Treffer: CPDGen: a Scalable Synthetic Dataset Generator for Container Port Operations.

Obtaining data for developing port studies and applications is awkward and costly. However, extending the basis of research in the field would be beneficial to improve its growth. Thus, we propose CPDGen (Container Port Dataset Generator), a synthetic data generator developed for container port simulation and experimentation, using a modular architecture which mimics the flow in a port. The implementation, in Python, is based on the SimPy framework and organizes the simulation process into five main phases: configuration, entities, actions, monitoring, and testing. Configuration is automatically performed processing simple JSON files, through which the end user can establish the infrastructural, temporal, and strategic parameters. The stochastic ship arrival engine implements the arrivals according to exponential, uniform, or weekly strategies, and the associated containers are dynamically generated by sampling characteristics from user-defined distributions. The simulation cycle manages the entire operational flow of arrival, mooring, unloading, and output updating. The monitoring module processes tables and graphs related to simulated ships and containers, while tests verify the consistency of the computed results. The modular division of the code allows for scalability, reusability, and extensibility. The tool has been designed to be integrated into experimental pipelines aimed at evaluating decision-making policies and generating synthetic benchmarks. [ABSTRACT FROM AUTHOR]