Agent-based Modeling Platform Research Articles

Agent-based model of the impact of higher influenza vaccine efficacy on seasonal influenza burden

Current influenza vaccines have limited effectiveness. COVID-19 vaccines using mRNA technology have demonstrated very high efficacy, suggesting that mRNA vaccines could be more effective for influenza. Several such influenza vaccines are in development. FRED, an agent-based modeling platform, was used to estimate the impact of more effective influenza vaccines on seasonal influenza burden. Simulations were performed using an agent-based model of influenza that included varying levels of vaccination efficacy (40-95 % effective). In some simulations, level of infectiousness and/or length of infectious period in agents with breakthrough infections was also decreased. Impact of increased and decreased levels of vaccine uptake were also modeled. Outcomes included number of symptomatic influenza cases estimated for the US. Highly effective vaccines significantly reduced estimated influenza cases in the model. When vaccine efficacy was increased from 40% to a maximum of 95%, estimated influenza cases in the US decreased by 43% to>99%. The base simulation (40% efficacy) resulted in∼28 million total yearly cases in the US, while the most effective vaccine modeled (95% efficacy) decreased estimated cases to∼22,000. Highly effective vaccines could dramatically reduce influenza burden. Model estimates suggest that even modest increases in vaccine efficacy could dramatically reduce seasonal influenza disease burden.

Can a Two-Dose Influenza Vaccine Regimen Better Protect Older Adults? An Agent-Based Modeling Study.

Older adults (age ≥ 65) are at high risk of influenza morbidity and mortality. This study evaluated the impact of a hypothetical two-dose influenza vaccine regimen per season to reduce symptomatic flu cases by providing preseason (first dose) and mid-season (second dose) protection to offset waning vaccine effectiveness (VE). The Framework for Reconstructing Epidemiological Dynamics (FRED), an agent-based modeling platform, was used to compare typical one-dose vaccination to a two-dose vaccination strategy. Primary models incorporated waning VE of 10% per month and varied influenza season timing (December through March) to estimate cases and hospitalizations in older adults. Additional scenarios modeled reductions in uptake and VE of the second dose, and overall waning. In seasons with later peaks, two vaccine doses had the largest potential to reduce cases (14.4% with February peak, 18.7% with March peak) and hospitalizations (13.1% with February peak, 16.8% with March peak). Reductions in cases and hospitalizations still resulted but decreased when 30% of individuals failed to receive a second dose, second dose VE was reduced, or overall waning was reduced to 7% per month. Agent-based modeling indicates that two influenza vaccine doses could decrease cases and hospitalizations in older individuals. The highest impact occurred in the more frequently observed late-peak seasons. The beneficial impact of the two-dose regimen persisted despite model scenarios of reduced uptake of the second dose, decreased VE of the second dose, or overall VE waning.

Elaborating on anti CTLA-4 mechanisms of action using an agent-based modeling approach

Anti CTLA-4 therapy is aimed at blocking the Cytotoxic T-lymphocyte antigen-4 (CTLA-4), a key cancer immunity cycle checkpoint. The mechanism of action of CTLA-4 may be described as a dynamic competition for the B7 ligand which, subsequently, interferes with the CD28-B7 costimulatory pathway. Anti CTLA-4 blockade enhances the process of cognate T cell activation and leads to a broadening of the T cell repertoire. In the present work, we used an agent-based modeling (ABM) platform of T cell immune response development, to explore hypothetical modes of anti CTLA-4 action. The model features a selected number of activated T cell clones, calculated based on combined random and chemotactically-driven encounters with antigen-presenting dendritic cells (DCs) and a distribution of individual T cell affinities to the antigen of interest. The proposed model can be used as a quantitative tool to explore various hypotheses on T cell immunity regulation and validate these against experimental data. A comprehensive ABM model analysis of immune response dynamic simulations revealed several putative anti CTLA-4 mechanisms of action, including: (i) an increase in the probability of primary activation of lymphocytes; (ii) T cell activation enhancement via a prolongation of short contacts with dendritic cells; and (iii) an increase in the maximum level of activation signal (or saturation), accumulated through a series of short contacts with DCs. The modeling work further demonstrates that it is only when considering jointly these various modes of anti CTLA-4 effects on the T cell immune response dynamics that a biologically meaningful increase in both the production of activated cells and the expansion of the T cell repertoire is observed. These model-based results are overall consistent with the collective biological knowledge on the functional role of CTLA-4. Furthermore, the ABM presented here may allow to interrogate various mechanistic scenarios underlying adverse events mediated by anti CTLA-4 pharmacologic therapies.

