Hello everyone and welcome. We're excited to have you join us for today's webinar, Explore Automation, Possibilities in Production Scattering with SAP P&G. As we kick off the new year, we think this is a great opportunity to look ahead and discover new ways to improve our scheduling processes. We have 90 minutes and our goal is to give you some good insights and inspiration to make you more efficient and resilient in production planning. Thank you for being here today. Now let's dive into the session. Before we get started, let me quickly walk you through today's agenda. We'll begin with a short welcome and recap from our last webinar, followed by a refresher on the basics of SAP P&G. Then we move into a live demo, hear insights directly from SAP and wrap up with a Q&A and closing remarks. If you have any questions along the way, please feel free to post them in the chat and we have some colleagues online to help answer questions as we go. And we'll also address additional questions during the Q&A at the end. Let's also start with a brief introduction to our Implement team joining us today. With me are my colleagues and production planning experts, Thomas and Mads. Thomas is a partner at Implement with over 20 years of experience leading PPDS projects. Mads joined Implement last summer and is a part of our Aarhus Denmark team. He primarily works on supply chain planning projects. I am Isabella from my Hellerup Denmark team. I also work in supply chain planning and I'm currently implementing a PPDS project with Thomas for a client overseas. I'd also like to highlight our colleagues, Sarah and Julian, who are supporting today's session with the interview and SAP demo part towards the end of today's session. We at Implement Consulting Group are more than 700 experts working with all areas across supply chain management. We support our clients from strategy to execution and we combine deep industry knowledge with hands-on delivery experience. This is our seventh webinar of our PPDS webinar series on PPDS and SAP digital manufacturing. We have worked with SAP PPDS for many years and want to share some of our knowledge to support our current and future clients. If you want to be updated on any of our future events or get access to previous webinar materials, just reach out to us. All right, please scan the code and go to menti.com and answer a few quick questions. This will help us understand who you are so we can tailor our content to your industry and needs and make our future webinars more relevant for you. We will give you a few minutes now to scan the code and answer all the questions and then we will be back shortly. Okay, and where are you in your PbDS journey? So this is a good mix. Some of you have just learned about PbDS. Some are moving from APO to S4HANA. And there's also quite a few of you looking to advance some existing PbDS. That's great to see. We'll try and adapt that for future webinars. All right, we'll go back to the presentation now. Thank you for providing all those answers. All right. Now we're moving on to the first topic for today. That's the recap from the last webinar. I will hand over to Thomas and Mads. Yes. Hello and welcome back. Great to see so many people here. And great. And great. I saw some of the ones that we have seen before. So that's great that you're following this topic. First, we will try to do a little bit of a recap of what we did in the last webinar. And as we saw before, this is number seven. So there's a lot of different ones. But I think the last one, we very much touched upon how to transition from SAP APO to PbDS. And one of the main points here was that we should really start focusing on the people who are there, what are they working with, and why they do what they do. And then we kind of focused over in the process, what is the kind of best practice process for PbDS and how do we understand the process and the complexity of the business to look into this. And then we kind of finally looked into working on the system side of this. And you can say when transitioning from, and this is a business slide, and last time we did a lot of discussions on this, but a lot of companies are primarily transitioning from an HSE system towards an end state where we have S4 together with PbDS in S4. And you can say the first three scenarios here was kind of saying you're coming from ECC, you are upgrading this to S4, and then you're applying advanced planning and PbDS on top of that. Or you come from a scenario where you have ECC and APO PbDS. You can actually also start by upgrading to S4, keeping your PbDS in APO, and then at a later stage moving that into S4 PbDS. So S4 can work perfectly together with APO as well. And kind of the third scenario was kind of to take everything in one go, so both upgrading to S4 and implementing and transitioning to S4 PbDS. And then some of the scenarios in number four here is very much the little bit more complex one where you actually end up with a landscape where you both have S4, PbDS inside S4, and IPP. So if you're coming from a scenario where you used APO both for scheduling but also for demand planning or for supply network planning, then the scenario could be that you start with S4 with PbDS but keep demand planning and supply network planning in APO. But there could also be scenarios where you replace the APO with IPP. And all that is supported in different kinds of ways. So it's kind of a little bit up to the process and where you are. The last thing we kind of touched upon or will touch upon again is that a lot of these projects really are started up because we want to have very advanced planning and we want to utilize AI to do scheduling and so forth. And this is kind of the roof of this house here. But I think the main thing to improve and start up with is, of course, to work on the planning processes, which we see in the bottom, and working on your master data and working on your housekeeping things. Because if these things are not in place, you will not gain the benefits of the top part. So I guess this was kind of a little bit of a recap of what we did the last time. And now for this session that is very much about different scheduling opportunities and planning opportunities, I'll hand over to Mesh to take a little bit about, I think, the slide part before we go to the demo. Yeah, exactly. So as we have now covered what we have been through during our previous webinars, as Thomas mentioned, we will now move on to more on the basics related to heuristics and optimizers in SAP. So supply chains are moving from automation now to augmentation by 2030 and toward autonomy after that. And we expect that this will reshape production planning and scheduling. And right now we are in a place where we are in automation with tech like RPA and Internet of Things, handling routine tasks so system proposed plans for planners to review. After this, we will move into augmentation where machine learning and digital twins will enrich supply chains so decisions get faster and better. And then over time, we aim for autonomy where AI will self-optimize within some guardrails, which will allow people to focus more on strategy and exceptions. And as the Gen Zs, at some point at least, will step into management, there will be expectations for intuitive data-driven tools to rise, which means that we have to, or which means that the push to modernize PBDS will accelerate. So therefore, with this path in mind, the simple message will be win with automation in PBDS today, build toward augmentation, and be ready for autonomy. So today you might find yourself in a state where lots of planning and sequencing activities are manual, which can be inefficient, it can be open to mistakes, and it may even have a dependency on local knowledge. Therefore, our ambition is to move toward a more standardized and optimal approach using automation. That means that we will save time in planning, we will balance capacity better, and increase efficiency. We will improve delivery performance, and we will make the process more robust. So in short, you could say that we are shifting from a person-dependent scheduling, or from person-dependent scheduling to