The most important thing for a route optimizing tool is delivering a good and realistic plan. In the past, we have already discussed how bad data quality can be recognized. In this blog, we will talk about how to avoid bad routing plans.

What's a good plan?

When you have created a plan, it is very important to know whether it is a good planning or not. Some sanity checks must be done to consider a plan as good and realistic. Here are some checks to recognize good plans when making operational use of a route optimizing tool.

Some route optimizing tools, like OptiFlow, offer a good visual representation of the created routes. This makes it easy to recognize unrealistic routes. If there are no logical explanations for these unrealistic routes, for example due to restrictions, then fix the routes in the plan board or continue the optimization, so the optimization tool can fix it itself.

A set of objectives can be used as a checklist that must be completed before regarding a plan good and realistic. By simulating an ASIS-plan, we imitate actual executed routes. In this way, we can generate KPI's that can be used as targets that must be fulfilled.

3. Did the optimization stabilize?

In modern applications, you can follow the evolution of the KPI's in time. When you see that these have stabilized for several minutes, you can conclude that an optimal solution is reached. If not, the optimization can still be improved.

4. Sign off of the planner?

In the end, the planner is the one who knows every constraint by hart. So, it is important to always have the planner's sign off. This person will recognize infeasibilities most efficiently.

It is important to mention that an optimization tool does not replace a planner. The job function shifts towards a consultant who asks smart questions, gaining valuable insights. When there are, for example, violations on time windows, the planner quickly discover how this could happen.

How to avoid a bad plan?

• Good starting solution

To optimize a routing plan, the application must at least have a starting point. This starting point is the first solution it creates. By generating good input data and by setting the optimization settings right, we can point the optimization in the good direction, and so, it will be more likely to have a good solution from the start. A good starting solution will benefit the final solution and will facilitate the time needed to create this final solution.

This picture shows a bad starting solution on the left and a start solution like it should be on the right.

• Input data

Data is the most important factor for algorithms and optimizations. Bad data quality makes it impossible for the algorithm to find a good solution. So, set up your data with all constraints and try to integrate all your business rules into your data. This will help you to create the most optimal plan and maybe most importantly, it will make you save costs. As mentioned above, integrating all constraints and business rules into your data will help the application to create a good starting solution. This will benefit the final solution.

• Optimization settings

It is possible to point the optimizations in the good direction by setting the Optimization setting right. Optimization setting can be seen as thinking paths for the application. An example of such a setting is clustering, Which asks the algorithm to combine orders in the same area.

• User-friendly interface

It is a surplus when it is easy to make modifications in the application itself. By modifying the current planning or by modifying the data in the optimization tool, it will be easier to discover solutions for infeasibilities. Especially when the application has a user-friendly interface, which makes it easy to make these changes.

• Service desk

Support is needed and it is possible for users of optimization tools to request help from the provider. Do not hesitate to make use of it, because this can lead to new insights which will benefit the planning solutions. Besides this, it is also important to question the planning by yourselves and you should not be afraid to question today's planning method. Maybe copying your planning method into an optimization tool is not necessary and may be difficult. Small modifications to your method may generate big benefits.