refactor(api)!: standardize dimensional keyword arguments

Use nvar and nobj across benchmark constructors, dimension metadata and catalog filters. Preserve DTLZ k and coupled dimensions, and expose SLCDT1 lambda with an ASCII keyword. Update tests, examples and concise scientific docstrings.

BREAKING CHANGE: positional dimension arguments and legacy n, m and λ keywords are no longer supported. Rename min_vars/max_vars and min_objs/max_objs filters to min_nvar/max_nvar and min_nobj/max_nobj; metadata fields default_n/default_m become default_nvar/default_nobj.

Author: Souza-DR

docs/src/guides/catalog.md

@@ -61,7 +61,7 @@ julia> small_defaults = filter_problems(
            name_pattern = r"^ZDT",
            has_bounds = true,
            has_jacobian = true,
-           max_vars = 10,
+           max_nvar = 10,
        )
 2-element Vector{String}:
  "ZDT4"
@@ -78,7 +78,7 @@ The [`ZDT1`](@ref) constructor accepts `n >= 2`, so it can still be used in the
 ten-variable experiment:
 
 ```jldoctest catalog_workflow
-julia> prob = ZDT1(10);
+julia> prob = ZDT1(nvar = 10);
 
 julia> prob.nvar
 10

docs/src/guides/evaluation.md

@@ -11,7 +11,7 @@ Construct the benchmark and evaluate all of its objectives with `eval_f`:
 ```jldoctest evaluation_workflow
 julia> using MOProblems
 
-julia> prob = ZDT1(10);
+julia> prob = ZDT1(nvar = 10);
 
 julia> x = fill(0.5, prob.nvar);
 

docs/src/problems/dtlz.md

@@ -7,21 +7,23 @@ Theoretical Advances and Applications* [DTLZ2005](@cite).
 
 ## Overview
 
-For all five constructors, `k >= 1`, `m >= 2`, `nvar = k + m - 1`, and
-`nobj = m`. Here, `k` is the number of trailing variables `x_m, ..., x_n` used
-by `g(x)`; equivalently, `k = nvar - m + 1`. The first `m - 1` variables
+For all five constructors, `k >= 1`, `nobj >= 2`, and `nvar = k + nobj - 1`.
+The public keyword is `nobj` by package convention; the mathematical
+formulation below retains the usual `m`, with `m = nobj`. Here, `k` is the
+number of trailing variables `x_m, ..., x_n` used by `g(x)`; equivalently,
+`k = nvar - m + 1`. The first `m - 1` variables
 determine the objective trade-off, whereas these `k` variables determine
 `g(x)`. Thus, `k` changes the number of decision variables without changing
 the number of objectives. DTLZ4 additionally requires `alpha > 0`. The default
 dimensions and componentwise bounds are shown below.
 
-| Problem | `k` | `m` | `alpha` | `nvar` | `nobj` | Lower bound | Upper bound |
-|:---|---:|---:|---:|---:|---:|---:|---:|
-| `DTLZ1` | 5 | 3 | — | 7 | 3 | 0.0 | 1.0 |
-| `DTLZ2` | 10 | 3 | — | 12 | 3 | 0.0 | 1.0 |
-| `DTLZ3` | 10 | 3 | — | 12 | 3 | 0.0 | 1.0 |
-| `DTLZ4` | 10 | 3 | 100.0 | 12 | 3 | 0.0 | 1.0 |
-| `DTLZ5` | 10 | 5 | — | 14 | 5 | 0.0 | 1.0 |
+| Problem | `k` | `nobj` | `alpha` | `nvar` | Lower bound | Upper bound |
+|:---|---:|---:|---:|---:|---:|---:|
+| `DTLZ1` | 5 | 3 | — | 7 | 0.0 | 1.0 |
+| `DTLZ2` | 10 | 3 | — | 12 | 0.0 | 1.0 |
+| `DTLZ3` | 10 | 3 | — | 12 | 0.0 | 1.0 |
+| `DTLZ4` | 10 | 3 | 100.0 | 12 | 0.0 | 1.0 |
+| `DTLZ5` | 10 | 5 | — | 14 | 0.0 | 1.0 |
 
