Package 'FoRDM'

Title: Forest Many-Objective Robust Decision Making ('FoRDM')
Description: Forest Many-Objective Robust Decision Making ('FoRDM') is a R toolkit for supporting robust forest management under deep uncertainty. It provides a forestry-focused application of Many-Objective Robust Decision Making ('MORDM') to forest simulation outputs, enabling users to evaluate robustness using regret- and 'satisficing'-based measures. 'FoRDM' identifies robust solutions, generates Pareto fronts, and offers interactive 2D, 3D, and parallel-coordinate visualizations.
Authors: Marc Djahangard [aut, cre], Rasoul Yousefpour [aut]
Maintainer: Marc Djahangard <[email protected]>
License: MIT + file LICENSE
Version: 1.0.2
Built: 2026-05-14 09:19:46 UTC
Source: https://github.com/mdjahan/fordm

Help Index

Build FoRDM Table

Description

Transfers the provided data table into the format for FoRDM analysis. The columns that represent management, sow (state-of-the-world, scenarios), and time have to be defined. All other columns are treated as objectives.

Usage

build_fordm_table(data, management, sow, time)

Arguments

data

A data.frame containing the input data.
management

The name of the management column.
sow

The name of the state-of-the-worlds (SOW) column.
time

The name of the time column.

Value

A list with the processed data for further use in the FoRDM analysis, including the input data, mapping for identification of columns and objective columns.

Build Objectives Data Frame for Regret Analysis

Description

Specify for which objectives regret-based FoRDM analysis should be applied. For each objective, define its name, direction, weight, time aggregation method (mean, sum, min or max), and discount rate.

Usage

build_objectives_regret(
  names,
  direction = rep("maximize", length(names)),
  weights = rep(1/length(names), length(names)),
  time_aggregation = rep("mean", length(names)),
  discount_rate = rep(0, length(names))
)

Arguments

names

Names of objectives as the column names in the provided data.
direction

Direction of objective function: 'maximize' or 'minimize'.
weights

Relative weights (0-1) for each objective, must sum to 1.
time_aggregation

Time aggregation across objectives: 'mean', 'sum', 'min' or 'max'.
discount_rate

Annual discount rates for each objective (e.g., 0.02 means 2% per year), applied during time aggregation.

Value

A data frame specifying objectives, directions, weights, time aggregation methods, and discount rates for use in FoRDM analysis.

Build Objectives Data Frame for Satisficing Analysis

Description

Specify information for satisficing-based FoRDM analysis. For each objective, define its name, time aggregation method (mean, sum, min or max), discount rate, threshold and direction.

Usage

build_objectives_satisficing(
  names,
  time_aggregation = rep("mean", length(names)),
  discount_rate = rep(0, length(names)),
  threshold,
  direction = rep("above", length(names))
)

Arguments

names

Names of objectives as the column names in the provided data.
time_aggregation

Time aggregation across objectives: 'mean', 'sum', 'min' or 'max'.
discount_rate

Discount rates for each objective (e.g., 0.02 means 2% per time step), applied during time aggregation.
threshold

Numeric value(s) defining the satisficing level for each objective.
direction

'above' if values should meet or exceed the threshold, 'below' if they should be lower.

Value

A data frame specifying objectives name, time aggregation method, discount rate, threshold and direction for use in satisficing FoRDM analysis.

Forest Many-Objective Robust Decision Making (FoRDM)

Description

This package provides a toolkit for supporting robust forest management under deep uncertainty. It provides a forestry-focused application of Many-Objective Robust Decision Making (MORDM) to forest simulation outputs, enabling users to evaluate robustness using regret- and satisficing-based measures. FoRDM identifies robust solutions, generates Pareto fronts, and offers interactive 2D, 3D, and parallel-coordinate visualizations.

Details

Main features:

  • Create FoRDM input tables from forest simulation outputs.

  • Calculate robustness measures based on Regret and Satisficing approaches.

  • Visualize results using 2D, 3D, or parallel-coordinate plots.

  • Explore trade-offs between robustness and performance.

Author(s)

Marc Djahangard [email protected] (author, maintainer)

Rasoul Yousefpour (author)

See Also

browseVignettes(package = "FoRDM")

FoRDM Regret-Based many-objective Robust Decision-Making Analysis

Description

Performs a regret-based (Type II or cVaR) many-objective robustness analysis for the provided data and objectives, providing a optimal robust management and the Pareto front.

