.. _auto_dataset_compute_pareto_frontiers:

*COMPUTE SORTED PARETO FRONT AND RETURN THE FRONTIER POINTS*
============================================================

Computes the Pareto frontier from a dataset by identifying non-dominated optimal records across one or more objective columns. Supports per-column min/max criteria, row-level constraints, and optional target values for parameter-identification use cases. Use this worker when you need to extract trade-off-optimal design points from a multi-objective dataset.

When to use
-----------

Classification: **process**.

Tagged: ``design_exploration``, ``frontier``, ``multi_objective``, ``non_dominated``, ``optimization``, ``pareto``, ``pareto_front``, ``trade_off``.

Inputs
------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20 20 20

   * - Label
     - ID
     - Type
     - Default
     - Required
     - Description
   * - Dataset
     - dataset1
     - dataset
     - —
     - 
     - Input tabular dataset containing the candidate records to evaluate; each row is a design point or result entry from which Pareto-optimal rows will be selected.
   * - Choose Columns
     - columns
     - scalar
     - —
     - 
     - Comma-separated list of column names (drawn from dataset1) that represent the objectives to optimize; at least one column must be specified.
   * - Objective
     - criteria
     - scalar
     - min
     - 
     - Comma-separated per-column optimization direction — 'min' to minimize or 'max' to maximize — matching the order of the chosen columns (e.g. 'min,max'); defaults to 'min' for all columns.
   * - Constraints
     - constraints
     - dataset
     - —
     - 
     - Optional dataset with columns [needle, condition, target] that define hard row-level filters applied before Pareto computation (e.g. needle=var1, condition=gt, target=2.0 keeps only rows where var1 > 2.0).
   * - Targets
     - targets
     - textarea
     - —
     - 
     - Optional colon-separated target values for each selected column used in parameter-identification mode, specified as 'col1:val1,col2:val2'; leave blank when pure Pareto optimality (no reference target) is desired.
   * - Group by
     - group
     - scalar
     - —
     - 
     - Group by this column. Each group will have a frontier

Outputs
-------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20

   * - Label
     - ID
     - Type
     - Description
   * - Frontiers
     - frontiers
     - dataset
     - Tabular dataset containing only the Pareto-optimal (non-dominated) rows from the input dataset, sorted along the frontier; preserves all original columns of dataset1.

Disciplines
-----------

- data.dataset.transform
- data.statistics
- design_exploration.optimization

Runnable example
----------------

A runnable example is registered for this worker. Open the example workflow on the d3VIEW canvas: `/api/workflow/example?id=dataset_compute_pareto_frontiers <https://www.d3view.com/api/workflow/example?id=dataset_compute_pareto_frontiers>`_

.. raw:: html

   <hr style="margin-top:2em">
   <p style="font-size:11px;color:#888">
   Auto-generated from <code>transformation</code> schema. Worker id: <code>dataset_compute_pareto_frontiers</code>. Schema hash: <code>a04dd2ca7cb0</code>. Hand-curated docs in <code>workerexamples/</code> override this page when present.
   </p>