Application of Fuzzy Cognitive Map for Policy Simulation - in the Agricultural Sector - Cungki Kusdarjito

Oct 26, 2026

Agricultural socio-economic research such as RRA and PRA must deliver timely results while managing subjective, ill-defined variables. Rural Rapid Appraisal gathers rural information systematically, while Participatory Rural Appraisal emphasizes community engagement. This research uses Fuzzy Cognitive Mapping to manage fuzzy variables via positive and negative concept links, integrating stakeholder perspectives. It presents a preliminary LLM-driven FCM model for oil-palm farming, focusing on middlemen roles, enhanced by Hebbian learning and evaluated through multiple scenarios. Results show effective policy impact simulation by tuning learning rate and squashing slope, with future work expanding stakeholder coverage.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages20

Reading Time50 min

AccessFree to download

Abstract

1. Introduction

1.1. Simulation Models in Indonesian Agriculture

1.2. Simulation Incorporating Qualitative and Subjective Variables

Application of Fuzzy Cognitive Map for Policy Simulation in the Agricultural Sector Cungki Kusdarjito Department of Agribusiness, Faculty of Agriculture, Janabadra University, Yogyakarta, Indonesia *Corresponding author: [ckusdarjito@janabadra.ac.id](ckusdarjito@janabadra.ac.id) ABSTRACT Agricultural socio-economic research, such as Rural Rapid Appraisal (RRA) or Participatory Rural Appraisal (PRA), often requires quick results while dealing with subjective data. RRA systematically collects rural information, while PRA emphasizes community participation and engagement. In this context, Fuzzy Cognitive Mapping (FCM) assists in managing subjective, ill-defined, and fuzzy variables. FCM creates a network of concepts within systems through positive and negative connections, integrating diverse stakeholder perspectives to support decision-making. This research presents a preliminary model using Large Language Models (LLM) to develop FCMfor oil-palm farming, specifically examining the role of middlemen. The FCMis built on subjective variables gathered through open interviews with two palm-oil farmers in Central Kalimantan. Hebbian learning enhances the accuracy of relationships between variables. Results show that FCM effectively simulates policy impacts and their outcomes across different scenarios by adjusting the learning rate and the slope of squashing function. Future research should incorporate more stakeholders to enhance LLM's capability in mapping diverse perspectives. Keywords: FCM, Hebbian Learning, LLM, PRA, RRA. 1. INTRODUCTION Agricultural programs use Rural Rapid Appraisal (RRA) and Participatory Rural Appraisal (PRA) for problem mapping based on qualitative data. RRA systematically gathers rural information, while PRA engages communities in identifying problems and solutions (Preece, 2006). Both methods help understand farming communities and design effective interventions (Chambers, 1994) . RRA relies on expert teams, whereas PRA emphasizes community involvement. Successful programs integrate participatory approaches to represent community aspirations (Brata et al., 2022) . However, rapid PRA adoption without understanding its principles can lead to shallow applications that fail to achieve results (Chambers, 1994) . The challenge of effectively integrating participatory approaches like RRA/PRA with conventional quantitative research methods restricts comprehensive data analysis (Campbell, 2002; Campbell, 2001) . Bridging this methodological divide would enhance both data richness and validity, helping researchers overcome the inherent limitations of either approach used in isolation (Koutra, 2010; Catley, 1997) . Studies examining PRA implementation in Indonesian villages reveal both advantagesand disadvantages. Research shows that PRA facilitates stakeholder collaboration more effectively than traditional top-down approaches. However, cultural and institutional contexts significantly influence PRA's effectiveness (Barat & Yusof, 2012) . Further assessment of PRA's impact on poverty alleviation indicates that it becomes less effective when implemented without adequate supportive local structures (Putri, 2007) . 1.1. Simulation Models in Indonesian Agriculture Various simulation models have been employed in Indonesian agriculture. The Oryza model simulates rice production in West Java by integrating weather, soil, and agricultural data, showing strong correlation between estimated and observed solar radiation for biomass production simulation (Sulistyowati, 2015) . In Yogyakarta, Soft System Methodology evaluates organic farming business strategies through focus groups and interviews, identifying key strategies: operational evaluation, managerial improvement, and sustainability indicators (Prasetyaningtyas et al., 2019) . System dynamics models successfully increased organic corn production in East Java by simulating land expansion scenarios, where smart agriculture practices enhanced product

FAQs

Why do RRA and PRA require tools for subjective and fuzzy data?

They involve quick appraisal and community perspectives that often produce subjective, ill-defined variables. FCM helps represent these fuzzy variables within a structured model.

How does the proposed model use LLM with Fuzzy Cognitive Mapping?

The study presents a preliminary model using Large Language Models to develop an FCM for oil-palm farming. The FCM is built on subjective variables collected through open interviews.

What method improves the accuracy of relationships between variables in the FCM?

Hebbian learning is used to enhance the accuracy of the relationships between variables, supporting more reliable simulation outcomes.

Oct 26, 2026

Upload by วิน