fresh-daugherty

An open, transparent, and fully reproducible reproduction of the modelling methods and case study of Daugherty (1991), built on ws3:

Daugherty, P. J. (1991). Credibility of Long Term Forest Planning: Dynamic Inconsistency in Linear Programming Based Forest Planning Models. PhD thesis, University of California, Berkeley.

Why this exists

Daugherty (1991) proves that linear-programming forest-planning models solved as open-loop formulations admit dynamically inconsistent plans — plans that fail Bellman's principle of optimality: a future planner, re-solving from the realized state under the same goals, would not follow the plan's tail. Such plans are not a credible basis for policy, and trade-off / shadow-price analysis derived from them is biased.

The thesis was never published peer-reviewed and is effectively inaccessible (archival hard copy only). As a result, the modelling trap it documents is repeatedly rediscovered, and the literature contains many LP forest-planning models presented (and cited) as rational dynamic-programming models that do not satisfy the Bellman time-stationarity condition. This project reproduces the thesis stack openly — open source, reproducible, transparent — so the result becomes citable and the trap detectable, and backs a short peer-reviewed paper.

What it does

• Reproduces the Daugherty (1991) open-loop harvest-scheduling LP in ws3 (Model I; the thesis's Model II formulation is not implemented), with the thesis's terminal constraints on every flow-constrained run.
• A sequential-replanning simulator that measures the occurrence and magnitude of dynamic inconsistency, with an objective-gap diagnostic that separates strictly suboptimal announced plans from alternate optima, under four replanning institutions (rolling or fixed horizon, reset or carried flow history).
• The case-study experiments across initial forest conditions, harvest policies (harvest-flow constraints), and interest rates, plus four extension experiments (declining discount rates, a calibrated harvest cap, a revenue-denominated flow, a rolling-mean flow). Consistent (closed-loop) solutions are not computed.

Status

v0.2.0 (phases P9-P19) is released and archived on Zenodo (concept DOI 10.5281/zenodo.21981433); it holds the records behind the manuscript (results/experiments/, documented in results/experiments/REPRODUCIBILITY.md; curated summaries in supplementary/). See ROADMAP.md and planning/v0.2.0-plan.md.

Quick Start

pip install -e ".[dev]"
fresh-daugherty --help

Related

• fresh-fuchs — the stochastic, risk-aware forest landscape-planning model whose inner-LP dynamic inconsistency motivated this reproduction (documented in its planning/dynamic-inconsistency-note.md).
• ws3 — the wood-supply engine this reproduction is built on.

Companion reference

The Model I/II/III formulations are documented in an accessible, citable source (cited alongside the thesis):

Gunn, E. A. (2007). Models for Strategic Forest Management. Chapter 16 in A. Weintraub, C. Romero, T. Bjørndal & R. Epstein (eds.), Handbook of Operations Research in Natural Resources, Springer, pp. 317-341.

Gunn (2007, pp. 331-332) cites Daugherty (1991) and describes the dynamic-inconsistency phenomenon directly (the "declining non-declining yield").

License

MIT, Copyright (c) 2026 UBC FRESH Lab.