<![CDATA[Operational bottlenecks in LPG refinery distribution can arise from single-manifold constraints, creating a critical trade-off between throughput maximization and product quality assurance. Existing scheduling approaches often address vessel evacuation and blending separately, leaving a gap in integrated decision models that simultaneously consider infrastructure constraints, production fluctuations, and quality specifications. This study aims to develops an integrated simheuristic decision-support framework combining a five-tier Lexicographic Heuristic Priority Rule, Fuzzy Logic membership functions, and a Zero-Sealing mechanism to prevent off-specification products from entering distribution. The framework enables refinery managers to determine achievable throughput through a single manifold while maintaining product quality. A discrete-event simulation with hourly time steps was developed over a 31-day operational horizon using historical vessel arrival, distribution, and production data from a refinery case. Model performance was evaluated through Monte Carlo simulation under four uncertainty dimensions: maritime arrival variability, batch-size variation, pipeline degradation, and upstream feed-quality fluctuations. Sensitivity analysis was conducted to identify critical operational thresholds. Across the tested scenarios, the framework recorded no overfill incidents or quality specification violations and achieved a mean domestic gas supply of approximately 1,355 MT/day, evacuating nearly all available upstream production. Pipeline rate degradation was identified as the most sensitive parameter, affecting both throughput evacuation and quality control when capacity declined below a critical level. The findings demonstrate that evacuation capacity and quality management are structurally interdependent in single-manifold systems. The proposed framework contributes provides a quantitative decision-support tool for reducing off-specification risks and informing refinery infrastructure investment decisions.]]>
