MizerScan: S3 class for the result of a parameter scan
In mizer: Dynamic Multi-Species Size Spectrum Modelling

MizerScan

R Documentation

S3 class for the result of a parameter scan

Description

scanModel() varies one aspect of a model over a range of values and measures a quantity on the attractor the model settles on at each of them. It returns its result as a MizerScan object, which is a data frame carrying, in addition, everything that plot() needs to draw it.

Usage

MizerScan(
  x,
  scan_name = NULL,
  scan_units = NULL,
  value_name = NULL,
  value_units = NULL,
  type = NULL,
  params = NULL,
  reference_lines = NULL,
  settings = NULL
)

is.MizerScan(x)

Arguments

`x`	A data frame with at least three columns, laid out as described above. For `is.MizerScan()`, any object to test.
`scan_name`	A string naming the quantity that was varied.
`scan_units`	A string giving its units, for example `"1/year"`.
`value_name`	A string naming the quantity that was measured.
`value_units`	A string giving its units, for example `"g/year"`.
`type`	The kind of quantity the measured values are, see array_types.
`params`	The `MizerParams` object the scan started from.
`reference_lines`	An optional named numeric vector of x positions to mark with vertical lines.
`settings`	An optional list recording the settings used.

Details

A MizerScan object behaves like an ordinary data frame with one row for each combination of scanned value and series. Its columns are, in order:

1: The scanned value. The column is named after the scan_name, so for example a scan over fishing effort has a column called "Fishing effort". Use names(scan)[[1]] rather than hard-coding it.
2: The measured quantity, averaged over the attractor. Named after the value_name.
Species: The series the row refers to. Named Species whatever the series are, because that is the column that mizer's colour and line-type machinery reads.
ymin, ymax: The smallest and largest value over the sampling window. On a fixed point these both equal the value; on a limit cycle they give the range of the oscillation.
attractor: What the state reached at this scan value is: "fixed_point", "limit_cycle" or NA for neither. This is the column that says whether the value in this row can be read as an equilibrium.
termination, converged: Why the run at this scan value stopped, and whether the solver met its own criterion. Both come from the "convergence" attribute that projectUntilSettled() attaches to its result, and neither is a claim about the state — see attractor for that.
period: The period of the limit cycle in years, or NA.
residual: How far the state still is from a fixed point, as a per-capita rate in 1/year, see getSteadyResidual().

The first three columns are the x, y and grouping variable in that order, which is the layout plotDataFrame() expects.

It also carries these attributes:

scan_name, scan_units – name and units of the quantity that was varied, used for the x-axis label.
value_name, value_units – name and units of the quantity that was measured, used for the y-axis label.
type – the kind of quantity the values are, see array_types.
params – the MizerParams object the scan started from, used for species colours and line types.
reference_lines – an optional named numeric vector of positions on the x axis to mark with vertical lines, for example c(F_MSY = 0.32).
at_max, max_value – for each series, the scanned value at which the measured quantity is largest, and the value it takes there. See the section below.
settings – a list recording the settings the scan was run with.

Value

A MizerScan object, which inherits from data.frame.

is.MizerScan() returns TRUE if x is a MizerScan object, FALSE otherwise.

Where the maximum is

The at_max attribute holds, for each series, the scanned value at which that series' measured quantity is largest. On a yield-versus-fishing-mortality scan that is F_{MSY}; on a scan over fishing effort it is the effort that maximises the quantity being plotted. max_value holds the value attained there.

This is the largest value among those that were scanned, not the maximum of the underlying curve. It is therefore only as good as the grid you gave in scan_values, and the way to sharpen it is to scan a finer grid near the maximum, not to interpolate a coarse one. Subsetting a MizerScan with [ recomputes both attributes from the rows that remain, so they never go stale.

Limitations

Because the object is a data frame subclass, its attributes survive base R subsetting with [ but are dropped by functions that rebuild the data frame, including dplyr::filter(), dplyr::mutate(), subset() and transform(). A scan that has lost its attributes can no longer be plotted. Subset with [, or rebuild the object with MizerScan().

Examples


scan <- scanModel(NS_params, scan_values = c(0, 0.5, 1),
                  set_func = scanEffort(), species = "Cod")
scan
summary(scan)
attr(scan, "at_max")

mizer documentation built on Aug. 24, 2026, 9:08 a.m.