Skip to contents

Write an equation as MathProg, GAMS or JuMP text

Source: R/mm_equation.R
mm_equation.Rd

Equations are the part of a model that is hardest to read as nested constructors: a cost balance takes several levels of ast_expression(). mm_equation() takes the statement in modeling-language notation and parses it into the same AST the constructors build.

Usage

mm_equation(model, text, desc = "", lang = c("gmpl", "gams", "jump"))

Arguments

model

A multimod model. Its declarations resolve the names used in text, so every set, parameter and variable the equation refers to must already be in it.

text

One equation statement. For lang = "gmpl" this is MathProg, name{indexing}: lhs <op> rhs with an optional s.t. prefix and trailing ;, where <op> is <=, >= or =. For lang = "gams" it is name(dims)$domain(dims) .. lhs <op> rhs with =l=, =g= or =e=. For lang = "jump" it is one @constraint(model, name[i in S], lhs <op> rhs) statement, with == for equality. Newlines are allowed in all three.

desc

Description to attach to the equation.

lang

The notation text is written in. MathProg is the default: it is an open language and uses the ordinary relational operators. JuMP and GAMS are read as well; all three produce the same AST.

Value

An equation node, ready to assign into model$equations.

Details

Declarations stay in R. new_set(), new_parameter() and new_variable() are flat and read well; only the expression needs the notation.

See also

new_equation() builds the same object from constructors, which is the route for an equation that is generated rather than written.

Examples

m <- example_models$toy

# MathProg (default)
eq <- mm_equation(
  m,
  "eqCapLim{(tech,region) in mTR}:
     vGen[tech,region] <= pAvail[tech,region] * vCap[tech,region]",
  desc = "generation is limited by available capacity"
)
cat(as_latex(eq))
#> \begin{flushleft}
#> \subsection*{\textbf{eqCapLim} -- \textit{generation is limited by available capacity}} 
#> \quad$\textbf{eqCapLim} 
#> _{\texttt{tech},\texttt{region}}$ 
#> $\mid\left\{\texttt{tech},\texttt{region}\right\} \in \mathit{mTR}_{\texttt{tech},\texttt{region}}$ \\
#> \begin{equation*}
#> \begin{adjustbox}{max width=\textwidth}
#> $\begin{aligned}
#> &\bm{\mathit{vGen}}_{\text{tech},\text{region}} \le \mathsf{pAvail}_{\text{tech},\text{region}}  \cdot  \bm{\mathit{vCap}}_{\text{tech},\text{region}} \\
#> \end{aligned}$
#> \end{adjustbox}
#> \end{equation*}
#> \vspace{1em}
#> \end{flushleft}

# the same equation in GAMS
eq2 <- mm_equation(
  m,
  "eqCapLim(tech,region)$mTR(tech,region) ..
     vGen(tech,region) =l= pAvail(tech,region) * vCap(tech,region)",
  lang = "gams"
)

# and in JuMP
eq3 <- mm_equation(
  m,
  "@constraint(model, eqCapLim[(h, r) in mTR],
     vGen[(h, r)] <= get(pAvail, (h, r), pAvailDef) * vCap[(h, r)])",
  lang = "jump"
)