Usage

fordm_analysis_regret(
  fordm_table,
  objectives,
  robustness = 0.9,
  method = "regretII"
)

Arguments

fordm_table

Output from build_fordm_table().
objectives

Output from build_objectives_regret().
robustness

Numeric (0-1) specifying the quantile of regret used to define robustness, e.g., 0.9 evaluates management performance that is at least as good as in 90% of SOWs.
method

the method used to evaluate robustness

  • "regretII": Regret type II (regret to best performing alternative) approach using the robustness quantile of scenario regrets.

  • "CVaR": Conditional Value at Risk, using the mean of the worst (1 - robustness) fraction of weighted regrets for risk-aware selection.

Value

A list containing the results of the FoRDM analysis:

  • optimal: The management strategy identified as most robust given the regret metrics.

  • pareto_front: The Pareto front of robust management strategies.

FoRDM Satisficing-Based many-objective Robust Decision-Making Analysis

Description

Performs a satisficing-based many-objective robustness analysis for the provided data and objectives, providing a optimal robust management and the Pareto front.

Usage

fordm_analysis_satisficing(fordm_table, objectives, robustness = 0.9)

Arguments

fordm_table

Output from build_fordm_table().
objectives

Output from build_objectives_satisficing().
robustness

Numeric (0-1) specifying the robustness level across SOWs, e.g., 0.9 evaluates management performance that meets objectives in at least 90% of SOWs.

Value

A list containing the FoRDM analysis results:

  • optimal: The management strategy that balances all objectives (Euclidean distance) while meeting the robustness threshold.

  • pareto_front: The Pareto front of robust management strategies.

Robustness Trade-Off Analysis (Regret-based)

Description

Analyzes what happens when you sacrifice robustness for better performance. Shows marginal benefits and losses for each objective when switching between management strategies across different robustness levels.

Usage

robustness_tradeoff_analysis(
  fordm_table,
  objectives,
  robustness_min = 0,
  robustness_max = 1,
  robustness_step = 0.05
)

Arguments

fordm_table

Output from build_FoRDM_table().
objectives

Output from build_objectives_regret().
robustness_min

Minimum robustness level (default: 0.0 = 0%).
robustness_max

Maximum robustness level (default: 1.0 = 100%).
robustness_step

Step size for robustness levels (default: 0.05 = 5%-steps).

Value

List containing the list of optimal managements at certain robustness levels, and a plot

Visualize 2D Pareto Front

Description

Plots a 2D plot of the Pareto front of management alternatives from FoRDM_analysis output.

Usage

visualize_fordm_2d(analysis_output, x, y, fordm_method)

Arguments

analysis_output

Output list from FoRDM_analysis_regret() or FoRDM_analysis_satisficing().
x

Name of the objective for the x-axis (string).
y

Name of the objective for the y-axis (string).
fordm_method

Either "regret" or "satisficing".

Value

A ggplot2 object showing the 2D Pareto front for the selected objectives.

Visualize 3D Pareto Front

Description

Plots a 3D plot of the Pareto front of management alternatives from FoRDM_analysis output.

Usage

visualize_fordm_3d(analysis_output, x, y, z, fordm_method)

Arguments

analysis_output

Output from FoRDM_analysis_regret or FoRDM_analysis_satisficing.
x

Name of the objective for the x-axis (string).
y

Name of the objective for the y-axis (string).
z

Name of the objective for the z-axis (string).
fordm_method

Either "regret" or "satisficing".

Value

A plotly object showing the 3D Pareto front for the selected objectives.

Visualize a Parallel Coordinates plot of the Pareto Front for FoRDM Analysis Results

Description

Creates a parallel coordinates plot showing the Pareto front from FoRDM analysis.

Usage

visualize_fordm_parcoord(analysis_output, fordm_method)

Arguments

analysis_output

Output from fordm_analysis_regret() or fordm_analysis_satisficing().
fordm_method

Either "regret" or "satisficing".

Value

A parallel coordinates plot object.

Visualize Parallel Coordinates Plot for a single selected management across all SOWs

Description

Creates a parallel coordinates plot showing SOW performance across objectives for a selected management strategy.

Usage

visualize_fordm_parcoord_management(
  fordm_table,
  objectives,
  fordm_method,
  management
)

Arguments

fordm_table

Output from build_fordm_table().
objectives

Output from build_objectives_regret() or build_objectives_satisficing().
fordm_method

Either "regret" or "satisficing".
management

Character string specifying which management to visualize.

Value

A parallel coordinates plot object.