Cloud-Based Stock Market Simulation and Trading Bots Competition Platform

This project enhances a cloud-based stock market simulation and trading bot competition platform originally developed in the previous year. The platform supports multi-language trading bots, simulates realistic market conditions, and evaluates bot performance in a fair and reproducible manner. This year's work focuses on improving automation, reproducibility, and developer usability, introducing C++/Java bot integration, containerised Jupyter Notebook environments, and improved orchestration of the execution lifecycle.

• ›C++ and Java trading bot integration via standardised Docker containers with multi-language kernel support.
• ›Behaviour-based dynamic analysis for enhanced bot validation and evaluation fairness.
• ›Average-score aggregation per bot with database persistence and live ranking tables.
• ›Containerised execution of the full competition platform for reproducibility and environment consistency.
• ›JupyterHub extended with Java and C++ kernels for unified development and testing workflows.

• ›End-to-end automation: all bot submission, execution, and scoring steps fully scripted.
• ›Reproducibility: Docker-based environment ensures identical execution across developer machines.
• ›Evaluation fairness: scoring robustness validated across multiple bot languages and strategies.
• ›Execution safety: containerisation prevents resource leakage and cross-bot interference.