洪水感受性低減のための自然基盤ソリューションの公平性を考慮した多目的最適化：再現コードと処理済みデータ
Reproducibility Code and Processed Data for "Equity-Aware Multi-Objective Optimization of Nature-Based Solutions for Flood Susceptibility Reduction" (原題)
Chibuike Chiedozie Ibebuchi
🤖 gxceed AI 要約
日本語
本パッケージは、洪水感受性低減のための自然基盤ソリューション（植樹・湿地回復）の配分を公平性を考慮して最適化する研究の再現用コードと処理済みデータを提供する。メリーランド州の4,079の国勢調査ブロックグループを対象に、混合整数最適化と強化学習（PPO）などのベンチマークを比較し、5つの公平性重みで集計リスク削減と脆弱性加重便益のトレードオフを評価する。
English
This package provides code and processed data for reproducing a study on equity-aware multi-objective optimization of nature-based solutions (tree planting, wetland restoration) for flood susceptibility reduction. It includes 4,079 Maryland census block groups, comparing exact mixed-integer optimization with benchmarks like PPO, and evaluating trade-offs across five equity weights.
Unofficial AI-generated summary based on the public title and abstract. Not an official translation.
📝 gxceed 編集解説 — Why this matters
日本のGX文脈において
日本では気候変動適応策としてグリーンインフラやNBSの導入が進むが、公平性を明示的に考慮した配分最適化の事例は少なく、自治体の適応計画や国土強靱化に示唆を与える。再現可能なコードとデータは、日本の研究者や実務者が同様の分析を実施する際の参考になる。
In the global GX context
Globally, this work contributes to the growing literature on nature-based solutions for climate adaptation, particularly by integrating equity considerations into spatial optimization. The reproducible code and dataset enable comparative studies and methodological advancements in other regions, aligning with international efforts on climate-resilient development.
👥 読者別の含意
🔬研究者:Provides a reproducible framework and dataset for equity-aware optimization of NBS, useful for extending to other regions or incorporating additional objectives.
🏢実務担当者:Offers a decision-support tool for prioritizing NBS investments that balance flood risk reduction and social equity, applicable to urban planning and climate adaptation.
🏛政策担当者:Demonstrates how to integrate equity metrics into infrastructure planning, informing policies that address both climate resilience and environmental justice.
📄 Abstract（原文）
This reproducibility package accompanies the study “Equity-Aware Multi-Objective Optimization of Nature-Based Solutions for Flood Susceptibility Reduction.” The archive contains the complete analysis code and the processed Census Block Group (CBG)-level dataset used to reproduce the optimization, benchmark, controlled-comparison, sensitivity, and diagnostic analyses reported in the study. The analysis evaluates the allocation of tree-planting and wetland-restoration projects under common suitability, intervention-capacity, and budget constraints. The revised framework explicitly represents the trade-off between aggregate flood-susceptibility reduction and vulnerability-weighted mitigation benefit across five equity weights. Exact mixed-integer optimization is used as the reference solution, while greedy, random feasible, and Maskable Proximal Policy Optimization (PPO) approaches are evaluated as benchmarks. The deposited files are: code/run_analysis.py — Complete Python analysis script. The script implements the exact mixed-integer optimization, greedy and random benchmarks, PPO benchmark, same-spending and matched-action comparisons, unweighted, area-weighted, population-weighted, and vulnerability-weighted outcome calculations, CBG inclusion-threshold sensitivity, budget sensitivity, intervention-parameter and capacity sensitivity, project-size sensitivity, bootstrap Spearman confidence intervals, PPO seed uncertainty, and checkpoint diagnostics. data/equity_nbs_master_data.csv — Processed analysis-ready CBG-level dataset used by the script. The file contains 4,079 Maryland CBG records and the processed variables required for the analyses reported in the manuscript. The dataset includes the modeled flood-susceptibility and social, demographic, land-cover, hydrologic, terrain, intervention-capacity, and related variables required by the final allocation framework. The archive intentionally contains the processed analysis dataset rather than redistributing large third-party raw geospatial products. Original public datasets remain available from the sources, including the Federal Emergency Management Agency, U.S. Geological Survey, U.S. Fish and Wildlife Service National Wetlands Inventory, Maryland iMAP, and the National Land Cover Database. To run the deterministic analyses: python code/run_analysis.py --mode full --skip-rl To run the complete analysis including the PPO benchmark: python code/run_analysis.py --mode full To perform a quick deterministic test: python code/run_analysis.py --mode smoke --skip-rl The complete PPO benchmark is computationally intensive because 10 independent seeds are evaluated for each county and equity-weight combination.
🔗 Provenance — このレコードを発見したソース
- openalex https://doi.org/10.5281/zenodo.22018526first seen 2026-08-23 04:49:25 · last seen 2026-08-23 04:49:27
🔔 こうした論文の新着を逃したくない方は キーワードアラート に登録（無料・3キーワードまで）。
gxceed は公開メタデータに基づく研究支援データセットです。要約・翻訳・解説は AI 支援で生成されています。 最終的な解釈・検証は利用者が原典資料に基づいて行うことを前提とします。