Simulation and optimization of organic fertilizer subsidy policy response in reservoir-type water source areas on agent-based model

To address the challenges of insufficient incentives and low coordination efficiency in the governance of agricultural non-point source pollution in reservoir-type water source areas, this study develops a multi-agent simulation framework that incorporates key stakeholders, including farmers, agricultural product buyers, water utilities, and government authorities. By accounting for farmer heterogeneity and fiscal incentive mechanisms, the model systematically evaluates the economic outcomes for different actors and the environmental impacts under various organic fertilizer subsidy scenarios. The simulation case results show that a moderate subsidy intensity (60%-70%) maximizes multi-party benefits, facilitates farmer transition, and encourages market participation. In contrast, low subsidy levels fail to drive change, while excessively high subsidies yield diminishing marginal returns despite improved environmental outcomes. The proposed model provides a quantitative tool for optimizing organic fertilizer subsidy policies, helping policymakers balance fiscal inputs with environmental governance outcomes, and supporting ecological protection and green development of water source areas.