a more tool-supported sequencing process, where planners will spend less time clicking around in a system, and have more time to making good decisions. Cool. And just to kind of set the basics in this, as Matt said, we are moving towards less clicking and more automated planning. We will give you a little bit of basics within what the options are within PBDS, and we will do this in the way that PBDS is actually kind of two things. It's PP, which is about order generation. I'll touch a little bit on that, and what are the options within PBDS for this. And then Matt will kind of continue on the DS side, which is the detailed scheduling of the orders. So when looking at how to generate the production plan or generate the plan orders, you could say there's some different sources for this. It could be from one end, very manual. You kindly type in the orders that you want, and of course that's on the lower part of the graph that Masjad showed. But it could also be that orders can be generated inside S4. We'll touch a little bit upon that. Or they can actually come from the IPP system. So if you're using IPP for either supply planning in time series or supply planning in order-based planning, the result of that can be integrated back into S4 and PBDS, and then you can work with the plan orders here. But this is not a PBDS, an IPP webinar. This is a PBDS. So we will kind of focus on how and the options that exist to generate the plan orders within PBDS. And here we kind of look into two things. There's something called heuristics, which is kind of similar to MRP. So it's calculations. I'll come a little bit into that. And then there is the more advanced one, which more support the end game in the graph before when we have automated planning, which is the PP optimizer. And this is also what you're going to see later in the demos that we have in this webinar. So to go in a little bit to show the difference between what the PP optimizer and you can call them the PP heuristics do, they kind of have the same, they have to fulfill the same, they have to create some plan orders, but they have different functionalities. And you can say that where the PP optimizer, there's some things about it. It works in buckets and not in time contingency. A little bit, if some of you are not Gen X like me, but from the old days where we have APO, then it was kind of like the S&P part here. So it works in buckets. It has the option to choose it because it's an optimizer. If you have several possibilities to produce, we have different routings or production versions in PPJs called production data structures, then the system, the optimizer can choose between these options. And then the main thing is that the optimizer is a constrained planning solution that works with the constraints that we have modeled, which could be capacity constraints and penalties of not delivering. And then it tries to make the orders based on that. While the PP heuristics, they are working time continuously. They are using rules. So if the rule is to use production version one, then it will take that. And then it cannot, depending on capacity, take another one. So this is an unrestrained plan. So always try to fulfill requirements, no matter if you have the capacity or not. And of course, that comes with a second step to do something about the orders then. So it's not an either or. We typically see that these things work together. It could be that you're using the PPO or the production planning optimizer for the more complex scenarios and maybe also for the short time horizon or for some products. And then you use, for example, the PP heuristics or just the MRP NS4 to generate orders maybe on a more long-term horizon, but also for the less complicated or complex scenarios where capacity constraints might not be that big a deal. So if you look at the PP optimizer, there's, of course, some things that it can do and there's some things that it cannot do. And basically, as I said before, it tries to optimize the production plan based on these constraints, cost and penalties for not delivering to the orders. It works in buckets. It can select between different sources of supply. It could also be that it selects between producing in-house or actually procuring product. It works on several levels in the bombs. So if you don't have the raw material or semi -finished good capacity for producing this, this will have an impact on the finished goods as well. So it actually works throughout the levels. It can work with procurement, as I said. It can work with consignment stock and contracts. And it actually also supports tank planning functionality and resource networks that exist in PPDS. And then it works very well together with the flexible constraint interface from IPP. What it cannot do, and of course, being a constraint plan, it cannot always fulfill all demands. So if we have more demands that we have capacity and we do not allow this to be late, then there will be some demands that are not fulfilled. It also is an optimizer. So it tries to do a lot of things that maybe is hard for the planners to understand. So it's not like it shows, OK, so we have this in stock and this in demand and so forth. Therefore, it creates an order like this. So it only kind of shows the result. Why has it put us as it was? And sometimes that's a little bit hard for planners or anyone to understand why have the optimizer optimized as it should. It creates orders within the bucket or it cannot create outside the bucket. And then there's some scenarios that are still not supported but in the roadmap. For example, working with engineering to order and so forth. And then it does not. It creates the plan orders, but it does not optimize the sequence of the plan orders. But you can actually activate the DS optimizer. So now we have a PP, production planning optimizer, and we have a DS optimizer for scheduling. I will just say a little bit about the DS scheduler before I go into MES that can explain some of the other options that exist for scheduling. So the PP optimizer is working on orders and master data and costs and penalties and trying to create a constraint plan that is fulfilling as much demand as possible, taking the capacity into constraints. And the DS optimizer that we see in the bottom here is working on the PP optimized plan. But then it's working on several other constraints like what kind of changeovers that we have between making one order after another one, also called a setup matrix. And then it's trying to actually still fulfilling the optimized plan, but also trying to minimize changeover time or duration of the entire plan. So, and again, these things can be combined. So running the PP optimizer first and then adding the DS scheduling optimizer on top of that to sequence the orders. So this was a little bit about setting the scene on the order creation and the optimizer way of doing that. Now we also kind of have other tools to do scheduling, which I'll hand over to you, MES. Thank you, Thomas. That was a great intro to the PPO. And as you just mentioned, we will now take a look at the different planning and scheduling heuristics in SAP. So in this webinar, we are focusing on scheduling heuristics, as mentioned by Thomas. However, in SAP PBDS, there are multiple other groups of heuristics as well, also focusing on procurement planning, process control and service. And here it's important to remember that each heuristic optimizes for something different. For example, it might minimize completion time, maximize resource utilization or protect due dates. So in PBDS, there are many predefined options as illustrated by the blue table on your right hand side. For instance, we can schedule sequences manually when needed. We can do stable forward planning to keep plans calm. We can run backward scheduling heuristics to hit the due dates. Or we can choose to work bottom up or top down depending on priorities. So choosing the right heuristic starts with how it works, why we use it and what to watch out for. As you can see here on the left hand side, we have the option to schedule sequences manually, which is the flexible planner driven choice. Here you can select the