AI Framework Optimizes Yunnan Pine Forest Management

Researchers have developed a coordinated optimization framework that integrates harvesting and replanting decisions for secondary Pinus yunnanensis forests, addressing a common limitation in traditional forest management models. Published in BMC Ecology and Evolution, the study focuses on secondary forests on Cangshan Mountain in Dali, where existing methods often treat harvesting and replanting as sequentially independent operations rather than spatially coupled decisions. The new approach uses a multi-agent deep reinforcement learning algorithm to solve the combinatorial problem of which trees to remove and where to add new trees within a unified framework, aiming to enhance the long-term structural benefits of replanting by anticipating future competition with existing trees.

The study identifies that conventional optimization approaches, such as genetic algorithms and simulated annealing, typically optimize harvesting first based on the current stand state and then add replanting using simple gap-filling rules. This sequential decomposition ignores critical interdependencies, specifically that replanting suitability depends on the spatial pattern of retained trees after harvesting. Furthermore, the long-term structural benefits of replanting depend on how future competition with existing trees is anticipated. Traditional single-stage optimization models are designed for a single decision variable set and cannot simultaneously solve the combinatorial problem of which trees to remove and where to add new trees within a unified framework.

To address these limitations, the researchers proposed a coordinated dynamic optimization framework based on MAPPO-BO, which stands for Multi-Agent Proximal Policy Optimization with Bayesian hyperparameter optimization. This framework employs a multi-agent deep reinforcement learning algorithm to achieve coordinated optimization through parallel decision-making and collaborative learning between harvesting and replanting agents.

In terms of replanting optimization, the study developed a slope-weighted replanting feasibility index, or SRFI, by incorporating slope weighting into the decision-making process. Replanting locations were dynamically determined in combination with stand density constraints for different tree species. The SRFI effectively guided replanting activities toward areas with relatively gentle topography.

The framework also introduced a stand structure prediction model based on two-period inventory data to establish a closed-loop feedback mechanism of optimization, prediction, and re-optimization. This mechanism enables the dynamic adjustment of stand structure over time.

Experimental results demonstrated that the MAPPO-BO framework improved the objective function by an average of 49.29% across five plots, outperforming the MADQN algorithm, which achieved an improvement of 47.17%. After dynamic optimization, the improvement in the objective function across three specific plots ranged from 44.43% to 64.69%. These results indicate that the proposed framework can effectively enhance the efficiency of coordinated harvesting and replanting decisions.

Multiple stand spatial structure indicators continuously improved following the application of the dynamic optimization framework, leading the stands to gradually approach the characteristics of natural forests. The study’s findings provide a new theoretical framework and methodological support for the precise management of secondary Pinus yunnanensis forests.

The study highlights the importance of treating harvesting and replanting as coupled decisions in forest management.

Subject of Research: Coordinated optimization of harvesting and replanting in secondary Pinus yunnanensis forests using multi-agent reinforcement learning.

Article Title: Dynamic stand structure optimization of secondary Pinus yunnanensis forests integrating two-period structural prediction and harvesting-replanting coordination

Article References: Wang, J., Lv, C., Yin, J., Chen, Y., Lv, S., & Wu, B. (2026). Dynamic stand structure optimization of secondary Pinus yunnanensis forests integrating two-period structural prediction and harvesting-replanting coordination. BMC Ecology and Evolution. https://doi.org/10.1186/s12862-026-02565-1

Image Credits: AI Generated

DOI: 10.1186/s12862-026-02565-1

Keywords: Pinus yunnanensis, Forest Management, Reinforcement Learning, MAPPO-BO, Stand Structure, Sustainable Forestry, Harvesting Optimization, Replanting Strategy, Cangshan Mountain, BMC Ecology and Evolution, Dynamic, stand