The encyclopedia · Engineering & Operations · Operational decision · 2020–2025
Delhivery's midmile routing algorithm cut logistics costs 7.3%
Delhivery automated midmile planning across 3,000+ locations with a hybrid evolutionary search, saving nearly 7.3% of route cost and picking new hub sites too.
Delhivery
the move
Delhivery, one of India's leading logistics companies, runs a midmile distribution network spanning more than 3,000 locations, with fleet mix, site dependencies and time windows.
Researchers from Delhivery and IIT Kharagpur formulated the problem as a novel mixed-integer linear program and built a hybrid evolutionary search algorithm plus an insertion heuristic that cuts computation time.
The algorithm achieved significant cost savings of nearly 7.3% and converges rapidly, automating the entire planning and operations process.
Its capabilities were extended to choose optimal locations for new hubs within the network, turning daily routing into network design.
why it works
- One algorithm replaced manual planning across thousands of locations.
- A 7.3% cost cut is large in a thin-margin logistics business.
- The same model answers both routing and hub-location questions.
- Rapid convergence made daily re-planning practical.
what transfers
When a network crosses thousands of nodes, replace manual route planning with a search algorithm that can also answer design questions like hub location.
what came after
Published in INFORMS Journal on Applied Analytics (2025), the work demonstrates OR techniques improving a large logistics network; Delhivery integrated the algorithm's planning into its operations.
references
- Optimizing Delhivery's Midmile Logistics Network Using a Hybrid Evolutionary Search Algorithm
- Optimizing Delhivery's Midmile Logistics Network Using a Hybrid Evolutionary Search Algorithm
spotted an error? The archive wants to know.