School Bus Routing for Pupils with Special Needs Under Uncertainty: An Averaging Approach

School Bus Routing can be affected by significant delays, particularly when transporting students with special educational needs and disabilities (SEND). The associated uncertainties can disrupt pickup schedules, negatively impacting students' commuting experience and readiness to learn upon arrival at their schools. These observations were further corroborated by our findings from stakeholders, where delays and unpredictability were consistently identified as significant pain points within existing SBRP solutions. In response to these challenges, this study, conducted in collaboration with an industry partner in North-West England, explores averaging techniques as a noise-handling strategy to mitigate the effects of stochastic disruptions. Specifically, the research investigates implicit and explicit averaging techniques to assess their effectiveness in improving the robustness of bus routes under uncertainty. Our results and findings using real-world datasets suggest that simulating delays and applying averaging techniques throughout the search process helps improve the robustness of solutions. We also examine the computational cost of repeated evaluations and discuss the trade-off with increased robustness. The concluding findings demonstrate the potential of our approach to reduce the effects of random disruptions that can be commonplace when designing school bus routes for pupils with special needs.