Conceptual ex ante simulation for Green Stormwater Infrastructure Adoption on private property using agent-based modeling.

This paper presents the Green Stormwater Infrastructure Social Spatial Adoption (G-SSA) model, a cellular automata agent-based model that simulates the behavior of private property owners responding to incentives to adopt on-site green stormwater infrastructure (GSI). Concepts such as small-world social networks, opinion dissemination, and diffusion of innovation are used to capture dynamic social, spatial, and temporal aspects of green stormwater infrastructure adoption. Demographic information, site constraints, GSI practice type and costs, and financial incentive information are integrated into modeling rules that influence adoption dynamics. A methodology is presented that describes how these concepts have been translated into an agent-based modeling platform that provides the opportunity to explore modeling dynamics and output. Model output confirms the viability of the methodology and produces results that inform future efforts to explore the G-SSA model. PRACTITIONER POINTS: Agent-based modeling (ABM) can evaluate the expected impact of market-based approaches for green stormwater infrastructure (GSI) adoption and O&M services. ABM can simulate years of implementation of GSI program incentives, which vary from stormwater fee reduction to subsidy payments to tradable credits revenues generated. Publicly available demographic data combined with behavioral economic relationships can build models to evaluate how municipalities can meet regulatory goals for urban retrofits using market-based approaches to encourage GSI adoption.

COVID-19 Modeling: High-Resolution Agent-Based Modeling of COVID-19 Spreading in a Small Town (Adv. Theory Simul. 3/2021).

Since 2020, COVID-19 has wreaked havoc across the planet, taking the lives of more than one million people. The uncertainty and novelty of the current conditions call for the development of theory and simulation tools that can support effective policy-making. In article number 2000277, Agnieszka Truszkowska, Maurizio Porfiri, and co-workers report a high-resolution, agent-based modeling platform to simulate the spreading of COVID-19 in the city of New Rochelle, NY-one of the first outbreaks registered in the United States. Image by Anna Sawulska, Agnieszka Truszkowska, Beata Truszkowska, and Maurizio Porfiri.

Quantifying Uncertainty in Exposure to Coastal Hazards Associated with Both Climate Change and Adaptation Strategies: A U.S. Pacific Northwest Alternative Coastal Futures Analysis

Coastal communities face heightened risk to coastal flooding and erosion hazards due to sea-level rise, changing storminess patterns, and evolving human development pressures. Incorporating uncertainty associated with both climate change and the range of possible adaptation measures is essential for projecting the evolving exposure to coastal flooding and erosion, as well as associated community vulnerability through time. A spatially explicit agent-based modeling platform, that provides a scenario-based framework for examining interactions between human and natural systems across a landscape, was used in Tillamook County, OR (USA) to explore strategies that may reduce exposure to coastal hazards within the context of climate change. Probabilistic simulations of extreme water levels were used to assess the impacts of variable projections of sea-level rise and storminess both as individual climate drivers and under a range of integrated climate change scenarios through the end of the century. Additionally, policy drivers, modeled both as individual management decisions and as policies integrated within adaptation scenarios, captured variability in possible human response to increased hazards risk. The relative contribution of variability and uncertainty from both climate change and policy decisions was quantified using three stakeholder relevant landscape performance metrics related to flooding, erosion, and recreational beach accessibility. In general, policy decisions introduced greater variability and uncertainty to the impacts of coastal hazards than climate change uncertainty. Quantifying uncertainty across a suite of coproduced performance metrics can help determine the relative impact of management decisions on the adaptive capacity of communities under future climate scenarios.

Exploring the common ground of landscape ecology and landscape archaeology through a case study from eastern Anatolia, Turkey

Landscape archaeology has a lot to offer to landscape ecology, being an interdisciplinary approach that emphasizes the study of long-term human–environment dynamics. We outline different conceptualizations of landscape in landscape archaeology and illustrate the potential of the approach for collaborating with landscape ecologists by describing a case study from the multi-period site of Arslantepe, located in the Malatya province of eastern Anatolia, Turkey. We use an agent-based modeling platform to understand the socio-economic transformations at Arslantepe during the Early Bronze Age-I. These simulations revealed long-term dynamics of grassland and woodland under different climate and population scenarios. It was found that both grassland and woodland responded most strongly to changes in population, with woodlands being more sensitive. Further, it becomes evident that the adapted site-tethered pastoralism could have brought more sustainable land use practices. The example shows the tremendous potential landscape archaeology has for studying long-term sustainability issues, especially related to modes of production. The landscape archaeological perspective can be linked with expertise provided by landscape ecologists, and we propose more in-depth collaboration of these two fields that offer diverse yet complementary perspectives.