operations and set the order and then SAP will adjust the timing based on your strategy settings. So in essence, you would say that this heuristic is easy to explain. So it works well for quick fixes or special cases, I would say. However, although this one seems rather simple, you have to be aware that it's not designed for mass scheduling and it can struggle when there's a lot of different alternatives. If we now take a look on your right hand side, we have the stable forward planning heuristic. And this one runs finite scheduling from now and forward across multiple levels. And it will take what the MRP created and turn it into some sort of feasible plan, which will clear backlogs and ease overloads. And I would say beside these points, it's also important to mention that this heuristic keeps anything with fixed due dates. It will keep anything with, you could say, fixed dates as is, meaning that you will preserve firm commitments. But then on the other hand, you have to be aware once again that the limitation of this heuristic is that it doesn't really handle advanced constraints such as blocks, synchronization or campaigns, meaning that you shouldn't expect this heuristic to solve those scenarios. And then if we move on to the next slide, we will have two more examples of scheduling heuristics. Here, if we look on the left hand side once again, we now have what we call the enhanced backward scheduling heuristic. This heuristic will plan from the due date backwards using available capacity to build a feasible short term plan. And in essence, you can say that this heuristic propagates feasibility up and down the bill of material through pegging. So resources are deloaded and operations are aligned, which will, yeah, at some point reduce work in progress and lead time. And then as we've done with the others, what you have to be aware of for this heuristic is that it more or less focus on capacity feasibility rather than perfect sequencing. And you also have to be aware that delivery dates are treated with some flexibility, I would say. If we now take a look at the right hand side, we have the bottom up or top down heuristic. And this heuristic uses pegging to reschedule connected orders starting at higher bump levels. So work is somehow released when components are truly ready. And although it quickly relieves overloads and stabilizes the plan around the real constraints by improving service levels and throughputs, it does require very good bottleneck identification and reliable pegging. Otherwise, the focus may mislead the plan. So now we have presented these different heuristics and Thomas has talked a bit about the optimizers. But eventually deciding where heuristics and optimizers bring the most benefits isn't always easy. And it doesn't always have to be a one of the other, as Thomas mentioned earlier. In this regard, we have seen many clients actually where they have a blend of both background optimization jobs and then they use heuristics for refinement. But I would say if we have to put just one sentence to the difference between heuristics and optimizers, it would be that heuristics will do exactly what you tell them. And then on the other hand, we have the optimizers where you really need to tell an optimizer what it should avoid doing and at what cost. And then it will try to find the most optimal result within those boundaries. And then you have to be aware that if you decide to run the optimizer and you think the result is wrong, then it's most likely because you didn't tell it that it wasn't allowed to do whatever it's done. Which means that these optimizers are not just plug and play and the quality of the results need to be evaluated. And I would say adding to these different points, you also need to factor in maturity and the quality of both transactions and master data. And I think, Thomas, you will elaborate a bit on this. I can just highlight up here in the end that you could say that this is kind of a letter about how advanced you will try to go. And as Mas also said, it's not an either or. You can't, of course, it makes sense when you are just starting up trying to use the less advanced models because then you have an overview of what's happening and what is it doing. But, of course, it will not help you in just having put your hands off and then let the system calculate. And, of course, that will happen when you try to have more knowledge about what is going on. So maybe you would start with this manual scheduling or the heuristic that Mas was talking about. So then you can kind of go to some of the other heuristics that are still kind of working with some rules, as you said. So if it tells us to try to do backwards scheduling to avoid that we are delayed, then it will try to do that. And then once that is in place, then trying to do the different types of optimizers or the optimizers with different settings. So it's not like you switch it on and then everything is just running and all the constraints are there. You can actually switch on the optimizer only focusing, for example, on minimizing the changeovers and, of course, delivering to orders. But it doesn't necessarily need to take into consideration what do we have in warehouse capacity or have we modeled our trucks and everything as well that you could add on top of it or you can then, again, add it even more on top of it to include working with shelf life and campaigns and so. So I think the takeaway from us here today would be that there is different planning opportunities, both for order creation and for scheduling. We can do it with different heuristics. That gives you some help. It's not just an MRP. It does different things. And then there's the optimizers, both for scheduling and for order creation. And you can combine them as you kind of want and on your journey. And then the idea would be to start with, of course, as much simple as possible, working towards a more automated way. And I think this was kind of the talk about the basics and what is going on. And now I think it's time for you to actually see some of the tools and actions. So we will hand it over to Isabella that will introduce the demos. All right. Yeah. So that was all we had on the basics. Now we're moving on and let's dive into our demo sessions. We have two different demos coming up. The first one is around the production planning optimizer. And this is not a new feature, but we would like to show how can simplify a planner's life. Let's go into the advanced scheduling board. Here we can see our current production on our scooter assembly line. If we open up the legend icon, we can see that the green bars indicate setup time. And the triangle in front of the green bar indicates that it's sequence dependent, meaning it takes into account the product that has been produced before. We have not maintained the red racing scooter in our setup matrix yet. The next step is to maintain the setup matrix for the individual characteristic. For that we go into maintain characteristic value matrix. Adding another color is fairly simple. We just click up here on generate setup transitions. And the system checks that we have our racing red city scooter now and therefore adds the color red. In this green, we can decide which rules the setup group generation should follow. Characteristic 1 for us is our scooter suspension type. Characteristic 2 is our scooter color. Generating the setup matrix is actually fairly simple. We go into CDBS MatGen and generate our setup matrix in our location P100. We now receive the final setup matrix based on our suspension, so either city and off-road. And our color, so green, yellow and also red. Off-road green is not used anywhere. And that is exactly where the setup matrix profile comes into play. We can see in this screen that the PDS-based determination is ticked. Essentially, this means that during the creation of our setup times, it takes into account all our characteristics, but it only looks for available production data structures. And as off-road green is not used anywhere, it does not create unnecessary changeovers. Now it has taken into consideration that we are now in this case changing over from the green