Route Optimization using ILS-RVND

Developed a complete optimization application that solves routing problems by implementing the ILS-RVND metaheuristic from scratch, allowing users to generate efficient travel routes through an interactive web interface.

Finding near-optimal routes for multiple destinations becomes computationally expensive as the number of locations increases, making brute-force approaches impractical for real-world applications.

Implemented the ILS-RVND optimization algorithm entirely from scratch in Python and deployed it as an interactive Streamlit application capable of solving routing problems with customizable datasets.

• Implemented the ILS-RVND metaheuristic from scratch instead of relying on optimization libraries to better control algorithm behavior.
• Separated optimization logic from visualization to improve maintainability.
• Designed modular neighborhood operators for easier experimentation and future extensions.
• Used Streamlit to rapidly transform an optimization algorithm into an interactive application.

• Implementing multiple neighborhood search operators correctly.
• Balancing solution quality with computational efficiency.
• Visualizing optimization results in an intuitive way for end users.

• Well-designed heuristics can produce near-optimal solutions within practical computation times.
• Algorithm implementation quality is as important as algorithm selection.
• Visualization greatly improves users' understanding of optimization results.

• Vehicle Routing Problem (VRP) support.
• Time Window constraints.
• Google Maps integration.
• Real-world traffic estimation.
• Benchmark against additional metaheuristics.