 Analytical Jacobians are registered for all five constructors. **Hessians are
 not registered**. The catalog metadata classifies every objective in `DTLZ1` through `DTLZ5`
@@ -175,7 +177,7 @@ g(x) = \sum_{r=m}^{n}(x_r-0.5)^2.
 ```jldoctest dtlz_usage
 julia> using MOProblems
 
-julia> prob = DTLZ2(k = 10, m = 4);
+julia> prob = DTLZ2(k = 10, nobj = 4);
 
 julia> x = fill(0.5, prob.nvar);
 

docs/src/problems/fds.md

@@ -6,7 +6,7 @@ was introduced in "Newton's Method for Multiobjective Optimization"
 
 ## Overview
 
-`FDS(n)` requires `n >= 1` and has `nvar = n` and `nobj = 3`; the default is
+`FDS(; nvar)` requires `nvar >= 1` and has `nobj = 3`; the default is
 `n = 5`. Its componentwise bounds are shown below.
 
 | Problem | `n` | `nvar` | `nobj` | Lower bound | Upper bound |

docs/src/problems/jos.md

@@ -20,8 +20,8 @@ How?" by Jin, Olhofer, and Sendhoff [JOS2001](@cite).
 
 ## Overview
 
-Both constructors have `n` variables and two objectives. `JOS1` requires
-`n >= 1`, whereas `JOS4` requires `n >= 2`. The default dimension of 50 matches
+Both constructors have two objectives. `JOS1` requires
+`nvar >= 1`, whereas `JOS4` requires `nvar >= 2`. The default dimension of 50 matches
 the dimension used for the `JOS1a` and `JOS4a` experiment instances reported
 by Fliege, Drummond, and Svaiter [FDS2009](@cite). Both constructors retain the
 componentwise domain ``[0,1]^n`` specified for the source test functions
@@ -96,7 +96,7 @@ g(x)=1+\frac{9}{n-1}\sum_{i=2}^{n}x_i.
 ```jldoctest jos_usage
 julia> using MOProblems
 
-julia> prob = JOS4(50);
+julia> prob = JOS4(nvar = 50);
 
 julia> x = fill(0.5, prob.nvar);
 

docs/src/problems/mgh.md

@@ -22,32 +22,33 @@ are fixed, coupled, or independently configurable.
 
 `MGH9` has fixed dimensions: the package exposes no dimension parameter and
 constructs a problem with `nvar = 3` and `nobj = 15`. These dimensions agree
-with `n = 3` and `m = 15` in both [MGH1981](@cite) and [Mita2019](@cite).
+with both [MGH1981](@cite) and [Mita2019](@cite).
 
-`MGH16` fixes `nvar = 4` and accepts `m >= 4`, with `nobj = m`; its default is
-`m = 5`. The original problem fixes `n = 4` and allows `m >= n`
-[MGH1981](@cite), while [Mita2019](@cite) uses `n = 4` and `m = 5`.
+`MGH16` fixes `nvar = 4` and accepts `nobj >= 4`; its default is
+`nobj = 5`. The original problem fixes four variables and allows at least four
+objectives [MGH1981](@cite), while [Mita2019](@cite) uses the same package
+defaults.
 
-`MGH26` accepts `n >= 1` and couples the dimensions as
-`nvar = nobj = n`; its default is `n = 4`. The original problem allows `n` to
-vary and defines `m = n` [MGH1981](@cite), while [Mita2019](@cite) uses
-`n = m = 4`.
+`MGH26` accepts `nvar >= 1` and couples the dimensions as
+`nobj = nvar`; its default is `nvar = 4`. The original problem
+[MGH1981](@cite) permits a
+variable number of variables and couples the number of objectives to it, while [Mita2019](@cite) uses `nvar = nobj = 4`.
 
