Berth Allocation Problem
The Berth Allocation Problem (BAP) addresses the allocation of berth space for vessels in container terminals with spatial and temporal decisions. It is central to terminal management, directly influencing turnaround time, congestion, and service quality. The challenge lies in efficiently allocating berths amid dynamic arrivals, vessel priorities, and physical limitations, ensuring both operational efficiency and fairness in port logistics.
Optimization Focus
Objective
- Minimize total vessel waiting and berth idle time.
- Balance utilization and fairness across terminals and services.
- Enable trade-offs between punctuality and operational cost.
Decision Variables
- Vessel–berth assignment (which berth each vessel uses).
- Service start/finish times (temporal allocation).
- Sequencing of vessels along the quay wall.
Constraints
- No overlap at the same berth segment (mutual exclusivity).
- Compatibility with length, draft/depth, and tidal windows.
- Earliest/latest start, precedence, and time-window limits.
Data Sources
- AIS trajectories and historical berth occupancy.
- Portbase/HaMIS and terminal operation schedules.
- Turnaround records and tide/depth constraints.
Main Assumptions
- Arrivals known within the planning horizon (static case).
- Each vessel occupies one berth for its service duration.
- Handling and travel times are deterministic and pre-estimated.
- No overlap at the same berth; exclusivity enforced.
- Vessel–berth compatibility by length and depth/draft.
- Tidal restrictions validated before assignment.
- Pilotage/towage availability modeled externally.
Modeling Approaches
The BAP is often formulated as a Mixed Integer Linear Programming (MILP) problem that captures both spatial and temporal decisions. For realistic scales, researchers rely on approximate or hybrid optimization methods:
- Metaheuristics (Genetic Algorithms, Tabu Search, Simulated Annealing).
- Decomposition and constructive heuristics for large instances.
- Rolling-horizon models for dynamic/real-time scheduling.
- Stochastic or robust formulations for uncertain arrivals/handling times.
Available Datasets
🧭 Real-World Datasets
| Dataset | Description |
|---|---|
| NOAA/BOEM Marine Cadastre – AIS Vessel Tracks (2019) | Shows the location and characteristics of commercial and recreational boats as a sequence of positions transmitted by an Automatic Identification System (AIS). |
| NOAA/BOEM Marine Cadastre – AIS Vessel Transit Counts (2019) | Navigation safety device that transmits and monitors the location and characteristics of many vessels in U.S. and international waters in real-time. |
| Marine Cadastre – Vessel Traffic | The data available in the table contains broadcast point data for U.S. coastal waters for calendar years 2009 through 2025. |
| PortWatch Tracker | Open platform designed to monitor and simulate disruptions to maritime trade flows. |
| European Data | Open data service (API) with real-time, free information about open bridges. |
| NOAA GeoPlatform | The data available in the table contains broadcast point data for U.S. coastal waters for calendar years 2009 through 2025. |
⚙️ Synthetic Benchmark Instances
| Dataset | Description |
|---|---|
| MathProg OR-Lib BAP Dataset | Standard benchmark for static/dynamic BAPs with varied sizes and configurations for algorithm testing. |
| DTU BACAP Dataset | Joint berth & crane assignment instances (quay geometry, handling times); suited for integrated optimization. |
| Kiel University SCM | Benchmark Instances for Berth Allocation and Crane Assignment. |
| Channel-restricted BAP datasets | Datatasets related to the computational study presented in the paper “The Berth Allocation Problem with Channel Restrictions”. |
Real-world data require AIS/PCS pre-processing, whereas benchmark sets enable controlled comparisons of optimization methods.