A deeper dive into the algorithm
Without getting too technical , let ’ s explore the logic behind our algorithm a little more to understand better how re-optimization works to adapt and improve dispatching in real time .
The bottom line : route planning is great in theory , but route re-optimization makes it great in reality
It is only possible to scale a modern delivery service using a route planning or dispatch software solution .
The dynamic nature of delivery services , including dealing with real-time traffic , drivers , and last-minute changes , and the need to accommodate ASAP and planned requests , means that plans can become obsolete almost as soon as they are ready .
One of the best shortcuts when re-optimizing a plan is eliminating solutions that failed before . This is called “ backtracking ,” and it allows us to start with fewer options , avoiding those that we already know do not work .
We can also implement “ domain reduction .” This takes into account the current state of the fleet as a starting point . It places constraints on the search for the optimal route based on the reality of the situation . For example , if a driver is about to end his or her shift , we don ’ t give additional out-of-the-way deliveries . If they are heading north , we avoid turning them around .
Another optimization is “ localizing ” the search . We swap a couple of routes at a time to see if the change generates a better result . If we see an improvement , we can propagate the change and test again .
Lastly , there are “ heuristics ,” which are the closest thing to common sense math has to offer . For example , we can ensure that we avoid changing routes for drivers who are already close to the pick-up point , cluster rides going in the same direction , disregard routes that are already very similar , and take into account the geography of the area .
Experience shows that companies that understand that and move beyond basic planning to adaptable , re-optimized dispatch schedules perform better in terms of efficiency . They provide better service with a smaller headcount and with more consistency .
Some of this may sound like math gymnastics . The bottom line is that for real-time , optimized route planning software to deliver , there is no getting around automated and optimized techniques that adapt routes to a changing reality .
If you want to learn more , feel free to contact us at contact @ autofleet . io CLDA
Seth Hochhauser is VP of Strategy at Autofleet , where he is responsible for Business Development , Partnerships , and Strategic Projects . He previously worked with one of the largest for-hire carriers in the US , building their digital trucking unit with advanced optimization capabilities . He is a fan of big dogs , big rigs , and Biggie Smalls . You can reach him at Seth @ autofleet . io . spring 2024 I customized logistics & delivery Magazine 25