← All experiments
Notebook / routing · algorithm · simulation

PARCEL / School bus routing

An early-stage research prototype for school bus routing. Explore clustering, assignment, and route construction on real street networks, with future research focused on local search, subcluster refinement, genetic algorithms, and simulated annealing.

Open interactive demo ↗

PARCEL is an experimental school bus routing and optimization project, currently in its alpha research stage.

This interactive lab accompanies the original PARCEL article, allowing visitors to explore how the algorithm clusters stops, assigns them to buses, and constructs routes on a real street network. The simulation exposes intermediate decisions to make the routing process more transparent and easier to evaluate.

Route construction and assignment decisions come from the instrumented algorithm, while road distances and travel estimates are obtained through the public OSRM demo endpoint.

Current limitations

The current implementation is an initial approach, not a globally optimal routing solution. While it demonstrates the core methodology, further experimentation, benchmarking, and validation are required to evaluate solution quality, scalability, and computational efficiency.

In particular, the initial clustering and assignment decisions may introduce constraints that limit the quality of the final routes.

Research directions

Future investigations may explore:

  • Individual-route optimization: Improving stop sequences and route geometry through local search and neighborhood operators.
  • Subcluster refinement and rescue: Reassigning individual stops or groups of stops between clusters to address inefficient initial assignments.
  • Genetic algorithms (GA): Exploring population-based optimization to improve assignments and routing configurations.
  • Simulated annealing: Investigating probabilistic search to escape local optima and discover better route configurations.
  • Hybrid optimization: Combining constructive heuristics with local and metaheuristic search for potentially better solutions.

These are proposed research directions, not implemented or validated improvements. Their effectiveness will need to be measured against the current baseline using objective metrics such as total travel distance, fleet utilization, constraint satisfaction, and execution time.

Project status

Alpha · Experimental · Active research

PARCEL is intended as an evolving research environment for investigating school bus routing strategies, visualizing algorithmic decisions, and comparing optimization techniques.

The current results should be interpreted as experimental outputs rather than production-ready or mathematically optimal solutions.

Read the full article ↗