Simulating long-term performance of regional distribution centers in archipelagic logistics systems

Indonesia has reevaluated its archipelagic logistics system to implement a strategic shift from a point-to-point network to a hub-and-spoke network, with a plan to establish seven new regional distribution centers (RDCs). To supplement the previous qualitative assessment of the plan and to potentially justify continuing with its implementation, this paper aims to evaluate the prospective long-term performance of the planned RDCs within a hub-and-spoke (HS) network. The chosen methodology for this analysis is a hybrid of optimization and simulation, using an agent-based modeling and simulation platform that incorporates a geographic information system (GIS) element to provide a realistic geographical context. This constitutes a relatively novel innovation in logistics research. Experiments were developed to estimate the logistics performance of the hub-and-spoke network and planned RDCs, as well as to identify potential problems such as bottlenecks at the RDCs and the relative (in)efficiency of the selected multimodal transport. Transport costs, inventory costs, and the order backlog of RDCs are evaluated based on the average results from 100 replications. Outcomes from the model are validated against empirical data on Indonesia’s logistics costs from the World Bank. Model results indicate that the hub-and-spoke network performs better than the point-to-point network when it comes to transportation costs. Since this is found to be the dominant element of total logistics costs, appropriate modal choice and route optimization are, therefore, critical to reducing logistics costs. The planned RDCs are found to have significant imbalance in their loadings, which is likely to produce bottlenecks. The paper concludes that a planned move to a hub-and spoke system is appropriate and that the planned investment in the RDCs is therefore justified. However, there is a need to streamline the predicted loadings at the RDCs to avoid bottlenecks. Our results suggest that the government should reopen the case and reevaluate the locations and coverage of each RDC. More generically, it is concluded that the application of hybrid optimization and simulation using agent-based modeling and simulation is feasible and the methodological approach adopted herein is generalizable to other archipelagic logistics systems.

A data-driven approach to characterize the impact of connected and autonomous vehicles on traffic flow

ABSTRACT This study presents a model to characterize changes in network traffic flows as a result of implementing connected and autonomous vehicle (CAV) technology based on traffic network and built-environment characteristics. To develop such a model, first, the POLARIS agent-based modeling platform is used to predict changes in average daily traffic (ADT) under CAV scenario in the road network of Chicago metropolitan area as the dependent variable of the model. Second, a comprehensive set of variables and indicators representing network characteristics and urban structure patterns are generated. Finally, three machine learning techniques, namely, K-Nearest neighbors, Random Forest, and eXtreme Gradient Boosting, are used to characterize changes in ADT based on network characteristics under a CAV scenario. The estimated models are validated and are found to yield acceptable performance. In addition, SHapley Additive exPlanations (SHAP) analysis tool is employed to investigate the impact of important features on changes in ADT.

Sociomathematical Norms for Integrating Coding and Modeling with Elementary Science: A Dialogical Approach

In recent years, the field of education has challenged researchers and practitioners to incorporate computing as an essential focus of K-12 STEM education. Integrating computing within K-12 STEM supports learners of all ages in codeveloping and using computational thinking in existing curricular contexts alongside practices essential for developing mathematical and scientific expertise. In this paper, we present findings from a design-based, microgenetic study in which an agent-based programming and computational modeling platform—ViMAP—was integrated with existing elementary science and math curricula through lessons co-designed and taught by the classroom teacher across a period of seven months. We present a dialogical re-positioning of coding, where disciplinarily grounded meanings of code emerge through the construction of computational utterances––i.e., computer models as well as complementary conversations and physical models that serve as mathematical and scientific explanations––through the use of socio-mathematical norms.

