RL-Driven Sustainable Land-Use Allocation for the Lake Malawi Basin

Ying Yao

cs.AIApr 4, 2026

RL-Driven Sustainable Land-Use Allocation for the Lake Malawi Basin

Authors: Ying Yao

Organizations: Georgia Institute of Technology

Abstract

Unsustainable land-use practices in ecologically sensitive regions threaten biodiversity, water resources, and the livelihoods of millions. This paper presents a deep reinforcement learning (RL) framework for optimizing land-use allocation in the Lake Malawi Basin to maximize total ecosystem service value (ESV). Drawing on the benefit transfer methodology of Costanza et al., we assign biome-specific ESV coefficients -- locally anchored to a Malawi wetland valuation -- to nine land-cover classes derived from Sentinel-2 imagery. The RL environment models a 50x50 cell grid at 500m resolution, where a Proximal Policy Optimization (PPO) agent with action masking iteratively transfers land-use pixels between modifiable classes. The reward function combines per-cell ecological value with spatial coherence objectives: contiguity bonuses for ecologically connected land-use patches (forest, cropland, built area etc.) and buffer zone penalties for high-impact development adjacent to water bodies. We evaluate the framework across three scenarios: (i) pure ESV maximization, (ii) ESV with spatial reward shaping, and (iii) a regenerative agriculture policy scenario. Results demonstrate that the agent effectively learns to increase total ESV; that spatial reward shaping successfully steers allocations toward ecologically sound patterns, including homogeneous land-use clustering and slight forest consolidation near water bodies; and that the framework responds meaningfully to policy parameter changes, establishing its utility as a scenario-analysis tool for environmental planning.

Figures & tables

Class	Type	ET (kg/m 2 /yr)
Water	Protected	616.93
Trees	Modifiable	933.57
Flooded	Protected	767.62
Crops	Modifiable	675.66
Built Area	Modifiable	648.95
Bare Ground	Modifiable	591.53

TABLE I: Per-Class Evapotranspiration (2024, MOD16A3GF)

Class	Type	ESV	Ratio
		(USD/ha/yr)
Water	Protected	554	1.00
Trees	Modifiable	238	0.43
Flooded	Protected	1,136	2.05
Crops	Modifiable	246 * 1.35 = 332	0.60
Built Area	Modifiable	295	0.53

TABLE II: Land-Cover Classes and Ecosystem Service Values. Crops carries the

1.35\times

regenerative-agriculture uplift used in the headline reward; the base value

246

is used only in the Appendix A ablation.

Fig. 2: Overview of the proposed RL framework. The agent observes a

10\!\times\!10\!\times\!5

land-use fraction grid, extracts spatial features via a shared GridCNN, and produces both a masked policy (Actor) and state value estimate (Critic). The environment executes the selected action and returns a reward combining ESV change and spatial coherence metrics.

Component	Symbol	Weight
Tree contiguity bonus	$w_{T}$	1.0
Crop contiguity bonus	$w_{C}$	4.0
Built-area contiguity bonus	$w_{B}$	2.0
Water buffer penalty	$w_{W}$	6.0 (anneal $1{\rightarrow}6$ )
Riparian-tree bonus	$w_{R}$	5.0

TABLE III: Spatial Reward Weights.

w_{W}

is annealed linearly from

1

6

over the first

60\%

of training; the other weights are held fixed.

Appendix figures & tables2 assets

Supplementary material from the paper’s appendix.

Appendix

	ExpI	ExpII	ExpIII
	Eco-only	Spatial-without-regen	Headline (Spatial+Regen)
Dominant conversion	$\rightarrow$ Built	Mixed	$\rightarrow$ Crops
Forest near water	Not obvious	Increased	Increased
Built expansion	Aggressive	Clustered	Clustered, restrained
Crop allocation	Moderate	Moderate	High
Ecological realism	Low	High	High

Appendix

TABLE IV: Qualitative Behaviour Across Reward-Design Ablations

Parameter	Value
Learning rate	$5{\times}10^{-5}\rightarrow 5{\times}10^{-6}$ (last 33%)
Rollout length ( $n_{\text{steps}}$ )	2,048
Mini-batch size	128
PPO epochs	10
Discount ( $\gamma$ )	0.99
GAE $\lambda$	0.95

Appendix

TABLE V: PPO Hyperparameters (spatially-aware runs).

