mooplot.plot_eaf

mooplot.plot_eaf(dataset, type='fill', percentiles=[], colorway=[], fill_border_colours=[], trace_names=[], line_dashes='solid', line_width=[], legend_preset='centre_top_right', template='simple_white', **layout_kwargs)

Plot attainment surfaces in 2D.

Parameters:

• dataset (ArrayLike) – The dataset argument must be Numpy array of EAF values (2 objectives and percentile marker), or it can be a dictionary of such values. The dictionary must have this format: {‘alg_name_1’ : dataset1, ‘alg_name_2’ : dataset2}.

• percentiles (list, default: []) – A list of percentiles to plot. These must exist in the dataset argument. If multiple datasets are provided, this can also be a list of lists - selecting percentile groups for each algorithm (dictionary interface)

• type (str, default: 'fill') – The type argument can be “fill”, “points”, “lines” to define the plot type (See plot_pf() for more details). If multiple datasets are used (dictionary interface) then this can be a list of types (equal to the number of different datasets provided).

• colorway (list, default: []) – Colorway is a single colour, or list of colours, for the percentile groups. The colours can be CSS colours such as ‘black’, 8-digit hexadecimal RGBA integers or strings of RGBA values such as rgba(1,1,0,0.5). Default is “black”. You can use functions colour.discrete_colour_gradient to create a gradient between two colours In case of multiple datasets (“dictionary interface”), you can use a single list to set the value for each set of lines, or a 2d list to set a value for each line within a surface

• fill_border_colours (list, default: []) – The same as colorway but defining the boundaries between percentile groups. The default value is to follow colorway. You can set it to rgb(0,0,0,0) to make the boundaries invisible

• trace_names (list, default: []) – Overide the default trace names by providing a list of strings

• line_dashes (str, default: 'solid') – Select whether lines are dashed. Choices are: ‘solid’, ‘dot’, ‘dash’, ‘longdash’, ‘dashdot’, ‘longdashdot’. A single string sets the type for all lines. A list sets them individually. In case of multiple datasets (“dictionary interface”), you can use a single list to set the value for each set of lines, or a 2d list to set a value for each line within a surface

• line_width (list, default: []) – Select the line width (default = 2) of the traces. Similar interface to line_dashes, colorway etc -> Enter a single value to set all traces. For single datset, a list sets size for all sets For a dictionary of datsets: a list sets the same value for all traces assosciated with that dataset. A list of list individually sets width for every trace in every dataset

• legend_preset (str, default: 'centre_top_right') – See “preset” argument for function apply_legend_preset()

• template (str, default: 'simple_white') – Choose layout template for the plot - see Plotly template tutorial . Default is “simple_white”

• layout_kwargs – Update features of the graph such as title axis titles, colours etc. These additional parameters are passed to plotly update_layout, See here for all the layout features that can be accessed: Layout Plotly reference

Returns:

Figure – The function returns a Plotly GO figure object Figure Plotly reference This means that the user can customise any part of the graph after it is created

Notes

For more background, see López-Ibáñez et al.[1].

References

[1]

Manuel López-Ibáñez, Luís Paquete, and Thomas Stützle. Exploratory analysis of stochastic local search algorithms in biobjective optimization. In Thomas Bartz-Beielstein, Marco Chiarandini, Luís Paquete, and Mike Preuss, editors, Experimental Methods for the Analysis of Optimization Algorithms, pages 209–222. Springer, Berlin / Heidelberg, 2010. doi:10.1007/978-3-642-02538-9_9.