Experiments and Agent Based Models of Zooplankton Movement within Complex Flow Environments

The movement of plankton is often dictated by local flow patterns, particularly during storms and in environments with strong flows. Reefs, macrophyte beds, and other immersed structures can provide shelter against washout and drastically alter the distributions of plankton as these structures redirect and slow the flows through them. Advection–diffusion and agent-based models are often used to describe the movement of plankton within marine and fresh water environments and across multiple scales. Experimental validation of such models of plankton movement within complex flow environments is challenging because of the difference in both time and spatial scales. Organisms on the scale of 1 mm or less swim by beating their appendages on the order of 1 Hz and are advected meters to kilometers over days, weeks, and months. One approach to study this challenging multiscale problem is to insert actively moving agents within a background flow field. Open source tools to implement this sort of approach are, however, limited. In this paper, we combine experiments and computational fluid dynamics with a newly developed agent-based modeling platform to quantify plankton movement at the scale of tens of centimeters. We use Artemia spp., or brine shrimp, as a model organism given their availability and ease of culturing. The distribution of brine shrimp over time was recorded in a flow tank with simplified physical models of macrophytes. These simplified macrophyte models were 3D-printed arrays of cylinders of varying heights and densities. Artemia nauplii were injected within these arrays, and their distributions over time were recorded with video. The detailed three-dimensional flow fields were quantified using computational fluid dynamics and validated experimentally with particle image velocimetry. To better quantify plankton distributions, we developed an agent-based modeling framework, Planktos, to simulate the movement of plankton immersed within such flow fields. The spatially and temporally varying Artemia distributions were compared across models of varying heights and densities for both the experiments and the agent-based models. The results show that increasing the density of the macrophyte bed drastically increases the average time it takes the plankton to be swept downstream. The height of the macrophyte bed had less of an effect. These effects were easily observed in both experimental studies and in the agent-based simulations.

Integrating Machine Learning With Microsimulation to Classify Hypothetical, Novel Patients for Predicting Pregabalin Treatment Response Based on Observational and Randomized Data in Patients With Painful Diabetic Peripheral Neuropathy.

PurposeVariability in patient treatment responses can be a barrier to effective care. Utilization of available patient databases may improve the prediction of treatment responses. We evaluated machine learning methods to predict novel, individual patient responses to pregabalin for painful diabetic peripheral neuropathy, utilizing an agent-based modeling and simulation platform that integrates real-world observational study (OS) data and randomized clinical trial (RCT) data.Patients and methodsThe best supervised machine learning methods were selected (through literature review) and combined in a novel way for aligning patients with relevant subgroups that best enable prediction of pregabalin responses. Data were derived from a German OS of pregabalin (N=2642) and nine international RCTs (N=1320). Coarsened exact matching of OS and RCT patients was used and a hierarchical cluster analysis was implemented. We tested which machine learning methods would best align candidate patients with specific clusters that predict their pain scores over time. Cluster alignments would trigger assignments of cluster-specific time-series regressions with lagged variables as inputs in order to simulate “virtual” patients and generate 1000 trajectory variations for given novel patients.ResultsInstance-based machine learning methods (k-nearest neighbor, supervised fuzzy c-means) were selected for quantitative analyses. Each method alone correctly classified 56.7% and 39.1% of patients, respectively. An “ensemble method” (combining both methods) correctly classified 98.4% and 95.9% of patients in the training and testing datasets, respectively.ConclusionAn ensemble combination of two instance-based machine learning techniques best accommodated different data types (dichotomous, categorical, continuous) and performed better than either technique alone in assigning novel patients to subgroups for predicting treatment outcomes using microsimulation. Assignment of novel patients to a cluster of similar patients has the potential to improve prediction of patient outcomes for chronic conditions in which initial treatment response can be incorporated using microsimulation.Clinical trial registrieswww.clinicaltrials.gov: NCT00156078, NCT00159679, NCT00143156, NCT00553475.

The Impacts of Emission Reduction Targets in Indonesia Electricity Systems

Climate change policy often contradicts the least-cost objective of electricity generation in developing countries. The objective of our study is to propose electricity generation mixes that can meet emission reduction targets in Indonesia. We estimate the optimal generation mix, costs, and emissions from three scenarios, namely existing power plant planning, and 11% and 14% emission reductions in Indonesia’s electricity sector. The estimations are based on linear programming, input-output analysis, and life-cycle analysis, integrated into an agent-based modeling (ABM) platform. The simulation results confirm the existing power plant planning, which is dominated by coal-based power plants, as the lowest-cost scenario in the short-term; however, this scenario also produces the highest emissions. Emission reduction scenarios have lower emissions due to a higher share of renewables and, therefore, the Indonesian electricity system is robust from fossil fuel price increases. In the long-term, costs incurred in the emission reduction scenarios will be lower than electricity generation costs under the existing power plant planning. Our findings should be a basis for re-evaluating energy policies, power plant planning, and the research agenda in Indonesia.
 Keyword: linear programming, agent-based modelling (ABM), input-output analysis, life-cycle analysis

