Fleet Route Optimisation: How It Works & Why It Matters

1. What Is Fleet Route Optimisation?

Fleet route optimisation is the process of finding the best assignment of jobs to drivers and vehicles while minimising total travel time, distance, or cost. It's a variant of the Vehicle Routing Problem (VRP) — one of the most studied problems in operations research.

For transport operators, this means: given a set of jobs with pickup and dropoff locations, time windows, and passenger requirements, and a set of drivers with vehicles at various locations — find the assignment that gets everyone where they need to be with the least wasted driving.

Manually, this is done by experience and intuition. Algorithmically, it can be done in seconds — and the results are often significantly better than human-planned routes, especially as fleet size grows.

2. How the Algorithms Work

Modern route optimisation algorithms use a combination of techniques:

Construction Heuristics

The algorithm starts by building an initial solution — assigning each job to the nearest available driver, or inserting jobs into routes in a way that minimises the total distance added. This gives a reasonable starting point.

Local Search & Improvement

The initial solution is then improved by trying small changes: swapping two jobs between routes, moving a job from one driver to another, or reordering stops within a route. Each change is evaluated, and improvements are kept. This continues until no further improvements can be found.

Metaheuristics

Advanced algorithms use metaheuristic strategies — simulated annealing, genetic algorithms, or large neighbourhood search — to escape local optima. These techniques temporarily accept worse solutions to explore more of the solution space, often finding significantly better global solutions.

Distance & Travel Time Matrices

The quality of optimisation depends on accurate travel time estimates. The best systems use road-network routing (actual driving directions) rather than straight-line distance. This accounts for one-way streets, highway access, speed limits, and real road geometry.

3. Single-Job vs Fleet-Wide Optimisation

There are two levels of optimisation that serve different needs:

Smart Assign (Single Job)

When you need to assign one job quickly, smart assign evaluates all available driver-vehicle combinations and ranks them by suitability. Factors include proximity to the pickup, estimated travel time, vehicle capacity, driver hours remaining, and passenger accessibility requirements.

This is ideal for ad-hoc jobs that come in during the day — you need the best available option right now, without re-optimising the entire fleet.

Fleet Optimisation (All Jobs)

Fleet-wide optimisation takes all pending unassigned jobs and all available drivers, then finds the globally optimal assignment. This considers interactions between jobs — assigning job A to driver 1 might free up driver 2 for a job they're better suited for.

This is typically run at the start of the day (or shift) to plan the optimal assignments before dispatching. The results can be previewed and adjusted before applying.

4. Real-World Constraints

Pure distance minimisation isn't enough. Real transport operations have constraints that the algorithm must respect:

• Time windows — passengers have appointment times. Pickups and dropoffs must happen within specified windows.
• Vehicle capacity — passenger count, wheelchair spaces, and cargo weight limits must not be exceeded.
• Vehicle type matching — wheelchair-accessible vehicles for passengers who need ramps or hoists.
• Driver hours — maximum driving hours per day, break requirements, and shift boundaries.
• Driver skills — specific qualifications needed for certain transport types.
• GPS availability — drivers or vehicles without current GPS locations may need to be excluded or handled with fallback logic.

A good optimisation system handles all of these as hard constraints (must be satisfied) or soft constraints (preferred but not required), with clear reporting on why certain jobs couldn't be assigned.

5. Measurable Benefits

Transport operators who switch from manual planning to algorithmic optimisation typically see:

• 15-30% reduction in total driving distance — less wasted travel between jobs
• More jobs per driver per day — tighter routes mean more capacity without adding drivers
• Fewer late arrivals — algorithms respect time windows and buffer realistic travel times
• Reduced dispatcher time — minutes instead of hours spent on daily planning
• Lower fuel costs — direct result of shorter total distances
• Better driver utilisation — balanced workloads across the fleet

The benefits compound as fleet size grows. Manual planning becomes exponentially harder with more drivers and jobs; algorithmic optimisation handles the complexity without additional effort.

6. Evaluating Optimisation in Software

When comparing dispatch platforms, ask these questions about their optimisation:

• Does it use road-network routing or straight-line distance? Road-network is significantly more accurate.
• Can you preview assignments before applying? You should see the proposed plan and be able to adjust before committing.
• Does it explain unassigned jobs? If a job couldn't be assigned, you need to know why — time window conflict, capacity exceeded, no suitable vehicle, etc.
• Does it handle time windows and capacity? These are non-negotiable for passenger transport.
• Does it consider existing committed jobs? Drivers may already have assignments. The optimiser should work around them.
• How fast is it? Fleet-wide optimisation should complete in seconds, not minutes.
• Is there a fallback? If the routing API is unavailable, the system should fall back to distance-based scoring rather than failing.

7. Getting Started

You don't need to optimise everything on day one. A practical starting path:

1. Start with smart assign — use single-job optimisation for incoming jobs. See how ranked options compare to your manual choices.
2. Try fleet optimisation for one day — run the fleet optimiser for tomorrow's jobs. Preview the assignments and compare to how you'd have planned manually.
3. Build confidence — as you trust the results, make fleet optimisation part of your daily planning routine.
4. Add constraints gradually — configure driver hours, vehicle types, and time windows to make optimisation results more accurate.

See how it works in practice on Waybill's fleet optimisation software — one-click, fleet-wide route optimisation with full decision transparency.

Route optimisation isn't about replacing dispatcher judgment — it's about giving dispatchers a better starting point. The algorithm handles the maths; the dispatcher handles the exceptions.