Intelligent Port Operations at Tanger Med

Harch Intelligence deployed a comprehensive AI-driven port orchestration platform — HarchOS Port — built on four integrated modules that address berth management, yard optimization, customs intelligence, and holistic port coordination.

The first module, dynamic berth allocation, replaces the static 72-hour planning cycle with a continuously updated real-time schedule. HarchOS ingests Automatic Identification System (AIS) data from every vessel within 1,000 nautical miles of Tanger Med, combining it with weather forecasts, tidal predictions, and historical arrival pattern data to predict each vessel's actual arrival time with 94% accuracy at the 24-hour horizon and 98% accuracy at the 6-hour horizon. When the system detects an impending deviation — a vessel delayed by Atlantic weather, a port congestion event at an upstream stop — it proactively reallocates berths, reassigns crane resources, and adjusts pilotage schedules to minimize the impact. The optimization algorithm evaluates over 12,000 possible berth configurations every 15 minutes and selects the one that minimizes total vessel waiting time while respecting draft restrictions, crane compatibility, and terminal operator preferences. In its first six months of operation, dynamic berth allocation reduced average vessel waiting time from 18.4 hours to 7.2 hours.

The second module, intelligent yard management, uses container connection graph analysis to optimize stacking decisions. When a container enters the yard, HarchOS immediately identifies its onward connection — which vessel it is destined for, which discharge port that vessel serves, and what other containers share the same connection. The system then directs the container to a stacking position that minimizes future housekeeping moves, clustering containers with the same destination vessel in adjacent slots. The optimization runs continuously as new containers arrive and vessel schedules update, dynamically reshuffling the yard layout to minimize the total number of internal moves required for upcoming loading operations. Housekeeping moves have been reduced by 67%, freeing straddle carrier capacity for productive operations and dramatically reducing yard congestion.

The third module, customs risk intelligence, replaces the outdated risk-scoring system with a machine learning model trained on 4.2 million historical declarations, 180,000 inspection outcomes, and real-time intelligence feeds including vessel routing anomalies, declarant behavioral patterns, and commodity risk profiles. The model classifies each incoming declaration into risk tiers with 91% accuracy, enabling customs to target physical inspections at the 18% of shipments that account for 94% of contraband and compliance violations. The reduction in unnecessary inspections — from 38% to 18% of total declarations — has freed 180,000 TEU-equivalents of yard capacity annually without any reduction in customs revenue or enforcement effectiveness.

The fourth module is the Port Control Tower — a unified dashboard that integrates all three operational modules with real-time vessel tracking, crane telemetry, gate operations, and weather monitoring into a single operational picture. The control tower provides AI-generated recommendations for every major operational decision, which human operators validate before execution. This human-in-the-loop architecture ensures that the port retains operational flexibility for exceptional situations while benefiting from AI-driven optimization for routine decisions.