(339b) Agent-Based Modeling Analysis of Biomass Feedstock Production System Dynamics and Resilience | AIChE

(339b) Agent-Based Modeling Analysis of Biomass Feedstock Production System Dynamics and Resilience


Rodriguez, L. - Presenter, University of Illinois at Urbana-Champaign
Hansen, A. - Presenter, University of Illinois at Urbana-Champaign
Ting, K. - Presenter, University of Illinois at Urbana-Champaign

An efficient and sustainable energy crop production system is essential for the success of the emerging biorefinery industry. However, there is currently a significant lack of understanding regarding the development and dynamics of the biomass feedstock market upstream of the biorefinery, which is a critical component of the value chain. This includes assessing the time-scale of its development as well as the impact of various engineering, technological, regulatory and social factors on the dynamics. Various uncertainties and disturbances can impact the functioning of this system, and understanding the resilience of the system in the presence of those is also important. The analysis is especially challenging due to the involvement of multiple stakeholders such as farmers, refinery, transportation companies and storage elevators that must participate and yet compete for a fraction of the profit.

The objective of this work is to conduct such as analysis using the complex adaptive systems theory approach. The work develops an agent-based model that incorporates various stakeholders (agents) within the feedstock production system. The model, developed in MATLAB®, uses an object-oriented programming approach by considering each agent-type as an independent class and a particular realization of that agent-type as an object (instance) of that class. The individual characteristics of an object are defined by a set of attributes. The behavior of each agent as well as its interaction with other agents is based on a set of rules (methods). Since the behavioral and social factors play an important role in the decisions made by individual agents, a mental model incorporating some of these factors is also developed for each agent. The set of decision rules are consequently based not only on the economic and technological rationale, but also on the unique mental models and cognitive processes. The learning/adaptive behavior exhibited by the agents is also modeled which enables the simulation of emergent behavior of the system.

The model has previously been used to study the development of the biomass market under different scenarios. The focus of this work is on understanding the impact of various disturbances and uncertainties on the dynamics of the system, which enables us to study the resilience of the system. Different possible disturbances such as loss of productivity, price fluctuations and infrastructure breakdown will be modeled. Desirable but sudden changes such as the availability of an efficient technology will also be considered. Simulation studies incorporating these disturbances will be conducted to study their impact on the dynamics of the feedstock production system. Preliminary studies using this model have illustrated that the development of this sector is gradual, nonlinear and may take up to 15 years. The impact of the time of disturbance along this development trajectory will therefore be critical and will also be studied. Simulations for different types of energy crops such as switchgrass and Miscanthus will also identify whether the dynamics are a function of the energy crop. The work will also enable us to evaluate different system resilience measures appropriate for analyzing the feedstock production system.