-`MGH33` accepts `n >= 2` and `m >= 2` independently, with `nvar = n` and
-`nobj = m`; its defaults are `n = m = 10`. The original problem allows `n` to
-vary but requires `m >= n` [MGH1981](@cite), while [Mita2019](@cite) uses
-`n = 10` and `m = 4`. Thus, the package supports the dimensions independently
-and includes `(n, m) = (10, 4)` as a nondefault configuration.
+`MGH33` accepts `nvar >= 2` and `nobj >= 2` independently; its defaults are
+`nvar = nobj = 10`. The original problem [MGH1981](@cite) permits a variable number of
+variables but requires at least as many objectives, while
+[Mita2019](@cite) uses `nvar = 10` and `nobj = 4`. Thus, the package supports
+the dimensions independently.
 
 The following table summarizes the dimension behavior implemented by the
 constructors.
 
 | Problem | Dimension behavior | Configurable parameters | Default `nvar` | Default `nobj` |
 |:---|:---|:---|---:|---:|
 | `MGH9`  | Fixed | — | 3 | 15 |
-| `MGH16` | Fixed `nvar`, variable `nobj` | `m >= 4` | 4 | 5 |
-| `MGH26` | Coupled | `n >= 1`, with `nvar = nobj = n` | 4 | 4 |
-| `MGH33` | Independent | `n >= 2`, `m >= 2` | 10 | 10 |
+| `MGH16` | Fixed `nvar`, variable `nobj` | `nobj >= 4` | 4 | 5 |
+| `MGH26` | Coupled | `nvar >= 1`, with `nobj = nvar` | 4 | 4 |
+| `MGH33` | Independent | `nvar >= 2`, `nobj >= 2` | 10 | 10 |
 