Cost-effective electric vehicle charging infrastructure siting for Delhi

Plug-in electric vehicles (PEVs) represent a substantial opportunity for governments to reduce emissions of both air pollutants and greenhouse gases. The Government of India has set a goal of deploying 6–7 million hybrid and PEVs on Indian roads by the year 2020. The uptake of PEVs will depend on, among other factors like high cost, how effectively range anxiety is mitigated through the deployment of adequate electric vehicle charging stations (EVCS) throughout a region. The Indian Government therefore views EVCS deployment as a central part of their electric mobility mission. The plug-in electric vehicle infrastructure (PEVI) model—an agent-based simulation modeling platform—was used to explore the cost-effective siting of EVCS throughout the National Capital Territory (NCT) of Delhi, India. At 1% penetration in the passenger car fleet, or ∼10 000 battery electric vehicles (BEVs), charging services can be provided to drivers for an investment of $4.4 M (or $440/BEV) by siting 2764 chargers throughout the NCT of Delhi with an emphasis on the more densely populated and frequented regions of the city. The majority of chargers sited by this analysis were low power, Level 1 chargers, which have the added benefit of being simpler to deploy than higher power alternatives. The amount of public infrastructure needed depends on the access that drivers have to EVCS at home, with 83% more charging capacity required to provide the same level of service to a population of drivers without home chargers compared to a scenario with home chargers. Results also depend on the battery capacity of the BEVs adopted, with approximately 60% more charging capacity needed to achieve the same level of service when vehicles are assumed to have 57 km versus 96 km of range.

An agent-based modeling optimization approach for understanding behavior of engineered complex adaptive systems

The objective of this study is to present a formal agent-based modeling (ABM) platform that enables managers to predict and partially control patterns of behaviors in certain engineered complex adaptive systems (ECASs). The approach integrates social networks, social science, complex systems, and diffusion theory into a consumer-based optimization and agent-based modeling (ABM) platform. Demonstrated on the U.S. electricity markets, ABM is integrated with normative and subjective decision behavior recommended by the U.S. Department of Energy (DOE) and Federal Energy Regulatory Commission (FERC). Furthermore, the modeling and solution methodology address shortcomings in previous ABM and Transactive Energy (TE) approaches and advances our ability to model and understand ECAS behaviors through computational intelligence. The mathematical approach is a non-convex consumer-based optimization model that is integrated with an ABM in a game environment.

An Agent-Based Model for Addressing the Impact of a Disaster on Access to Primary Care Services.

Hurricane Sandy in the Rockaways, Queens, forced residents to evacuate and primary care providers to close or curtail operations. A large deficit in primary care access was apparent in the immediate aftermath of the storm. Our objective was to build a computational model to aid responders in planning to situate primary care services in a disaster-affected area. Using an agent-based modeling platform, HAZEL, we simulated the Rockaways population, its evacuation behavior, and primary care providers' availability in the aftermath of Hurricane Sandy. Data sources for this model included post-storm and community health surveys from New York City, a survey of the Rockaways primary care providers, and research literature. The model then tested geospatially specific interventions to address storm-related access deficits. The model revealed that areas of high primary care access deficit were concentrated in the eastern part of the Rockaways. Placing mobile health clinics in the most populous census tracts reduced the access deficit significantly, whereas increasing providers' capacity by 50% reduced the deficit to a lesser degree. An agent-based model may be a useful tool to have in place so that policy makers can conduct scenario-based analyses to plan interventions optimally in the event of a disaster. (Disaster Med Public Health Preparedness. 2016;10:386-393).

