Choosing the Best Route Optimization Engine | Part 3

With the high demand and expectations for delivery services, businesses often feel the pain of scaling up their services and improving the order ETAs. When it comes to route optimization and fleet dispatching, few solutions are truly able to scale with the increasing volume and demand. Common pain points are:

1. Waiting for ages before you get a response from the solver.
2. Solution quality degenerated with increasing volume.
3. Software can only fit a certain number of orders into one request.

Keep reading to learn more about the benefits of choosing the best route optimization engine— LogisticsOS.

How to Know Your Vehicle Routing Solution is Scalable?

The scalability of the Vehicle Routing software can be measured in the following ways when dealing with large datasets (5,000 stops or beyond):

1. The capability of finding near-optimal solutions.
2. The capability of handling complex constraints (time windows, traffic, etc).
3. Time spent on finding a solution.

The Experiment

Let’s take a look at a real example of last mile delivery in the great Toronto area (GTA). There are 9999 orders to be delivered by 200 drivers. Each order has a time window, service duration, and two-dimensional delivery quantities (weight and volume). Each driver/vehicle has a shift schedule (availability time slot) and capacity. All of the drivers shall be dispatched from the same warehouse and dismissed after their last delivery. Let’s run some simulations with different options together.

Minimize Total Distance with TomTom map data and Historical daily average traffic patterns:

Set the optimization goal for the solver to minimize total distance for all drivers (Solution details see pic). We do not know the exact optimal solution to this dataset, but we are confident that our solution is not far from optimality by looking at the plot. The distance/duration matrix was computed with TomTom map data with historical average traffic patterns. We use straight lines between stops and the lines between the start depot and the first orders on each route were omitted for better visualization.

Minimize Total Time with TomTom map data and Historical daily average traffic patterns:

Set the optimization goal for the solver to minimize the total time (traveling and waiting) for all drivers. Compared to the previous solution, the plan has less total time and more distance. Also, the average speed for the vehicle increased since the solver navigated the driver to faster routes (with less traffic) instead of routes with shorter distances.

Minimize Total Time with HERE map data and Predictive traffic patterns:

Switching to the HERE map with predictive traffic patterns, set the dispatch time to 3:30pm the next day. We re-optimized the example instance and got a new solution. The first thing to be noticed is that vehicles tend to travel faster in the current solution. There are certainly deltas in the map data between providers. However, traffic patterns dominated the differences, as we can notice that there is a huge increase in distance and a huge decrease in travel time.

Straight line plot VRP solution

Within 15 minutes, we finished 3 runs with the test case. We iterated through different optimization goals, map providers, and traffic patterns. This is why LogisticsOS optimization engine is deeply loved by customers. Instead of waiting for ages before getting a response, dispatchers now can solve multiple times to find the best dispatching strategy.

About LogisticsOS

LogisticsOS is the most powerful route optimization system on the market. Our flexible, hosted route optimization and planning/replanning APIs require no knowledge of optimization techniques. High-quality results are returned within seconds— even at scale. From time reduction and vehicle capacity to depot automation and cost customization, our feature-rich software can create the efficiency and cost savings that you need. Contact us today for a free product demo.