KC-BFPRL: Knowledge-Guided Multi-UAV Collaboration for Grassland Restoration via Bilevel Formerpointer-Based Reinforcement Learning
Aug 17, 2026Dongbin Jiao, Xianyi Wang, Yuchen Yuan +5Reinforcement Learning Multi-Agent Coordination
School of Information Science and Engineering, Lanzhou University, Lanzhou, 730000, P. R. China · Key Laboratory of Tourism Information Fusion Processing and Data Ownership Protection, Ministry of Culture and Tourism, Lanzhou University, Lanzhou, 730000, P. R. China · School of Automobile, Chang’an University, Xi’an, 710064, P. R. China +2
EcoFair-CH-MARL: Scalable Constrained Hierarchical Multi-Agent RL with Real-Time Emission Budgets and Fairness Guarantees
Mar 15, 2026Saad AlqithamiConstrained MARL Multi-Agent Reinforcement Learning
Computer Science Department, Al-Baha University, Albaha 65779, Saudi Arabia
Auction-Based RIS Allocation With DRL: Controlling the Cost-Performance Trade-Off
Feb 18, 2026Martin Mark Zan, Stefan SchwarzWireless Resource Allocation Reconfigurable Intelligent Surfaces
Institute of Telecommunications, TU Wien, Vienna, Austria

Aug 17, 2026J/K move · Enter open · S save

KC-BFPRL: Knowledge-Guided Multi-UAV Collaboration for Grassland Restoration via Bilevel Formerpointer-Based Reinforcement Learning

Dongbin Jiao, Xianyi Wang, Yuchen Yuan, Weibo Yang+4

School of Information Science and Engineering, Lanzhou University, Lanzhou, 730000, P. R. China · Key Laboratory of Tourism Information Fusion Processing and Data Ownership Protection, Ministry of Culture and Tourism, Lanzhou University, Lanzhou, 730000, P. R. China · School of Automobile, Chang’an University, Xi’an, 710064, P. R. China+2

Reinforcement Learning Multi-Agent Coordination UAV Trajectory Optimization Multi-Robot Task Allocation Hierarchical RL

Open paper AskPDF arXiv

Multi-unmanned aerial vehicle (UAV) systems provide scalable service platforms for large-scale environmental tasks, such as grassland ecosystem restoration. However, coordinating fleet operations requires solving the restoration area maximization problem (RAMP). This non-linear combinatorial optimization challenge is complicated by payload-dependent energy dynamics and heterogeneous ecological degradation. We propose a novel knowledge-guided collaborative bilevel formerpointer reinforcement learning framework (KC-BFPRL) to address this complexity. Using a hierarchical paradigm, KC-BFPRL decomposes RAMP into global task allocation and local restoration planning, with the latter further divided into upper-level trajectory planning and lower-level restoration area allocation. Our specialized architecture pairs featuring a Transformer-based encoder that fuses static environmental features with dynamic UAV states, and a Pointer Network decoder trained via a robust actor-critic framework. By embedding ecological priority rules and heuristic logic, KC-BFPRL achieves a structured warm-start, solving the RL cold-start problem while ensuring strict constraint satisfaction. Extensive experiments demonstrate that KC-BFPRL consistently outperforms state-of-the-art baselines, achieving superior objective values and efficiency. It maintains a

0.00\%

optimality gap in the most complex scenarios U8-R160 and operates nearly three times faster than MAPDP, validating its robustness, scalability, and real-time applicability for large-scale automated ecological restoration.

Abstract

Explore similar work

KC-BFPRL: Knowledge-Guided Multi-UAV Collaboration for Grassland Restoration via Bilevel Formerpointer-Based Reinforcement Learning

EcoFair-CH-MARL: Scalable Constrained Hierarchical Multi-Agent RL with Real-Time Emission Budgets and Fairness Guarantees

Auction-Based RIS Allocation With DRL: Controlling the Cost-Performance Trade-Off