MOVE
Last-Mile Delivery Routing
Optimize daily delivery routes for last-mile operations, dynamically adjusting vehicle paths to minimize costs, meet time windows, and handle real-time constraints like traffic and new orders.
Understanding the Problem
Last-mile delivery represents 40-53% of total shipping costs despite being the shortest segment of the supply chain. With over 60% of consumers expecting same-day or next-day delivery, companies must optimize routes dynamically. The challenge extends beyond distance minimization to balancing real-time traffic, delivery time windows, vehicle capacity, and sustainability goals. Poor route planning costs $17.20 per failed delivery and results in excessive fuel consumption and customer dissatisfaction.
THE CHALLENGE
What Makes it Hard
Delivery companies must determine the best sequence of stops for each vehicle while respecting time windows, capacity limits, and driver hours. Routes must adapt continuously as traffic conditions change, new orders arrive, and deliveries fail.
WHO FACES IT
The Vehicle Routing Problem with Time Windows (VRPTW) is NP-hard: a fleet of 50 vehicles serving 500 locations has astronomical possible combinations
Routes must adapt continuously to changing traffic, last-minute orders, and delivery failures, making static pre-planned routes obsolete within hours
Businesses must simultaneously minimize costs, maximize service quality, and reduce emissions: objectives that often conflict
BUSINESS IMPACT
Route optimization reduces delivery costs by 20% and distance by 28%. UPS ORION saves 100 million miles and 10 million gallons of fuel annually.
These gains compound daily across every route in your fleet.
DHL Cost Reduction
20%
Greenplan Results[1]
DHL's Greenplan dynamic routing algorithm achieved a 20% reduction in delivery costs through optimized route planning.
UPS ORION
100M miles
Saved Annually[2]
UPS's ORION system saves 100 million delivery miles annually, saving 10 million gallons of fuel per year.
Route Optimization
28%
Distance Reduction[3]
A 3PL cut average delivery distance by 28% and saved $450K annually in fuel through route optimization.
How We Solve It
Classic Vehicle Routing Problem with Time Windows (VRPTW) combining discrete routing decisions with continuous scheduling. Binary variables select edges while continuous variables track arrival times. Objectives minimize distance, vehicles, and emissions while respecting capacity, time windows, and route duration constraints.
Hybrid Compute
What We Bring
Exact MIP solutions for smaller fleets (under 100 stops) within minutes
Production-scale heuristics for 1000+ stops using adaptive large neighborhood search
Real-time local search for mid-route adjustments as conditions change
Multi-objective balancing of cost, service quality, and sustainability
FUTURE POSSIBILITIES
The
Quantum Horizon
VRP is a canonical optimization problem that quantum researchers use to benchmark new algorithms. However, current quantum hardware handles only toy problems (6-13 locations) while classical solvers optimize 1000+ customers in real-time.
Exploratory Work
Academic research on small-scale VRP instances to understand quantum algorithmic approaches. No practical path to quantum advantage within a decade for production-scale delivery optimization.
Current Research Directions
QAOA implementations achieving 13 locations using 156 qubits
D-Wave quantum annealing limited to fewer than 7 customers due to embedding overhead
Quantum-inspired metaheuristics showing no advantage over mature classical methods
Interested in quantum research?
Explore proof-of-concept implementations with our team.
Ready to solve this problem?
Talk to our experts about how Strangeworks can help with last-mile delivery routing.
SOURCES