Routing Optimization

The client faced challenges in optimizing delivery routes due to the complex nature of incorporating multiple restrictions, such as terrain conditions, vehicle availability, and size. The manual planning process led to inefficiencies, including suboptimal customer coverage and extended delivery times. Therefore, our main motivation was to automate the generation of near-optimal routes for their account managers and sales representatives.

To address these challenges, we developed a routing tool using a combination of algorithms and technologies: Custom Clarke & Wright Algorithm: This was employed to generate initial routes by maximizing efficiency in terms of travel time between Points of Sale. Google OR-Tools: Leveraged to refine the initial routes by incorporating various business-specific constraints, such as time limits and vehicle capacities. 2-Step TSP Algorithm: Applied as a final optimization step, this method further improved each route by minimizing travel distances and enhancing overall efficiency. We also utilized clustering techniques like DBSCAN to group customers within a 500-meter proximity into single locations, simplifying the routing process. The solution was designed to adapt dynamically, with the ability to test various visit durations and adjust routes accordingly.

Exceptional results were achieved through innovative solutions, turning the challenge at hand into measurable success.