A Modelling of Multi-Objective Sustainability Palm Kernel Supply Chain Based on Hybrid NSGA-II and Reinforcement Learning

Abstract

Palm kernel shell supply chain is a partial network of integrated palm oil supply chain network having potential business to be developed in order to reduce waste in the network. Sustainability and uncertainty are important challenges in the supply chain network design to make the business network sustain. This research work aims to design a Strategy of Sustainable Multi-Echelon Supply Chain based on Hybrid NSGA-II and Reinforcement Learning Under Uncertainty in Palm Kernel Shell Material Flow. We design a mathematical model with triple objectives in palm kernel shell supply chain network to meet sustainability criteria with economy, social, and environmental impact from the supply chain. To solve tripe objective function Hybrid NSGA II is designed to reach Pareto optimal solution combined with Reinforcement Learning to reach beneficial solution during optimization processes. The hybrid methodology found 28 alternatives strategy and outperform other method in multi-objective model solution processes. The calculation shows environmental reduction (in kg carbon reduction) to 252 kg and economic impact in complex palm kernel shell supply chain network rather than previous research. The solution also suggests the network configuration fo palm kernel shells material flow.