data.frame with sets that define the dimension of the constraint.

rhs

a numeric value, list or data frame with sets and numeric values for each constraint. Note: zero values will be replaced with replace_zerros to avoid dropping them by the interpolation algorithms.

defVal

numeric. The default value for the rhs. It is recommended to set the default value for the rhs of every constraint to avoid unexpected behavior. If not specified, the default value is 0, and the warning is issued.

interpolation

character. Interpolation rule for the constraint. Recognized values, any combination of "back", "inter", "forth", e.g., "back.inter" or "forth.inter", indicating the direction of interpolation. The default value is "inter", meaning that the interpolation is done for years between the specified values. The "back" and "forth" values induce backward and forward interpolation of the rhs values, respectively.

replace_zerros

numeric value to replace zero values in rhs and defVal. Default is 1e-20.

type

character; the constraint kind, which selects the variable the left-hand side is built from: a name containing "inp" or "out" constrains technology input or output, "capacity" / "newcapacity" the capacity variables, and "balance" the commodity balance. The two special values "tax" and "subs" instead delegate to newTax() / newSubsidy().

for.sum

named list of sets to SUM the left-hand side over, as opposed to for.each, whose sets index one constraint row each.

rule

accepted for backward compatibility and currently unused.

comm

commodity name; read only by the "tax" and "subs" types, which pass it on to newTax() / newSubsidy().

cout, aout

logical; for an output-type constraint, include the main (vTechOut) and the auxiliary (vTechAOut) output variables.

cinp, ainp

logical; for an input-type constraint, include the main (vTechInp) and the auxiliary (vTechAInp) input variables.

Value

Object of class constraint.

Details

Custom constraints extend the functionality of the model by adding user-defined constraints to the optimization problem. If the predefined constraints are not sufficient to describe the problem, custom constraints can be used to add linear equality or inequality constraints to define additional relationships between the variables. In many cases this can be done without writing constraints in the GAMS, Julia/JuMP, Python/Pyomo, or GLPK-MathProg languages by using the constrant class and the newConstraint function. To define a custom constraint with the newConstraint function, the user needs to specify the name of the constraint, the type of the relation (equality, less than or equal, greater than or equal), the left-hand side (LHS) terms of the statement, and the right-hand side (RHS) value. The dimension of the constraint is set by the for.each parameter. The 'lhs' terms are defined as a list of linear terms (summands). Each summand consists of a variable, a multiplier, and a set of sets for which the summand is defined.

See also

Other class constraint policy: class-constraint, class-costs, newCosts(), newSubsidy(), newTax(), subsidy-class, tax-class

Examples

# cap total exports in 2020 at 1 unit
cnst <- newConstraint(
  name = "EXP_CAP",
  desc = "Total exports in 2020",
  eq = "<=",
  defVal = 1,
  for.each = data.frame(year = 2020L),
  total = list(variable = "vExportTot", for.sum = list(region = "R1"))
)
cnst
#> Name: EXP_CAP, eq: <=, defVal: 1
#> desc:  Total exports in 2020 
#> 
#>  for.each 
#>   year
#> 1 2020
#> 
#> Term 1:
#> variable: vExportTot, defVal: 1
#> for.sum set:
#> $region
#> [1] "R1"
#> 
#> $comm
#> [1] NA
#> 
#> $timeslice
#> [1] NA
#>