 Analytical Jacobians are registered for all four constructors. Hessians are
 not registered. The catalog metadata classifies every objective in each
@@ -125,7 +126,7 @@ The following example uses the linear rank-1 dimensions reported in
 ```jldoctest mgh_usage
 julia> using MOProblems
 
-julia> prob = MGH33(n = 10, m = 4);
+julia> prob = MGH33(nvar = 10, nobj = 4);
 
 julia> x = zeros(prob.nvar);
 

docs/src/problems/qv.md

@@ -8,7 +8,7 @@ parameter domain under the `QV1` name [Huband2006](@cite).
 
 ## Overview
 
-`QV1(n)` requires `n >= 1` and has `nvar = n` and `nobj = 2`. The default
+`QV1(; nvar)` requires `nvar >= 1` and has `nobj = 2`. The default
 `n = 16` is the dimension used by Quagliarella and Vicini. The constructor
 retains the componentwise domain specified in both sources.
 

docs/src/problems/slcdt.md

@@ -6,8 +6,7 @@ Schütze, Laumanns, Coello Coello, Dellnitz, and Talbi [SLCDT2008](@cite).
 ## Overview
 
 Both constructors have fixed dimensions and registered componentwise variable
-bounds. `SLCDT1` takes a perturbation coefficient `λ`, which selects the
-bounds; `SLCDT2` takes no parameters.
+bounds. `SLCDT1` takes perturbation coefficient `λ`, which selects the bounds; `SLCDT2` takes no parameters.
 
 | Problem | `λ` | `nvar` | `nobj` | Lower bound | Upper bound |
 |:---|:---|---:|---:|---:|---:|
@@ -75,7 +74,7 @@ julia> values1 = eval_f(prob1, x1);
 
 julia> J1 = eval_jacobian(prob1, x1);
 
-julia> prob1b = SLCDT1(λ = 0.0);
+julia> prob1b = SLCDT1(lambda = 0.0);
 
 julia> prob2 = SLCDT2();
 

docs/src/problems/toi.md

@@ -27,17 +27,17 @@ numbering.
 and therefore two objectives, matching the instance of [Mita2019](@cite).
 
 `Toi8`, `Toi9`, and `Toi10` are stated for a variable dimension in
-[Toi1983](@cite), and the constructors accept `n >= 2` and couple the
-dimensions accordingly: `nvar = nobj = n` for `Toi8` and `Toi9`, and
-`nvar = n`, `nobj = n - 1` for `Toi10`. Their defaults reproduce the instances
-of [Mita2019](@cite), namely `n = 3` for `Toi8` and `n = 4` for the other two.
-
-| Problem | Dimension behavior | Configurable parameters | Default `nvar` | Default `nobj` | Bounds |
-|:---|:---|:---|---:|---:|:---|
-| `Toi4` | Fixed | — | 4 | 2 | ``[-2,5]^4`` |
-| `Toi8` | Coupled | `n >= 2`, with `nvar = nobj = n` | 3 | 3 | ``[-1,1]^n`` |
-| `Toi9` | Coupled | `n >= 2`, with `nvar = nobj = n` | 4 | 4 | ``[-1,1]^n`` |
-| `Toi10` | Coupled | `n >= 2`, with `nvar = n` and `nobj = n - 1` | 4 | 3 | ``[-2,2]^n`` |
+[Toi1983](@cite), and the constructors accept `nvar >= 2` and couple the
+dimensions accordingly: `nobj = nvar` for `Toi8` and `Toi9`, and
+`nobj = nvar - 1` for `Toi10`. Their defaults reproduce the instances
+of [Mita2019](@cite), namely `nvar = 3` for `Toi8` and `nvar = 4` for the other two.
+
+| Problem | Dimension behavior | Configurable parameters | Default `nvar` | Default `nobj` | Lower bound | Upper bound |
+|:---|:---|:---|---:|---:|---:|---:|
+| `Toi4` | Fixed | — | 4 | 2 | -2 | 5 |
+| `Toi8` | Coupled | `nvar >= 2`, with `nobj = nvar` | 3 | 3 | -1 | 1 |
+| `Toi9` | Coupled | `nvar >= 2`, with `nobj = nvar` | 4 | 4 | -1 | 1 |
+| `Toi10` | Coupled | `nvar >= 2`, with `nobj = nvar - 1` | 4 | 3 | -2 | 2 |
 
 Analytical Jacobians are registered for all four constructors. Hessians are not
 registered. The catalog metadata classifies every objective in each default
@@ -99,7 +99,7 @@ f_i(x)=100\left(x_{i+1}-x_i^2\right)^2+\left(x_{i+1}-1\right)^2,
 ```jldoctest toi_usage
 julia> using MOProblems
 
-julia> prob = Toi9(n = 4);
+julia> prob = Toi9(nvar = 4);
 
 julia> x = fill(0.5, prob.nvar);
 

docs/src/problems/zdt.md

@@ -16,8 +16,8 @@ the package provides no constructor for it.
 
 ## Overview
 
-Every constructor takes the number of variables `n` as its only parameter and
-requires `n >= 2`, giving `nvar = n` and `nobj = 2`. The default dimensions and
+Every constructor takes the number of variables `nvar` as its only parameter and requires
+`nvar >= 2`, giving `nobj = 2`. The default dimensions and
 componentwise bounds are shown below.
 
 | Problem | Source function | Default `nvar` | `nobj` | Lower bounds | Upper bounds |
@@ -175,7 +175,7 @@ g(x) = 1 + 9\left(\frac{1}{n-1}\sum_{i=2}^{n}x_i\right)^{0.25}.
 ```jldoctest zdt_usage
 julia> using MOProblems
 
-julia> prob = ZDT1(30);
+julia> prob = ZDT1(nvar = 30);
 
 julia> x = fill(0.5, prob.nvar);