to the red scooter. And that we are producing afterwards our off-road scooter in yellow. Good. If this short recap of our last webinar was maybe too fast or maybe confusing, I would highly recommend watching the recording of our last webinar, which we held back in September. It was called Boost Production Scheduling Efficiency with SAP PPDS. And my colleagues Chris, Johanna and I, we were talking a lot about the setup matrix there. You can easily watch the recording by receiving the slide from us after this event. And then simply following the link that is behind the webinar. And then you can watch it from home, from work or wherever you feel comfortable to do so. For now, however, I would like to guide you back to our current scenario, which we are also using for this webinar. And that is the production of our city scooters and our off-road scooters. We produce both types of scooters on two assembly lines, where, however, also in today's scenario, we do not want to use the second assembly line. We only want to use the primary assembly line. They consist, of course, of different semi-finished goods. In our case, the baseboard and the handlebar of the scooter for both types. The only difference is the suspension that we use for the different scooters, where at the city scooter we use, of course, the city suspension. And for the more bumpy roads, we use the off-road suspension on our off-road scooter. The semi-finished goods have their own assembly lines. So, the baseboard assembly line, which we will, however, not take into consideration today, as well as the handlebar assembly line. What are we going to cover in this demo? Just very briefly, we are going to look at the basic settings of the PP-Optimizer. So, we will guide you through the material master and the cost setup. Then, we will discuss a very simplified scenario, where we only have a look at two different scooters and different demand types. And we will also compare a heuristic with an optimizer. And then, finally, we will have a more complex and more realistic scenario, which is then also more interesting for the more complex real-life use cases. So, stay tuned. It's going to be interesting. To jump right into it, let's start with the basic settings. And we do not want to guide you through all of them, but we want to guide you through at least the material master. So, we are going to our Fiori and to display material. We are going to look for our city scooter in green. And we go to the advanced planning tab. Because in the advanced planning tab, if we scroll all the way to the bottom, we can see different penalties and costs and also different demand penalties. As we have told you, the PP optimizer is working based on cost. It's not real cost. It's funny money. But in planning, it is trying to avoid those costs, of course, trying to minimize the overall cost. In this first very simple scenario, we have set it up that we do not consider any procurement costs or any safety stock penalties. Also, no max days of supply penalties. But we don't want to have excessive stock. We want to avoid heavily overstocking. That's why we just put in a bit of cost on the storage. Additionally, you can see that we have two different types of demand, which have different penalties. So, if we delay a customer demand, where the optimizer is going to calculate with penalty cost of a thousand. And if it's not delivering to the requested order, it's even five thousand. On the other hand, if we neglect the forecast and are too late, the penalty is ten. And at the same time, the penalty for non-delivery is 50. This way, we are prioritizing customer demand over forecast. So, actually realized demand is more important to us than forecast. And as I said, in the very first very simple scenario, we have only two scooters. So, if I go to product view, and I will also start with the city scooter in green, I can see that we have a forecast requirement for the mid of April of 20 pieces. And our second scooter, if I go to the city scooter in racing red, I can see that also for the mid of April, for the same day actually, we have two different sales orders. One with six units and one with nine units. So, what are we going to do with these? In the first scenario, we want to compare a heuristic with the PP optimizer. So, let's first run the heuristic. For that, we go on edit and run a product heuristic. We'll run a lot for lot heuristic and just adopt. And as we can see, it has created two planned orders. One for six units matching the first sales order. And one for nine units matching the second sales order. Now, we can do the same for the city scooter in green. And of course, we would in real life not do this product by product. But this is just for demo purposes. We're going to run the exact same heuristic on a lot for lot basis. Going to adopt it. And it has created a planned order for 20 units. Now, we have created three planned orders. Let's go on save. And let's have a look what this means for our capacity utilization. We go into monitor capacity utilization for our scooter assembly line. So, this is our scooter assembly line. And we need to expand the horizon as we are in January. But the demand is in April. And we can see that all the demand falls into the same week. And unfortunately for us, that means that in this specific week, on April the 17th, we are overloaded by 4% and our capacity utilization is at 104%. Now, we told you that the PP optimizer plans in time buckets. And we also said that the PP optimizer cannot create orders that are longer than this bucket. So, let's see what the PP optimizer does compared to the heuristic. Let's get back out again. And let's execute the production planning optimizer. We have already set up a variant in order not to waste time in this webinar. We have just selected a more specific planning horizon. Selected the different products in our location. And we have also selected an optimizer profile to refine the settings. We have defined our time buckets. And we have also defined the different demand types. Let's hit execute. And then it's going to run for a few seconds. And we're going to see that it has run without an error. And we can watch the result now in the product view. In the product view, if we go to our city scooter in Racing Red, we can see that it has now combined the two sales orders into one planned order. So, different to the lot for lot heuristic, it has now of course created only one planned order. At the same time, if we now look at our city scooter in green, we can see that it has split the forecast requirement of 20 into two planned orders. One with two units and one with 18 units. And why has it done so? Well, the answer delivers our capacity monitor. So, if we go back into our scooter capacity monitor and select the 26 weeks horizon again, we can now see that instead of having the overload of 104% in calendar week 16, we now have a load of 98, while we also have a 6% load in calendar week 15. So, if we look at the different materials, we can see it's producing the city scooter in green already in calendar week 15, at least the order with the two units. And the reason for that is quite simple. The production planning optimizer does not exceed the length of the time bucket. And, in order to avoid storage costs, which we have set up in the material master, it is now only producing two pieces in week 15. So, in total, it's splitting the one forecast requirement into two pieces and producing only two pieces in calendar week 15. And now the question is, why have we not produced the Racing Red City scooter first and have two units of that stored in our inventory? The reason for that is quite simple. With the production planning optimizer, the term priority means something different. It does not mean priority that you come first or that you need to be produced first, but priority means you are the last one to be removed from the bucket, so to be moved from one bucket into a previous or into a later bucket. Of course, it does not move into a later bucket here, because we have set up a lot of delay and non-delivery cost. Good. Let's come to the final scenario. What we have shown you so far was very simple, because