This demonstration considers a metal works use case with typical equipment, different workflows and constraints. The goal is to reduce tardiness, eliminate delayed deliveries and reduce time lost in setup and other process complexities.
We start by loading data that could also be extracted from an ERP.
Let’s generate a complete schedule using the EDF policy.
This only takes a fraction of a second.
This method has poor performance.
We see almost 500 hours of total tardiness, 26 late jobs, 240 hours lost in setups.
To improve that, instead of planning manually, we’ll generate 50 different schedules just with a simple click.
The best schedule already improves the KPIs a lot, but we’ll go further with another simple click.
AI will now iterate though 100s of schedules, combining them and trying to come as close as possible to the optimum that can be reached in this situation.
As you can see, the situation now looks much better, but there is still room for improvement.
Let’s try to find the root cause for the delays and eliminate any remaining tardiness.
You can see here only a couple of production orders have delays.
On the shift utilisation chart we see that, although many stations are highly loaded, adding a few extra shifts on some of the lathes will probably solve our problem.
Now we do a little further optimisation and we can eliminate every delay and every late job. There we are.
In a few seconds ORITAMES started with generating a simple plan, then generating a whole population to already improve, using AI for optimisation, identifying the bottlenecks to allow the user to add just extra capacity where required to come up with the perfect schedule.
If you need a quotation for ORITAMES, else if you need an APS implementation plan or you want to try ORITAMES with your own data, please contact our partner MangoGem and reference this pminsight post.