Quay-Crane Scheduling Problem
The Quay-Crane Scheduling Problem (QCSP) determines how to sequence container handling tasks across multiple quay cranes assigned to a vessel, while respecting non-crossing and interference/safety-margin constraints. Objectives typically include minimizing vessel turnaround/makespan and smoothing crane workloads. QCSP is closely related to (and often integrated with) berth allocation and quay-crane assignment.
Optimization Focus
Objective
- Minimize vessel handling time (makespan/turnaround).
- Reduce crane interference and empty travel; smooth workloads.
- Support multi-objective goals (e.g., energy or emissions) when needed.
Decision Variables
- Assignment of bay tasks to cranes over time (who does what, when).
- Sequence of tasks per crane and task start/finish times.
- Feasible crane moves along the quay (respecting safety margins).
Constraints
- Non-crossing: adjacent cranes cannot pass each other on the same rail.
- Interference/safety margins: minimum bay distance between cranes.
- Precedence between bay tasks; travel/hoist times and crane speeds.
Data Sources
- Bay-plan/stowage (import/export moves, work content per bay).
- QC performance curves (hoist/slew/trolley), rail distances, safety margins.
- TOS logs/telemetry; integrated BAP/QCAP inputs if coupled.
Main Assumptions
- Vessel’s bay-level work content is known (static QCSP variant).
- Assigned cranes operate on the same rail with non-crossing limits.
- Interference is modeled via safety margins (minimum bay gaps).
- Travel/handling times deterministic or scenario-based if robust/stochastic.
- Precedence constraints reflect stowage rules and operational logic.
Modeling Approaches
QCSP is commonly formulated as a MILP with interference and non-crossing constraints; large instances often use heuristics and matheuristics.
- Exact/Matheuristics: B&B, logic-based Benders, set-partitioning, ALNS (for integrated cases).
- Metaheuristics: GA/TS/SA, GRASP, hybrid evolutionary; recent MDP/RL and energy-aware models.
- Integrated models: with berth allocation & crane assignment (BACAP) for stronger realism.
Available Datasets
🧭 Real-World Datasets
| Dataset / Case | Description |
|---|---|
| QCSPgen | Supplementary research data for tandem-QC operations; includes input sets used to validate an exact algorithm for real terminal scenarios (License: CC BY 4.0). |
| Tandem-QC | Open dataset coupling tandem-QC scheduling with truck/slot planning. Useful if the QCSP interacts with horizontal transport (License: CC BY 4.0). |
| Berth-and-QC | Weekly berth-and-QC planning instances. Good for experiments that link tactical berth plans with operational QC scheduling (License: CC BY-NC). |
⚙️ Synthetic Benchmark Instances
| Dataset | Description |
|---|---|
| QCSPgen | Official generator used in the literature to create the standard QCSP instance families. Page also links file format and benchmark results. |
| QCSPgen project page | Direct project page for the QCSP instance generator (configure bays, cranes, precedence, and safety margins). |
Real-world data are often proprietary (accessible via collaborations), while public benchmarks (Kiel sets, Kim & Park family, QCSPgen) enable reproducible comparisons.