we only had the city scooter in red and in green. Additionally, we also only had one assembly line open. Now, we are not only going to add the city scooter in yellow and the off-road scooter in yellow, but we are also going to open up our second assembly line. Additionally, we have modeled various demands. For now, we only had forecast on one of the scooters and only sales orders on the other scooter. This will change now. Now we will have a mixed demand for all of them. Let's see first what the heuristic does. Actually, we are going to jump a bit and go directly into our scooter capacity monitor, because we have prepared this beforehand. And what we can see is that the utilization in calendar week 15 is at 240%. Of course, because we have not scheduled anything yet and because the production planning heuristic that we just used does not consider any bucket sizes or any capacity constraints yet. Also, we have a slight overload in calendar week 16, which is, of course, not ideal. If we look at our second assembly line, then we can see in the same time horizon, it's actually empty. So the second assembly line is not used, because as a standard, we want to produce on our first assembly line, and therefore the heuristic does not create any orders on the second assembly line. Let's have a look at what it looks like if we do the same thing with the production planning optimizer. So now I am actually going to open up both of our assembly lines. First, let's have a look at assembly line 1. And on assembly line 1, we can see now we have a more balanced picture, even though due to rounding errors, we have a slight overload of 1 % in calendar week 14. And we can also see, if we look at the material, that the orders are now more split out. If what we can also observe is, if we look at the legend, our off-road scooter is actually missing from our assembly line 1. And that's fairly simple, because it's produced on our assembly line 2. So basically what the optimizer has done is, it has moved, due to the lack of capacity, the off-road scooter to our assembly line 2. It could have also produced the off-road scooter on assembly line 1 in the weeks before. However, as we have model storage cost, now it is moved to the other line. And as you can see, calendar week 15, we reach exactly our capacity limit, while then we have other orders moved to calendar week 14 and 16. And as you can see, we're only producing our off-road scooter in yellow. All right, we're back. That was all we had on production planning optimizer. I hope you found that demo as interesting as I did. And now in a moment, we'll display the next demo of the detailed scheduling optimizer. Now that we have created the orders for the scooters by use of the PP optimizer, it is time to schedule the orders, as the PP optimizer is not able to. It only creates orders as mentioned earlier. Therefore, in this demo, we will initially show how the planned orders generated by the PP optimizer is received at the scooter assembly line. And in order to showcase the difference between scheduling heuristics and optimizers in SAP, we will first show how to schedule the orders by use of a scheduling heuristic, whereafter we will show the basic settings of the DS optimizer and how to apply it for scheduling orders. In Fiori or GUI, if we go into the advanced scheduling board app and choose our display horizon up here, which in our case could be from the 27th of March until the 20th of April, say apply, we can then choose our resource, which in this case would be the scooter assembly line, as we see right here. When we say OK and go, we will be able to see when the planned orders are scheduled at the assembly line here. What you see right here are some blue columns. If we go over here and find the legend, we will be able to see that these are planned orders. Also, the triangle we see right here means that we have a sequence dependent setup. If we close the legend and look at the line here under the orders, we know from one of our previous webinars that this line tells us that the orders are overlapping. By clicking the arrow on the resources over here, we are able to see the overlaps. So for instance, right here at the scooter assembly line, at the same time, we are scheduled to assemble the city scooter yellow, the city scooter green, and the city scooter racing red. This does cause us some issues as our scooter assembly line has a maximum capacity of one, meaning that we can only assemble one scooter at a time. Therefore, in order to ensure that all the planned orders are assembled in time, we have to schedule the orders. As explained earlier in the webinar, orders can both be scheduled by use of heuristics and by use of the DS optimizer. To illustrate the two methods, we start out by scheduling the orders by use of a heuristic, whereafter we do so with the DS optimizer. In order to use the heuristic, we start off by selecting our resource over here, which is the scooter assembly line. We then click on automated planning from where we select heuristics. In here, we click on the heuristics tab, from where we will see a list with various different scheduling heuristics, including the ones we have covered during this webinar. For our demo here, we will choose the stable forward planning heuristic, which will allow us to resolve the capacity overloads. Let's try to find it. It's right here. Just click it, then it's selected, and we say OK. And when we have confirmed that we have chosen the right heuristic, we say execute, which will start the heuristic run. Now, when the green tick mark appears down here in the bottom, we have an indication of that our heuristic has run successfully. If we now go back to the advanced scheduling board by clicking up here on the back tab, we can now see that our orders has been scheduled and that we have no more capacity overload. As the stable forward planning heuristic creates a feasible plan by random sequencing of the orders, the plan is not necessarily cost optimal. Therefore, in order to get an even better plan with respect to cost or time, we can run the DS optimizer instead. In order to do so, we have to once again select our resource, go into all that made a planning, but this time we choose optimize instead. Here we start off by choosing our optimization horizon. Here we can just choose the same horizon as we did earlier, which was the 27th of March till the 20th of April. However, adding to this, an important note is that the start of the optimized schedule should be within the horizon and cannot be before. Otherwise, it will not allow us to continue. And here for this scenario, we will just set it equal to the start, which is 27th of March. When we have chosen our horizon, we click optimize. And we then arrive at this screen, where we start off by choosing the range at the top right here, which allows us to change the optimization settings. As mentioned earlier in the webinar, the longer runtime, the potentially better results. And right here in the top of the basic settings, we're able to adapt our runtime. So although a bit, a longer runtime may be better for this demo purpose, we'll just set it to one minute. If you scroll a bit down, we are able to adapt all the different parameters at which we want to optimize. Here it is possible to make the objective function consider parameters such as make span, setup times are cost, delay cost, and mode cost. All with different rates represented by the circle diagram here to the right. So right now we see an equal split between make span, total setup times, and total setup cost. For this demo, we will schedule the orders by optimizing on setup times and delay costs respectively. So let's start out with setup times. In order to show the difference, we will set setup times to 100 and let the others be 0. Do that here. 0, 0, and 100 for total setup times. And when we have done that, we will click down here in the bottom right corner on the green tab, which then will update the objective function. In this case, only to consider setup times. Whenever we see the change here, we know that the update has been completed and we can exit it up here in the right hand corner. And in