Work

Hybrid Blockchain-Enabled Customary-Rights-Aware Dammar Price Index with Multi-Agent Reinforcement Learning for Credit-Risk-Adjusted Supply-Chain Finance in Pesisir Barat

Year: 2026
Type: article
Abstract: Background: The coastal province of Pesisir Barat supplies the world's largest volume of dammar resin, yet small-scale harvesters lack formal credit because their hak ulayat rights remain unregistered under the Indonesian Civil Code (KUHPerdata). Objectives: To design a permissioned blockchain registry that legally recognises customary land claims and to develop a multi-agent reinforcement learning (MARL) model that dynamically adjusts credit limits through a credit-risk-adjusted price index. Methods: A Hyperledger Fabric ledger records hak ulayat claims, links each claim to IoT telemetry (temperature, humidity, GPS) and enforces Pasal 1243 penalties via smart contracts. The MARL environment defines state Sₜ = (qₜ, lₜ, pₜ, wₜ, τₜ) (resin quality, legal status, market price, wave height, temperature deviation) and action Aₜ = (cₜ, rₜ) (credit allocation, routing). Reward Rₜ = -alpha·defaultₜ - beta·delayₜ - gamma·spoilageₜ + δ·profitₜ is calibrated (alpha = 0.6, beta = 0.3, gamma = 0.05, δ = 0.05) to reflect the 60 % platform/40 % harvester risk-sharing schedule. Boundary conditions (Hs <= 2.5 m, transport <= 4 h, τ <= +4 degC) are hard-coded in contract logic. Results: Field data (average yield 12.4 kg/harvester/month, informal interest 15.2 %/month, formal 12.8 %/year, unregistered land value IDR 1.2 trillion) feed the model; MARL reduces default risk by 18 % and logistics delay by 22 % versus a static credit rule, while maintaining spoilage below 2 % under the temperature threshold. Published in: Journal of Coastal Frontiers & Applied Development (JCFAD), Vol. 1 No. 1 (2026), pp. 1–12. Official publisher: PT CHAIN ASIA CONECT Research & Development Division.
Language: English
Open Access status: green
OpenAlex ID: W7235486143
FWCI:
Cites:
Cited by:
Related to: 10