Make the Efficient Transition From Cut-off to Continuous Planning.

Due to increasing logistics businesses demands, tighter window times, and changing laws and regulations regarding equipment, transport restrictions are increasing, and daily transport planning is becoming more and more challenging. This is why transportation and logistics companies are currently looking into advanced planning and scheduling (APS) software.

Advanced planning solutions are very capable of calculating optimal planning by using algorithms while taking all conceivable rules and restrictions into account. An APS can be used not only to support the daily operational planning process but also for strategic and tactical planning optimization. It can be used to answer questions such as "Is it smart to start using a warehouse in a certain region?" "Do I have enough resources to handle seasonal traffic?" "Does this new customer fit into the "balance" of our network?"

Despite the high investment associated with using an APS system, the potential benefits far outweigh it. Unfortunately, in practice, we see that implementing an APS-solution does not always produce the desired results. It often happens that after an intensive implementation there is little or no automatic planning, which was the main reason to invest in an APS.

Logistics dynamics   

The cause lies in the extreme dynamics that characterize logistics operations. Last-minute incoming orders, undefined constraints only known to planners, and outdated standard times on the basis of which the APS will calculate are common reasons why automatic planning does not work. Moreover, an APS assumes that in the time between scheduling and execution, the world remains unchanged. Of course, the opposite is true; delays due to traffic jams and wait times are commonplace.

A driver calling in sick can turn the entire schedule upside down. What seemed an optimal schedule during planning (the cut-off time) is no longer realistic once the first truck has left.

There's a reason why planners and trip supervisors shuffle orders, text drivers, and call charters all day. Because all or part of that wonderful planning schedule created the night before has gone haywire, and they still want to make sure all orders are delivered that day. When you look back at the end of the day, you see that the completed trips are different than originally planned.

APS as a spider in the web

To harness the true power of an APS and use it for increasingly complex and detailed scheduling, it is necessary to interface with execution. In many cases, that means linking the APS to the on-board computer system or to a driver's app to gain insight into the current status of the trip and its current location. Often, an APS system is able to calculate an arrival time based on that information and determine whether a vehicle is on time or running late.

The challenge lies in taking all circumstances into account. For example, is another traffic jam expected on the route? Or perhaps there is a threat of delays at intermediate loading and unloading locations due to congestion?

Execution management   

In the case of multi-carrier operations, in which different parties with different systems are responsible for the execution of one schedule, it is a major challenge to gain continuous insight into the situation on the ground. This is where execution management platforms come in, which are specifically designed to manage multi-shipper and multi-carrier operations based on different real-time data sources. Moreover, the impact of rescheduling or late delivery, for example, can be made immediately transparent through the entire chain via such platforms, from shipper to receiver, from planner to taxi.

Environment for continuous planning

Using this real-time information from various data sources in planning creates a so-called "continuous planning environment." There is no cutoff time. Planning is updated based on any change, no matter how small. By continuously responding to changes manifesting on the ground and taking into account newly arrived or changed shipments, the initial plan is updated each time.

Continuous scheduling is ideal in theory but has some practical drawbacks. Consider what would happen if the schedule could be changed at any time during the day—drivers would go crazy.

An intermediate form is that you "freeze" trips after they are forwarded to the driver, but you can still make changes within those trips during execution. This so-called continuous rescheduling can be useful if there is a traffic jam somewhere that you can avoid by changing the route (and thus the order of stops).

This makes this form of rescheduling quite complex. Planners and trip managers do this by heart, but you would actually want to use some form of automated support for this as well. After all, changes within a trip have so many ramifications and are subject to so many constraints that calculating all possible outcomes would take a chess grandmaster.

Machine Learning Algorithms

To respond to ever-changing ground conditions in a continuous planning environment, machine learning algorithms can help. They are able to factor historical data into planning calculations. For example, the default load time at a given address can be overridden if it turns out that on Wednesday afternoons, we load on average 30 minutes longer. Or even better: an intelligent, self-learning APS system can use this information to specifically not load this shipment on Wednesday afternoons.

It is clear that a dynamic continuous plan optimization solution goes much further than a static APS that makes the ideal plan at a certain time of the day (the "cut-off moment"). Of course, it depends very much on the situation as to what best suits your logistics process. If you're interested in what an AI-based continuous plan optimization solution can do for you, check out the Norma LIVE route optimization solution.

You can expect that such a system will be able to make sure that, based on continuous feedback from the execution, the APS learns to plan better and better and can tell the difference between deviations caused by incidents and deviations that were actually predictable based on known parameters. 

More about Norma LIVE

Norma LIVE is a digital route planner based on artificial intelligence. It combines scientific and evolutionary computing with machine learning to support transportation, logistics, the last mile, and home delivery operations through process automation and optimization. Norma LIVE helps its users improve, achieve more, save money, and scale.

Norma LIVE connects via API technology to both internal systems (e.g., TMS, execution management systems, and onboard computers) and external data systems, such as road network maps and traffic forecasts. By analysing all the information, it is able to create the best-case pick-up and delivery plan for orders within seconds. So, there is no one-time cut-off schedule but a continuously updated schedule based on real-time execution data and the most up-to-date shipments.

Download this white paper to learn more about tips for last-mile delivery options you can use to scale your transportation and logistics business.