Evolutionary AI Platformer Controller
Overview
The AI Game Controller project is a C++/SFML platformer that integrates an artificial intelligence controller capable of learning gameplay behaviour through evolutionary techniques. The project combines 2D game development with AI experimentation, using a playable platformer environment to test how automated agents can adapt to movement, decision-making, and level navigation challenges.
The project was developed to explore how evolutionary approaches can be applied to game AI, with a focus on agent behaviour, fitness evaluation, and iterative improvement over time. Rather than presenting only a platformer game, the project demonstrates how gameplay systems can be used as a test environment for AI learning, making it both a game development project and an AI-focused research piece.
My Role
I developed the AI-driven platformer controller in C++ using SFML, including the gameplay systems, autonomous agent behaviour, and evolutionary learning workflow. My work focused on training agents through fitness scoring, mutation, and adaptive gameplay behaviour, while using the platformer environment as a testbed for AI experimentation. I also structured the project around object-oriented C++ design, simulation-based testing, gameplay analytics, and debugging tools to support future expansion and performance optimisation.
Key Engineering Features
- Artificial intelligence for autonomous gameplay
- Genetic algorithm and neuro-evolution principles
- Custom-built 2D platformer mechanics
- Simulation and automated training systems
- Object-oriented C++ architecture
- Gameplay analytics and debugging tools
Technical Highlights
The project combines a custom C++/SFML platformer with an AI controller designed for autonomous gameplay. Evolutionary learning techniques are used to train agents through fitness scoring, mutation, and repeated simulation, allowing behaviour to improve across generations. The system is structured around object-oriented C++ architecture, with separate gameplay, AI, training, and debugging components to support experimentation, performance optimisation, and future expansion.
Development Goals
The project was developed to strengthen my understanding of artificial intelligence in games, evolutionary learning techniques, and simulation-based training systems. A key goal was to explore how agents can learn gameplay behaviour through fitness scoring, mutation, and repeated testing within a custom platformer environment. The project also focused on combining game development with AI experimentation, creating a modular foundation for future work in autonomous gameplay, adaptive behaviour systems, and intelligent game agents.
Tools and Technologies
C++, SFML, Visual Studio, artificial intelligence, genetic algorithms, neuro-evolution principles, autonomous gameplay systems, simulation-based testing, fitness scoring, mutation systems, object-oriented programming, game loop architecture, state management, gameplay analytics, debugging tools, and performance optimisation.