The Science behind Milk Collection Route Optimisation

Route optimisation problems are present in most industries. Think of the growing amount of parcel deliveries, couriers and logistics services providers. Food, medicine, waste management and even flowers all require detailed planning and organising to ensure collections and deliveries are completed on time.

Agriculture – especially the dairy industry – faces a number of transportation challenges on a daily basis and having optimal route plans can provide a key to competitiveness.  For smaller operations, it’s relatively easy to plan with paper and pencil. But even with a few dozen sites to visit, it’s difficult to find a near-optimal solution among an exponentially growing number of options.

The simplest route optimisation problem is the traveling salesman problem (TSP), where a vehicle visits all locations from a given starting point with the lowest possible travel cost.  The vehicle routing problem (VRP) is just like TSP, but with multiple vehicles. There are many variants such as time windows, multiple depots, heterogeneous fleet of vehicles, capacity limits, etc.  What all of these variants have in common is that they are NP-hard, which essentially means that only the small problems can be solved to optimality in a reasonable amount of time. For larger problems, today’s foremost mathematical minds cannot guarantee solutions within a reasonable timeframe and therefore a heuristic approach is required.

Vehicle routing is a well-studied field in the academic space. There are numerous algorithms, solvers and publications out there providing great results. The simplest form of TSP has a number of direct applications, but real world VRP problems usually have some unique properties which make applying research results a bit trickier.

What makes Milk Collection optimisation unique

So what we have is essentially a VRP with multiple depots, heterogeneous fleet, time windows, capacity limits on factories and vehicles and some restrictions. But what makes milk collection different from optimising a taxi fleet, a bike messenger service or parcel deliveries?

Unlike passengers and parcels, milk collected from farms usually does not have a predetermined destination. The destination of the loads depends on the production requirements of the factories at a given time. In addition, milk trucks can make several trips in one day, which makes the consecutive trips temporally dependent on each other.

The Solution

There is no silver bullet for finding the optimal solution; however finding a significantly better solution to that which a human is capable of is absolutely possible. So how do milk collection planners know if their optimised plans are, well, optimal? The simple answer is they don’t know for sure, just like the mathematicians cannot say for sure. However, by using the appropriate optimisation techniques for milk collection it is still possible to make quite significant savings in costs, distance travelled and related carbon emissions.  Usually these improvements quickly help planners and hauliers to overcome any lingering disappointment in modern mathematical advancements!

To learn more about OptaHaul’s modern approach to milk collection optimisation, visit www.optahaul.com.

Leave a Comment

Your email address will not be published. Required fields are marked *