order to make the optimization run, we click on the icon up here in the top, lift further range. When the optimization run has completed, we will be taking automatically to this page where the green ticks over here indicate that our run has completed successfully. If we go back to the advanced scheduling board by clicking up here once again, we can now see that the sequence of the orders on the accepted line has been rescheduled. Here with respect to setup times. In the example we just showed, you may wonder how the setup times are reflected and can be validated. For this, we have another short example. As we can see by the legend over here, the setup times corresponds to the green columns. However, depending on the setup length compared to the production time, you may have to be aware that it can be difficult to see the setup time in the scheduling board because it becomes too small compared to the remaining. Therefore, in order to sanity check that everything is in order, we can use the zoom function up here. Then we can close the legend, click the zoom and as we narrow it down, you become able to see the green columns right here between the different orders reflecting the setup times. As mentioned earlier, we also want to illustrate a scenario where the objective function solely considers delay cost. Therefore, in order to do so, we will have to choose the resource again, go back into automated planning, optimize, choose our horizon, 27th of March to the 20th of April and set the start equal to the beginning of the horizon. Say, optimize, click on the wrench once again and this time we will once again set make span equal to zero zero. Now, total setup times is zero so is the cost and now we will put 100 for the delay cost and as we click the green tick down here in the bottom corner we will see that the diagram updates to only consider total delay cost. the optimizer so if we close this one and run the optimizer once again we are automatically taken to this page once again whenever the optimization run has finished. If we go back into the advanced scheduling board we now see that orders has been rescheduled once again and this time with respect to delay cost. So, summing up these two examples were both very extreme as we only considered one parameter at a time in the objective function. In this regard or however it is also possible to include multiple parameters at a time making the objective function reflect your specific business needs. All right. That was all we had for our demo sessions. Now we have seen how to create orders and how to schedule them using the optimizers and heuristics. Now we move on to some questions that we have sent to SAP and they have kindly responded. We will play the interview shortly and afterwards we will kick off the Q&A session and dive into some of the questions that you have placed in the chat. So remember to get your questions in if you have any burning questions left. Welcome Bernhard and Christian. I'm very happy to have you guys here and thanks for taking the time to join our exciting webinar. Maybe you guys could spend just a couple of minutes introducing yourself and also talking a bit about the roles that you guys have at SAP. Yeah. Thank you, Sarah. Happy to do so. So my name is Bernhard Trebels. I have been with SAP for a longer time. Started back in 99 in development as a developer then spent many years in consulting and pre-sales before I joined the product management some years back. Always in the area of supply chain planning and manufacturing. Yeah, in my current role, I lead the product management team for the S4 HANA planning and manufacturing product managers. So this comprises topics like PP, QM, PPDS, and PEO. Christian, over to you. Thank you, Bernhard. Then I continue. So hello, my name is Christian Halt. I'm, yeah, product manager for production planning and detailed scheduling. Yeah, I've also quite a long SAP history, so with SAP since 2001. Mainly also in consulting in different roles, but as Bernhard, always to the topic of supply chain planning and, yeah, now I'm co-responsible for evolution of our product, so production planning and detailed scheduling. So for sure we do roadmap definition and, yeah, at the end it's our goal to create business value for our customers. Yeah, thanks Christian and Bernhard for that. So we wanted to ask you, with PPDS now being embedded in S4HANA, do you see more customers exploring the benefits of PPDS compared to the days where you needed a separate APO solution? And for those considering PPDS, could you just put a few words around what problem it solves or where it brings value in the end-to-end supply chain planning maybe? Yeah, happy to do so. Thank you Sarah for the question. So these were a couple of questions. So let's take it one by one. Customer adoption, yeah. So we are very happy that we see our investments into PPDS well resonating in the market, yeah. So that compared to APO times where PPDS was a component of the APO that we see a significant rise in number of customers. So also new customers adopting PPDS who do not come out of our customer base of APO. We see that the strategic direction, the decision taken like 10 years back to embed PPDS into S4L works. one target was to simplify the integration of the ERP system with PPDS. And this for instance if you take examples it reflects in one master data right so that certain integration steps are much more simplified so you do not have to generate integration models but you can simply set a flag to integrate or we have one MRP planning across PP and PPDS. it's one technical system to administrate if you run it yourself. These are examples and customers appreciate that and for sure they also appreciate the significant investments which we have done in the modernization of the PPDS application in the different dimensions. I'll come to that in a minute or actually I can continue there right away. Take that over because this was also your second question. I mean what's different compared to APO right so we set a strong focus on business innovations. So our customers business processes continues to develop further and we want to then also reflect them adequately in our system. May it be such concepts like demand driven MRP or may it be time dependent stock levels while in the past it oftentimes was just fixed values right. yeah it reflects apart from process innovations also in the modernization so nowadays user experience is defined basically on the basis of fury right. we also see architectural modernization like data extraction and stuff like that which then also CDS use which then allow analytics and SAC right. So these are examples for modernization. Kristen will later touch the point of AI also a area of modernization with high customer interest and then last but not least we invest strongly into integration right. the world around us changes quite a bit and we need to see that we are well connected to our neighbor applications to contribute our part to the end-to-end supply chain planning solution of SAP. In October we've seen the release of 2025 FPS Zero which was a featured packed release and that clearly shows that the capabilities of S4 HANA PBDS are growing very fast. With the Q4 2026 release having lots of new features planned can you tell us a bit more about the release cycles? Yes, yes, sure. I mean, basically that's correct. We are focusing on FPS Zero and FPS 2 releases which are coming once a year basically but in general we are following the feature pack cycle of S4 HANA which means we are also considering feature pack 1 and feature pack 3. So this allows us to bring innovations or come out with innovations more regularly and it also helps our customers to upgrade, get new features more regularly without doing regular upgrade to a new release. And yeah, we are also coming up with important new innovations with the irregular feature pack so to say. So for example in feature pack 1 we intend to come out with the first version of the embedded analytics a key milestone I would say in our journey and then this will allow the planners then to really monitor KPIs while they work in the system, while they do interactive planning. So