(174) Development of an agent-based modeling and simulation platform for improving therapeutic outcomes in patients with painful diabetic peripheral neuropathy

(174) Development of an agent-based modeling and simulation platform for improving therapeutic outcomes in patients with painful diabetic peripheral neuropathy

Agent-Based Modeling of Biological Pathways - A Case-Study on mRNA Export and Quality Control Mechanism

Currently, a substantial gap exists between the methods utilized for modeling biological pathways. While some methods rely heavily on assumptions that prevent them from capturing the spatial detail and stochastic behavior observed in nature, others are designed to account for deeper levels of spatial details and stochasticity but at the cost of increased computational time and resources. Agent-based modeling (ABM) is a highly promising approach that brings together the advantages of detailed spatiotemporal approaches and computationally inexpensive methods. ABMs consist of a collection of agents with governing rules that dictate local behavior and interactions with adjacent agents, eventually predicting the complex behavior that may not be obvious from the individual rules. Our recently developed ABM platform, called BioABM, is primarily designed to capture the dynamics of molecular-scale complex systems by associating diffusion and binding/unbinding constants of biological species to movement and interaction probabilities in ABM. In order to showcase its potential impact as a predictive tool, we apply BioABM to an important, yet not well understood, biological pathway, i.e. messenger ribonucleic acid (mRNA) export and quality control mechanism. mRNAs are transported to the cytoplasm to transfer genetic information and direct synthesis of functional proteins. Multiple proteins and complexes are involved and the process is meticulously quality controlled by various sophisticated, yet highly efficient, mechanisms in eukaryotic cells. Using BioABM, we explored the dynamics of mRNA export and the mechanism through which aberrant mRNAs are distinguished and retained inside the nucleus. Our results shed light on different aspects of this pathway; addressing factors such as the significance of the expression level of mRNA export factor and the length of mRNA as well as presence of a nuclear basket-associated quality control complex on mRNA export dynamics.

Evaluating Off-Peak Pricing Strategies in Public Transportation with an Activity-Based Approach

Public transportation authorities across the world are implementing various peak and off-peak pricing strategies to manage travel demand and improve the overall system performance. In this study, an activity-based demand framework was used to evaluate two off-peak pricing strategies currently in use in Singapore. These strategies consisted of a free prepeak travel on mass rapid transit (MRT) and an off-peak discount for an integrated transit (public buses and MRT). Smart card data collected before and after the implementation of the first policy were used to calibrate the behavioral models involved, to capture travelers’ preferences and choices properly. To evaluate both pricing strategies, a comprehensive set of key performance indicators was considered and included the changes in peak ridership, average trip fare, operator’s revenue, the number of public transportation trips, and mode share. The results indicate that off-peak discount pricing strategies are a viable policy option for spreading demand peaks and that they are more effective during the afternoon peak period. This study also demonstrates the capabilities and the advantages of an agent-based modeling platform, SimMobility, as a tool for policy analysis.

An interactive land use transition agent-based model (ILUTABM): Endogenizing human-environment interactions in the Western Missisquoi Watershed

Forest Transition Theory (FTT) suggests that reforestation may follow deforestation as a result of and interplay between changing social, economic and ecological conditions. We develop a simplistic but empirically data driven land use transition agent-based modeling platform, interactive land use transition agent-based model (ILUTABM), that is able to reproduce the observed land use patterns and link the forest transition to parcel-level heuristic-based land use decisions and ecosystem service (ES). The ILUTABM endogenously links landowners’ land use decisions with ecosystem services (ES) provided by the lands by treating both lands and landowners as interacting agents. The ILUTABM simulates both the land use changes resulting from farmers’ decision behaviors as well as the recursive effects of changing land uses on farmers’ decision behaviors. The ILUTABM is calibrated and validated at 30m×30m spatial resolution using National Land Cover Data (NLCD) 1992, 2001 and 2006 across the western Missisquoi watershed, which is located in the north-eastern US with an estimated area of 283 square kilometers and 312 farmers farming on 16% of the total Missisquoi watershed area. This study hypothesizes that farmers’ land use decisions are made primarily based on their summed expected utilities and that impacts of exogenous socio-economic factors, such as natural disasters, public policies and institutional/social reforms, on farmers’ expected utilities can significantly influence the land use transitions between agricultural and forested lands. Monte Carlo experiments under six various socio-economic conditions combined with different ES valuation schemes are used to assess the sensitivities of the ILUTABM. Goodness-of-fit measures confirm that the ILUTABM is able to reproduce 62% of the observed land use transitions. However, the spatial patterns of the observed land used transitions are more clustered than the simulated counterparts. We find that, when farmers value food provisioning Ecosystem Services (ES) more than other ES (e.g., soil and water regulation), deforestation is observed. However, when farmers value less food provisioning than other ES or they value food provisioning and other ES equally, the forest transition is observed. The ILUTABM advances the Forest Transition Theory (FTT) framework by endogenizing the interactions of socio-ecological feedbacks and socio-economic factors in a generalizable model that can be calibrated with empirical data.