Value allocation mechanism for multi-agent data sharing in digital innovation network: Agent-based modeling

To address the challenges of value distribution in multi-agent data sharing within digital innovation networks, this study proposes a fair and efficient allocation mechanism designed to optimize data resource flows and incentivize active participation among innovation actors. A comprehensive framework is constructed by integrating agent-based modeling (ABM) with complex network theory to capture the interactive dynamics among heterogeneous agents. The classical Susceptible-Infected-Removed (SIR) model is extended through the addition of a secondary diffusion layer to simulate both behavioral evolution and data-sharing propagation. Distributional fairness is quantitatively evaluated using the Lorenz curve and Theil index. Simulation results indicate that the incentive coefficient substantially enhances profit growth and adoption speed, while the profit-sharing coefficient critically shapes income distribution between platforms and innovators. The cost coefficient exhibits a nonlinear effect: moderate levels improve system efficiency, whereas excessive cost burdens diminish returns. Multiple linear regression analysis confirms the robustness of the model, and statistical significance tests validate the reliability of the results. These findings demonstrate the model’s effectiveness in balancing incentives and efficiency across heterogeneous stakeholders, providing both theoretical insights and practical guidance for strategic decision-making in the digital economy.