overall answering your question, yes we are following the annual cycle but we also come up with very important innovations in the feature pack 1 and feature pack 3. So multiple times a year. Okay. And which features from the October release are you most proud of and what are you most excited to see in the future release? You just already mentioned one but is that also the one you're most excited about? Yeah, I mean it's quite difficult to tell you which are the most excited ones but I mean in general I think you are also talking or you have talked a lot about PPO and I think that's one of our key innovation also for release 2025. So we are enabling the consideration of flexible constraints and in the past we had also a lot of innovations around consideration of calendars, resource, networks and earlier also purchase with and without source of supply quota arrangements and also MRP areas. This is one area of innovation but I mean our range is much broader I would say. So we have also our focus topics like UX and usability and there I would like to mention especially the advanced scheduling board which enables now our color configuration. We also support the tank planning topic with the combined fill level chart and we are coming up with quite a lot of monitoring apps to help the planner to really monitor the planning situation, monitor the system. So for example monitor product status and monitor receipts. and I mean last but not least I would also like to mention innovation areas like AI in which we have quite some content on basic stuff like transactional navigational but then also assistance to help the planner to understand and work within the system. So that means a lot of different areas and yeah I cannot tell you which is the most excited one. Lots of great stuff Christian. Yeah I agree. Everybody is talking about AI and how AI can reduce the workload and support our supply chains of the future. Can you tell us a bit about where does AI have a role to play in the PBDS space maybe? Yes yes sure happy happy to do so and I mean I can tell you that AI is relevant across the overall planning cycle supported by production planning and detailed scheduling. Maybe let me touch first of all on the PPO the production planning optimizer on which we also plan for an explanation component an explanation assistant similar that we already have it for the detailed scheduling optimizer so an assistant which helps the planner to understand the quite complex planning results of the optimizer and also helps to understand issues which could not actually be resolved by the optimizer engine and we provide root cause analysis options multi-level analysis to really point quickly point the planner to the root causes of an issue which can be lying two three five levels down the bomb structure right and it's usually very hard and time consuming for the planner to get there and to understand the root cause so this is an area in which we are heavily investing and also heavily evolving I would say as that PPO explanation component is around the corner but we are also having already some quite basic stuff what we call navigational and transactional so the dual assistant is able to answer questions using our monitoring app so for example please show me all the products within alert or please show me all overload situations on a specific resource so this can quickly be answered by the dual assistant and at the end also helps the planner to navigate in the system and to quickly get to the most important issues on the current planning situation looking into the future we plan to come up with the planning assistant to more actively also support the production planner in the daily work we want to start with one use case on the resolution of overload situations capacity overload situations but then for sure we also think about other use cases like handling of component material shortages which is also occurring quite regularly and yeah I mean looking forward last but not least we also plan to help the planner in to some more long-term planning we plan we plan for a simulation assistance to really help the planner to define and create simulation scenarios simulation environments and to do simulations actively there and this will then so to say round the AI picture which is then supporting the overall planning cycle from the nightly planning run over the analysis interactive planning simulation to the execution so quite a lot of interesting stuff ahead of us I can say and yeah we already started our journey with quite some interesting use cases which are already there that's wonderful thank you for that answer very interesting topic yeah then Bernhard maybe the next question to you in the latest release we've seen developments with flexible constraints which can be considered by the PPO what use case do you see where IVP and PBDS work together considering both IVP and PBDS can be used for production planning and when does it make orders versus defining the guard rails for PBDS to create them basically we are very happy here as the product team for PBDS that since about two years ago we offer two integration patterns between IVP and PPDS so you might remember that until two years ago there was just the order based integration so IVP creating orders and these integrations being replicated one-to-one or integrated into PPDS where then just scheduling mostly was done sequence optimization this kind of things and since two years ago we can also integrate guardrails from IBP to PBDS that means not integrating individual orders but quantities per time bucket maybe day or week or months and this is a more loosely coupled integration which also brings its advantages both sides both ways of integrating have their justification their advantages and it is certainly up to the use case of the customers which one is the most best applicable one yeah so when would you rather look into the order integration well if the supply planner or in classic APO times the SNP planner has a strong say in how things are being supplied if there's not too much of detailed planning required in a factory in a plant right so then you can already make a pretty concrete order based plan in SNP which is just to be scheduled time wise and within the day in PPDS and that's it on the other hand side it comes with let's say certain challenges which we believe to where we are better able to address it with the so-called key figure based integration which follows more the hierarchical or cascading planning approach where IBP looks at the rough cut planning size this quantity in this time bucket say week and then PPDS can break down the production lot for that week into daily production lots or even shift wise production lots because for the huge week order you never find a slot on any of the production lines to squeeze it in right you might have to break it down into multiple production lots and then place one production lot on the one production line and another production lot on the other production line right so in the end then this reflects more the mindset of the supply network IBP being so to speak the customer requesting supply from a production plant and the production plant just has to make sure that it delivers in time and quantity but how it breaks down the individual quantities that's up to the factory responsible right and that is oftentimes what we also have seen at customers right that production knows it best how it has to manage its plants they just have to make sure that as the supply chain requests quantities at times that they provide these supplies right so in a way separation of responsibilities separation of concerns yeah so and this is what I really like about this second approach yeah and overall there is no right or wrong about those approaches we want to offer possibilities to our customers so that they can find the best possible way for them how to model their requirements in the system so that brings us to the almost end of this webinar we hope you learned something useful today we're now ready to answer some of the many questions you shared in the chat so thank you for sending them all I know my colleagues online have worked very