Dimensions and automatic names

The dimension of the state, the control and the variable is given by the domain of the declaration: R for a scalar, R^n for a vector of dimension n.

@def begin
    t ∈ [0, 1], time
    x ∈ R^3, state       # dimension 3
    u ∈ R², control      # dimension 2, same as R^2
    v ∈ R, variable      # dimension 1: a scalar
    ...
end

Digit exponents

R², R³, … are other spellings of R^2, R^3, …. They exist up to R²⁰; for a larger dimension write R^21.

Symbolic dimension: R^n

The exponent can be any expression whose value is known when the block is evaluated: a global, an argument of the enclosing function, or an expression such as n + 1.

function problem(n)
    @def begin
        t ∈ [0, 1], time
        x ∈ R^n, state
        u ∈ R^(n + 1), control
        ...
    end
end

If n is not defined, Julia reports UndefVarError: n not defined.

Rⁿ is not R^n

Rⁿ is a single identifier whose last letter is a superscript n. Only the digit exponents are recognised, so a symbolic dimension cannot be read from it, and

x ∈ Rⁿ, state          # ParsingError: unknown syntax

is rejected on purpose. Write x ∈ R^n, state.

Names of the components

By default the components of x are named x₁, x₂, …, and can be used in the block under these names or as x1, x2, …; they are also accessible as x[1], x[2], …. A name given by the user, x = (q, v) ∈ R², state, adds q and v for the same components (see Declarations).

Limit of the automatic names

When the dimension is a literal number (R^12, R¹²), every component has its automatic names. When it is symbolic (R^n), the dimension is not known when the block is parsed, and only the components x₁ to x₉ (and x1 to x9) are named. A component beyond the ninth is reached by its index, x[10], or by a name given in the declaration. A name such as x₁₀ is not an error at parse time: it is an undefined variable, and Julia reports it when the model is evaluated.