School bus routing problem in the stochastic and time-dependent transportation network

Area of Science:

• Operations Research
• Transportation Science
• Logistics Management

Background:

• Real-world road networks face travel time uncertainties due to accidents, weather, and congestion.
• These uncertainties impact individual quality of life and overall transportation system reliability.
• The school bus routing problem is particularly challenging in stochastic and time-dependent environments.

Purpose of the Study:

• To address the school bus routing problem in stochastic and time-dependent road networks.
• To develop a model that generates cost-reliable routes considering travel time uncertainties.
• To ensure timely student pick-up and drop-off despite dynamic network conditions.

Main Methods:

• Formulated the problem as a variant of the extended Vehicle Routing Problem.
• Developed a Robust Optimal Schedule Times (ROST) model with hard time windows.
• Utilized Robust Optimization to minimize the worst-case total cost (schedule delays and travel time disutility).

Main Results:

• The proposed ROST model effectively generates cost-reliable school bus routes.
• The model can be simplified for deterministic networks, reducing computational complexity.
• A case study confirmed the model's validity and the importance of considering network stochasticity.

Conclusions:

• The ROST model provides a viable solution for school bus routing in dynamic and uncertain environments.
• Considering stochastic and time-dependent network characteristics is crucial for reliable transportation planning.
• The approach offers practical applicability for real-world school bus systems.