hard to answer all of them we might not cover all of your questions but we'll make sure to respond to them after the session now I would like to invite my colleagues Thomas and Mess to join me in for answering some of your questions all right when installing the optimizer which approach does I am usually take and I think this one is probably going to Thomas okay thank you and if I should touch a little bit upon this now it does say installing and of course that is a very technical part about putting it on a server and making the connection but I don't think that's necessarily exactly what this question was about it's more probably about how we go about with implementing the optimizer and what to take into consideration and I think I think we should go small so of course there could be a lot of constraints focusing on and a lot of different costs to focus on so I think a little bit similar to what Julian was showing in the demo start with kind of making sure that we have penalties for not delivering and for delay delivery kind of that's the penalty cost that controls the solution and then working with cost for inventory that's typically also one that we will try and then of course differences between the different types of demand elements and then building that in a small scenario and saying okay then we can kind of understand so if we both have sales orders and forecast we should have that the optimizer is prioritizing the sales orders for example and then that gives you some kind of idea about how it works because if we include everything about production capacity and storage cost and shelf life cost and all that then it's very hard to kind of understand the results so I think we have some ideas and also how to make the cost model because as Julian was also saying is that we don't necessarily we don't use the real cost so it's not the what does it cost to have one product in stock a day for example it's more relatively costs and we have some we've predefined some ideas on how to do this so we kind of balance the inventory cost with the non-delivery cost so I think the question here would be start with a simple scenario and then build on top of that I think I hope that was an answer to this or else we can do it we can talk later I think that was great thank you Thomas when does it make sense to use the optimizer instead of a heuristic mess would you like to take this yeah I will this is a very good question thank you and I think there's actually a lot of considerations to be put into this I guess to give an example you could say that if you have a system landscape where IPP is responsible for the for the PP part then you can just select the the DS optimizer to schedule the orders and the PP optimizer won't be required here but as we have also covered during this this webinar there are some complexity related to applying these these optimizers so I don't repeat everything we went through but I would say that when we're dealing with these optimizers there's a lot of dependencies first of all on your system health meaning that as some of the stuff Thomas described we really have to make sure that our transactional data and master data is in place and then beside this which may be even more important is the maturity because although the optimizer will come up with some results there needs to be some control for the users meaning that you have to be ensured that you have the right capabilities in terms of maturity so I would say summing up in scenarios or situations where you have your data in place and you have the right capabilities I believe it would be a good solution potentially to use the optimizer I hope this answers the question great thank you is it possible to make your own heuristic and here the short answer is yes there is no change from APO to S4HANA it's the same but usually SAP has their standard heuristics that you will normally go for but if they don't match your standard requirements then you can make your own heuristic and customize it but of course this also needs to be maintained please elaborate on the statement you need a reliable pegging and I think this was a comment to Thomas' section earlier yeah a little bit broad I think it was related to when we talked about the different scheduling heuristics there was something about the top down bottom up heuristic where you I think it was saying that you need to have reliable pegging here and I think it could be a little bit technical but I think I would mention something about pegging so I think the idea is that the PPDS is kind of having this pegging engine that ensures that everything is kind of connected so if I have a sales order demand and that is covered by a planned orders then those two are called pegged so they are linked together and the system automatically opens and changes these pegging depending on demand and the supply elements in the system and of course if that's in place then it can and that goes throughout the entire bill of material also so we could actually from the raw material purchase order see what planned order this goes into what planned on semi-finished goods what that goes into finished goods and what that goes into demand and I think of course we need reliable pegging to make sure that if I then move my purchase order then I make sure that I'm also moving my plan orders and maybe getting an alert or a hard stop saying I cannot move the plan order more because then I will violate the due date of the sales order so I think reliable pegging is like this but of course maybe it goes back to the housekeeping itself also because if I have something in stock that is not correct then of course the system will think yes that can be pegged to a certain demand and thereby it will not generate new planned orders for this so I think the reliable pegging is to maybe goes a little bit back to the housekeeping the top down heuristic will take the pegging into consideration I hope that was what else please just write in the chat then we can try to elaborate it to maybe one-on-one at a certain point because it is a little bit technical thank you great and we have one more question now any recommendations about moving from the classical DS board to advanced planning board Thomas yeah I can take that as well and a good question I think we hear it we heard a lot this and we have heard it a lot and actually Christian from SAP also talked about it in the question saying that SAP is actually developing a lot in the advanced FURI scheduling board so now we have the color coding as well and we have this multi-split screen for tank planning but I think we get the questions a lot because if you are coming from APO and had the GUI advanced platinum board you had all the opportunities to do it exactly like like to have it with colors and everything else while the FURI was much more standardized and strict we see now that some of these things are being developed into the FURI so my short question to this is FURI will be the end game that have to go to but in some cases we still recommend that do something on GUI but then if do something on GUI don't make it too different from what the end game would be and that's what we typically do when we do projects and still work in GUI we try to configure and customize it so it's as close to what the thing in FURI will end with so for example there's some predefined ways how a fixed order is shown I think should adapt to that when doing something in the GUI as well but I think it's a big question and I think I'm a consultant so I can say it depends and of course it does where are thank great I think that was all for our Q&A session we will answer any questions that we have in the chat later on and then we move on now to the end wrap up we would like to leave with the remark that we regularly post articles and blog posts and other relevant SAP related content on LinkedIn and on our website so feel free to add us on LinkedIn to not miss out on any of our future updates on SAP also as previous mentioned for this webinar all relevant documentation including recordings slide presentations will be made available to after the webinar and then I just like to say thank for me as well thank so much