problem ( i . e ., try out every possible route to find the best one ). But with so many possibilities , it quickly becomes impossible to solve this problem in any reasonable amount of time .
Planning - Creating efficient routes in real-life
As a leader in fleet optimization solutions , Autofleet employs advanced algorithms that provide real-world solutions to optimize delivery and scheduling .
We help you find a near-optimal solution — one that can be calculated in a fraction of the time , one that is viable , and a better option than waiting days for the ultimate answer .
We begin by calculating a viable and goodenough plan and adapting it by implementing incremental changes . It is easier to solve the problem once you have an initial solution , and it can be done efficiently and with great results .
The solution is supported by constraints based on your customer fields and desired optimization ( e . g ., you cannot deliver packages that need signing to a household at 5 AM or ask a driver to work a 20- hour shift ). We eliminate any solution that breaches those constraints quickly without letting them run their entire course .
Solutions that have already failed and ones that are similar to them are avoided . And solutions so close to the current plan that it won ’ t make much difference are also ignored .
Combining these techniques creates a near-optimal and efficient plan that can be used to dispatch drivers .
The Autofleet way : Route planning re-optimization on the go
Even with a good plan , new variables come into play - traffic , driver availability , last-minute cancellations , new orders , and more . This happens even when you do all the calculations and create a solid schedule . It never plays out as planned in the real world , and you need a new plan . But if optimized route planning is a difficult problem to solve once - before the beginning of a shift , imagine how hard and time-consuming it is to come up with a new solution each time the situation changes .
The naive solution is to simply add any change to the existing plan . While adding a single stop point may not make much difference , these changes pile up and accumulate . They can quickly degrade any plan to the point that it is unusable . Moreover , changes like a driver calling in sick or a major roadblock have a ripple effect that can make any planned schedule completely irrelevant .
The problem is meeting the need to reoptimize plans on the fly and adapting them to changing circumstances . And how to tell which of the plans is better ? The current one ? Or the new re-optimized solution ? Figuring that out is not a simple task ( remember , the number of possible options is huge !).
The evaluation process has several steps . First , the re-optimized plan needs to meet certain constraints , such as meeting all arrival windows on time .
Only plans that are within the parameters set by the constraints are moved on to the next step , where they are scored according to specific KPIs . And only the highest-scoring plan is dispatched , making sure the best plan is always in play .
24 customized logistics & delivery Magazine I spring 2024