Skip to contents

Changelog

Source: NEWS.md

multiscape 1.1.0

Solution architecture

  • Simplified the public result architecture so that solve() consistently returns a SolutionSet, including for single-objective problems.
  • Removed the internal Solution class from the public API and documentation. Individual run-level solutions remain available only as internal components of a SolutionSet.
  • Added stable solution_id identifiers to distinguish stored solutions from attempted runs identified by run_id.
  • Updated run, solution, and summary tables to preserve run_id and solution_id consistently across extraction and analysis functions.
  • Improved internal finalization of solution metadata and identifiers after solving.
  • Updated SolutionSet printing and documentation to reflect the distinction between run attempts and stored solutions.

Base conservation-planning workflow

  • Added automatic support for base conservation-planning problems when no explicit actions or effects are supplied.
  • Problems without explicit actions are now interpreted as binary conservation decisions, where each planning unit can be conserved or not conserved.
  • Added an implicit conservation action and internally generated feature contributions based on the feature amounts stored in dist_features.
  • The implicit conservation model uses amount_after to represent the feature amount obtained when a planning unit is conserved.

Multi-objective run design and controls

  • Revised the multi-objective run-design resolver used by weighted-sum, epsilon-constraint, and AUGMECON methods.
  • Improved support for automatic and manually specified run designs through run_grid() and run_manual().
  • Added and documented common multi-objective execution controls through mo_control().
  • Improved handling of infeasible runs, missing solutions, solver errors, and slack upper bounds.
  • Standardized the storage of run-design parameters and objective values in the SolutionSet run table.
  • Preserved infeasible runs in the run history while assigning solution_id only to runs that produced a stored solution.
  • Improved objective evaluation and model preparation across multi-objective methods.
  • Fixed weighted-sum objective preparation for implicit conservation benefit objectives.
  • Improved validation of objective coefficient vectors and objective-specific error messages.

Result extraction

  • Added get_runs() to extract run-level status, runtime, gap, design parameters, solution identifiers, and objective values.
  • Added get_objectives() to extract objective values in long or wide format.
  • Added get_objective_specs() to extract registered objective aliases, objective types, model types, optimization senses, and creation metadata.
  • Updated objective extraction to include both run_id and solution_id.
  • Updated existing extraction functions to work consistently with the unified SolutionSet architecture.
  • Improved handling of infeasible runs and missing objective values during extraction.

Solution-set management

  • Added solution_filter() to return a coherently filtered SolutionSet.
  • solution_filter() can filter by run_id, solution_id, solver status, or feasibility.
  • Added optional filtering of non-dominated solutions using moocore.
  • Added support for selecting the objectives used to evaluate dominance.
  • Ensured that filtering updates run tables, design tables, stored solutions, and summary tables consistently.
  • Added cloning of SolutionSet objects before modification to prevent reference-based mutation of the original object.
  • Added solution_append() to combine compatible SolutionSet objects generated from the same planning problem.
  • solution_append() verifies compatibility of planning units, features, actions, effects, targets, constraints, locks, spatial relations, and objective definitions.
  • Appended runs and solutions are automatically assigned unique run_id and solution_id values.
  • Added support for combining solution sets obtained from different multi-objective methods or run designs applied to the same planning problem.
  • Added solution_unique() to retain one representative from groups of equivalent solutions.
  • solution_unique() can identify repeated solutions using either complete decision vectors or objective values.
  • Added numerical tolerance controls for identifying equivalent points in objective space.
  • Preserved runs without stored solutions when removing duplicated solutions.

Frontier analysis

  • Added frontier_extremes() to identify the observed minimum and maximum values of each objective.
  • Added classification of observed bounds as best or worst according to each objective’s optimization sense.
  • Added support for returning all tied extreme solutions or only the first representative.
  • Added frontier_distances() to calculate normalized distances to observed ideal and nadir points.
  • Added automatic transformation of maximization objectives into a common minimization space for frontier calculations.
  • Added range normalization so that objectives measured in different units contribute comparably to distance calculations.
  • Added Euclidean, Manhattan, and Chebyshev distance metrics.
  • Added rankings based on proximity to the observed ideal point and distance from the observed nadir point.
  • Added original-scale ideal, nadir, and objective-range metadata to the returned distance tables.
  • Clarified that frontier reference points are calculated from the solutions contained in the supplied SolutionSet.

Selection analysis

  • Added selection_frequency() to calculate how frequently each planning-unit/action assignment is selected across stored solutions.
  • Standardized selection analysis around a canonical planning-unit/action representation.
  • Simple conservation-planning problems are represented using the implicit conservation action.
  • Added selection_similarity() to quantify structural similarity among solutions.
  • Added Jaccard similarity for comparing selected planning-unit/action assignments.
  • Added Hamming similarity for comparing complete binary assignment vectors, including shared non-selections.
  • Added long-format and matrix-format similarity outputs.
  • Added internal helpers to construct consistent long-format and matrix-format selection representations.

Documentation and website

  • Reorganized the pkgdown reference index into dedicated sections for:
    • result extraction;
    • solution-set management;
    • frontier analysis;
    • selection analysis;
    • multi-objective workflow configuration.
  • Removed public documentation references to the internal Solution class.
  • Updated function documentation to use the unified SolutionSet terminology.
  • Updated examples and cross-references for run_id, solution_id, objective extraction, filtering, frontier analysis, and selection analysis.
  • Added moocore as an optional dependency for non-dominance filtering.
  • Updated GitHub Actions configurations for current Codecov and Node.js runner requirements.
  • Updated Codecov uploads to use the PyPI CLI and avoid binary signature-verification failures.
  • Updated GitHub Actions versions for Node.js 24 compatibility.

multiscape 1.0.7

CRAN release: 2026-04-30

  • Updated native routine registration to resolve additional CRAN LTO/gcc-ASAN checks.
  • Revised examples and package metadata for CRAN compliance.

multiscape 1.0.6

CRAN release: 2026-04-28

  • Fix CRAN submission issues
  • Revise examples and DESCRIPTION for CRAN resubmission

multiscape 1.0.5

  • Release candidate for CRAN.

multiscape 1.0.4

  • First CRAN release of multiscape.
  • Provides a modular workflow for exact multi-objective spatial planning based on mixed-integer programming (MIP).
  • Introduces the core Problem, Solution, and SolutionSet classes. The public result architecture was later unified around SolutionSet in version 1.1.0.
  • Adds support for modular problem construction through create_problem(), add_*(), set_*(), and solve().
  • Supports atomic objective registration and multi-objective solution methods, including weighted-sum, epsilon-constraint, and AUGMECON.
  • Includes support for spatial relations such as boundary, rook, queen, k-nearest neighbours, and distance-based relations.
  • Supports commercial and open-source solvers, including Gurobi, CPLEX, CBC, and SYMPHONY.
  • Adds user-facing extraction and visualization tools for planning units, actions, features, targets, spatial outputs, and trade-offs.
  • Includes substantial updates to documentation, package structure, and contribution guidelines.