An agent-based model of a cellulosic biofuel system involving multiple stakeholder communities

Abstract:

The success of the cellulosic bioenergy sector will depend on the transition from the current food crop dominated agricultural system to one with diversified land uses. The transition will affect and be affected by the various stakeholder communities directly and indirectly involved in this cellulosic bioeconomy, including farmers, refinery plants, local communities, government and non-government agencies. More specifically, the complex interactions among stakeholders and the spillover effects, i.e., the ‘stakeholder synergy’, largely control the evolution of cellulosic bioenergy development. Previous studies investigating such a complex system usually adopted on a top-down approach that is driven by economic benefits, and hardly considered the heterogeneous decision-making processes at the individual-level under the various economic, environmental, and social conditions. The aim of this study is to test stakeholder synergy and simulate its impact on a cellulosic biofuel system through a bottom-up agent-based modeling (ABM) approach. The ABM simulates heterogeneous decisions made by the various stakeholders, the interactions and feedbacks between the stakeholders, and the interactions between human agents and the environment. Agent behavior rules are derived using a combined machine and human intelligence approach that takes advantage of both empirical data such as land-use survey and expert knowledge. A case study in the Sangamon River Basin in Central Illinois shows the impact of stakeholder synergy and evolution of the cellulosic biofuel system triggered by different policy, technology, and behavior factors.

The figure provides a demonstration of the agent-based model simulation showing the spatial distributions of land use and bio-refinery decisions at the end of simulation (a), and temporal evolution of farmers’ average attitude toward environment and perennial grass (b), total biofuel production (c), total nitrogen loading (d), and total acreage of different land uses (e).