Efficient public transport scheduling is at the heart of every public transport company. Yet, this task is a complex puzzle—balancing demand, traffic situations, public transport crowding and other factors like the weather. Traditional methods often fall short in tackling these challenges with robust punctuality, and lack insight. That’s where HIPE (Hybrid Intelligence Planning Engine) comes in, an explainable AI solution co-developed by our team and EBS, a leading public transport company.
Over a 4-month collaboration, Gradient built HIPE to empower planners with an intelligent yet flexible tool that enhances their decision-making processes while maintaining room for human insights. The result? A machine learning-driven scheduling solution that outperforms traditional approaches while seamlessly incorporating planners’ expertise.
What is HIPE?
HIPE stands for Hybrid Intelligence Planning Engine, and it’s more than just a scheduling tool. At its core, it combines advanced machine learning models with human expertise to create optimal schedules for buses, trams and metros. The system doesn’t just generate schedules—it provides deep insights into how various factors like time, day, season, public transport crowding, traffic, and weather influence travel times.
This hybrid approach ensures that planners remain in control, using the machine learning model as a guide to make informed decisions. HIPE’s iterative design allows for quick adjustments based on planners’ insights, ensuring the schedules are not only efficient but also practical and aligned with reality.
Key Features of HIPE
- Explainable Machine Learning: HIPE leverages explainable machine learning models trained on historical data, providing accurate predictions for travel times based on complex variables like traffic and weather conditions.
- Hybrid Intelligence: HIPE strikes the perfect balance between machine and human intelligence. Planners can adjust, override, or validate the model’s suggestions, creating a true partnership between AI and human expertise.
- Insight into Effects: Beyond predictions, HIPE offers transparency. It breaks down the impact of factors like passenger crowding or seasonal changes, helping planners understand the “why” behind the suggested schedules.
- User-Centric Design: Developed in close collaboration with EBS planners, HIPE prioritizes user experience. Input sessions and user experiments shaped its intuitive interface, ensuring it integrates seamlessly into existing workflows.
- Fast Iterations: With its lean development framework, HIPE is a custom software that can be easily adapted to any public transport organisation’s needs.
Our Development Journey
The development of HIPE was a deeply collaborative effort. From day one, we worked closely with experienced planners at EBS, conducting input sessions to understand their challenges and priorities. This partnership ensured that HIPE wasn’t just a technical solution but a practical tool tailored to planner’s needs.
Through iterative design and continuous feedback, we rapidly refined HIPE, conducting user experiments to test its usability and effectiveness. Within just four months, we developed a tool that outperformed traditional methods, proving its value in optimizing public transport schedules.
What We Learned
- Collaboration is Key: The success of HIPE is a testament to the power of close collaboration. Engaging planners throughout the process ensured the tool met their needs and was embraced by its users.
- Transparency Matters: Planners were more likely to trust and adopt HIPE because it provided insights into the “why” behind its suggestions. This transparency is critical for building trust in AI-driven tools. The explainability of AI is vital.
- Hybrid Intelligence is the Future: By combining machine learning with human expertise, HIPE demonstrated that AI doesn’t replace people—it enhances their capabilities and improves schedule punctuality beyond any model or planner could do independently.
- Iterative Design Yields Results: Fast iterations and user experiments enabled us to create a robust, user-friendly tool in a short time frame.
Looking Ahead
HIPE has already demonstrated its potential to revolutionize public transport planning. As we continue to refine and implement the tool, we envision expanding its capabilities—incorporating other data sources, integrating new variables, and scaling to other use cases.
This project is just the beginning of what’s possible when AI and human intelligence come together. With HIPE, we’re not just optimizing schedules; we’re trying to shape the future of public transport.
Want to learn more about HIPE or collaborate on your next project?
Reach out to us to discover how HIPE can transform your scheduling process!