Decision Support Tool for Rollingstock Maintenance
Image generated by DALL-E
I led the development of a decision support tool to improve the scheduling of rolling-stock maintenance. This involved collecting and analysing historical maintenance data to build predictive maintenance models using Python libraries such as pandas, sklearn, and matplotlib. I used the Harris’ Hawk Optimisation technique to optimise feature selection and maximise accuracy, which significantly improved the predictive models.
One of the notable achievements of this project was accurately predicting the duration of brake maintenance, which helped streamline maintenance scheduling. I also regularly held meetings with stakeholders to gather information on business rules, constraints, and requirements, which helped me create mathematical models for multi-objective optimisation of maintenance scheduling tasks.
To expand the scope of the project, I created a simulation-based optimisation model that included What-if scenario analysis to evaluate the maintenance schedule thoroughly. My contributions to this comprehensive project have significantly improved the efficiency and reliability of rolling-stock maintenance operations.