Matching Vehicles to Trips: The Daily Assignment Problem

Every passenger transport operator runs the same daily routine: a list of trips that need to run, a list of vehicles and drivers available, and someone, usually the most experienced planner, deciding by hand which vehicle and driver goes on which trip. When that planner is on leave, the whole operation slows down with them.

This is the assignment problem, and it sits one rung above getting your trips out of manual planning (covered in the pillar post, Fleet Scheduling: The Limits of Manual Planning). It is also where most of the daily planning hours go. Here is what software does with it, and what you need before it works.

In short: Matching vehicles and drivers to trips is a constraint puzzle that gets exponentially harder as you add vehicles, trips, and rules. A human picks a "good enough" answer and moves on. Scheduling software checks thousands of valid combinations in seconds and proposes a better one, while respecting capacity, licensing, shifts, and time windows. It is worth doing once your assignment is too big to hold in one head, but only if your trip, vehicle, and driver data is clean enough to feed it.

What does "matching vehicles to trips" actually mean?

It is deciding, for every trip you run, which specific vehicle and which specific driver is assigned to it, so that every trip is covered and no resource is double-booked. With five trips it is easy. It gets hard quickly.

With a handful of vehicles and trips, a planner can see the whole board. Add more vehicles, more trips, split shifts, and rules about which driver can drive which vehicle, and the number of possible valid assignments explodes past anything a person can check. The planner is no longer finding the best plan, just the first plan that does not break.

Why does manual planning break at this?

Most operators hold the plan in a spreadsheet, and a spreadsheet stores the plan but does not understand the rules.

• It will happily let you assign one driver to two trips that overlap.
• It will not warn you that a 40-seat coach is going out on a 12-passenger run while the minibus sits idle.
• It cannot tell you that today's plan used 30 percent more overtime than yesterday's for the same trips.
• Every change is a manual re-shuffle, and every re-shuffle is a chance to introduce an error.

A spreadsheet is fine for recording the plan. The moment your constraints matter, it cannot keep up.

How does software do it better?

It treats the assignment as a problem to solve, not a grid to fill in. Give it the trips, the vehicles, the drivers, and the rules, and it searches through valid combinations to find a cheaper one.

What a planner does by hand	What scheduling software does
Holds the rules in memory	Stores the rules explicitly, applies them every time
Finds the first plan that works	Searches many valid plans, proposes a low-cost one
Re-shuffles manually on every change	Re-solves in seconds when something changes
Hard to replace if they leave	Runs the same way regardless of who is on shift

The planner stays in the loop, reviewing a proposed plan and handling the exceptions instead of building every plan from a blank sheet. The combinatorial search is the part software does better than people.

What constraints does it have to respect?

This is where a real operation differs from a textbook. Useful scheduling has to handle your actual rules:

• Vehicle capacity and type: the right size vehicle for the trip, and only vehicles fit for the job.
• Driver licensing and competency: who is allowed to drive what.
• Shifts and rest rules: legal driving hours, breaks, and shift lengths.
• Time windows: pickups and arrivals that have to land inside a contracted window.
• Depots and positioning: where vehicles and drivers start and end.
• Fairness: spreading work and overtime so the roster does not burn people out.

A tool that ignores your real constraints produces plans you cannot use. This is exactly why operators with unusual rules often outgrow off-the-shelf products and end up wanting something built around how they actually run.

What do you need before this works?

The honest blocker is rarely the algorithm. It is the data. Before software can match vehicles to trips, you need:

• A clean list of your vehicles, with capacity and type.
• A clean list of drivers, with what each is licensed and available for.
• Your trips in a structured form, with times and requirements.
• Your real constraints written down, not just remembered.

If that data lives in five WhatsApp chats and three spreadsheets, the first job is consolidating it, not optimizing it. Getting this wrong is the most common reason scheduling projects stall, which is why it deserves its own attention before you buy or build anything.

What to do next

Measure four things:

• How many vehicles, drivers, and daily trips you run.
• How long the assignment takes each day.
• How often it has to be redone.
• How exposed you are if your best planner leaves.

The next rung up is the roster: see Driver Rostering and Shift Planning.