Our second product
Mizan Routes CI · Route Optimization
A dispatcher drops in the day's orders and fleet; the app refuses to compute if the data is doubtful, then produces optimal routes, an interactive map, and a quantified comparison against the naive route.
The workflow
From order to optimized route
01
Ingestion
Import of the day's orders and available fleet
02
Quality (lock)
Insufficient capacity, aberrant coordinates, negative quantities: blocked before any computation
03
Optimization
OR-Tools (CVRP) + real distances via OSRM, never as the crow flies
04
Delivery
Interactive map, km/FCFA comparison vs naive route, export, history
Concrete proof
The same doctrine as DataGov CI
"You can't decide, or model, on corrupted data." In Mizan Routes CI, concretely: the quality module blocks optimization on doubtful data, never a silently wrong result, and the OSRM fallback blocks by default rather than silently substituting a straight-line distance.
OR-Tools
CVRP solver + time windows (VRPTW)
OSRM
self-hosted routing engine, real distances
Real-time traffic
optional congestion/incidents (TomTom), auto-fallback to OSRM
km · FCFA
quantified comparison, naive vs optimized route
What's next
Foundations built with V2 in mind
The data schema already provisions the following extensions, without activating them in V1:
- Multi-depot support (today: a single active start depot, selectable on a map)
- Automatic geocoding of order addresses (today: GPS coordinates required as input for orders; address search already exists for configuring a depot)
No firm delivery date, except under specific contractual agreement.
Another product from the studio
Your routes are reliable. Your data can be too.
Are your routes costing more than they should?
Let's talk about your fleet and delivery constraints in Abidjan.
Request a demo →