After defining inventory as the number of flow units in the system, flow rate the number of flow units flowing through the system per unit of time, and flow time the time it takes a flow unit from entering the system to leaving the system. The purpose of this session is to find out what are the drivers behind these three variables. In this session, we will introduce a concept of a bottleneck, which is probably the most important definition in this course. We'll also talk about process flow diagrams, which are really maps describing how the flow units goes from being an input into the process, from leaving the process as a finished unit of output. To see these concepts in action I would like you to join me once again and let's take a walk together to our local Subway restaurant and see how the operations look inside the restaurant. Now time to go inside the restaurant. From some training materials that Subway kindly shared with me, I now know that there are a couple of tasks that need to carried out in order to make a sandwich. In this picture here you see how there are three workers, three stations that are making Subway sandwiches in the store at this hour. The work starts at station one with greeting the customer. That takes four seconds followed by five seconds to take the order. If we add up the activities that are carried out by the first station, we get a total time of 4 plus 5 plus 4 plus 3 plus 12 plus 9, which is equal to 37 seconds per customer. We refer to this time as the processing time. Notice that processing times are always expressed in seconds per customers or more generally in units of time per floor unit. In the same way I can compute that it is taking 46 seconds per customer at station two. And 37 seconds per customer at station three. Again, those numbers are called the activity times. Throughout this course, I will use the words activity times and processing times interchangeably, so pardon the ambiguity when I just said activity times. I really was thinking about processing times. So be aware of that in the future. Activity times and processing times are one in the same thing. I want to be a little bit of hand waving here around the toasting activity, keeping in mind the toasting doesn't require the direct hands-on work by a worker but instead is automated through a toaster. If there should be a limited amount of toasting capacity, that is indeed something that we would consider. But for nor, be hand-waving over this matter. Now this is pretty useful because now we know there are three workers carrying out the work. We know all the things that need to be carried out to make a sandwich. And we know that the total time it takes to make a sandwich is 37 plus 46 seconds plus 37 seconds. Now let me propose a slightly different representation of the same data. This way we can draw a little picture that captures the flow of a flow unit through the process here at Subway. This picture is called a Process Flow Diagram. Our experience with Subway starts with a waiting line and in general we will refer with triangles as pictures capturing wherever there's waiting lines or inventory. Boxes will capture activities and so we know that the processing time, at the first box, there's going to be 37 seconds per unit. The work is then handed off to the next worker, which might potentially include some delay, that's why I draw a little triangle here, where there are another 46 seconds of work. From there, potentially another little waiting, before you hit the checkout. There is a third station where, again, we're going to have 37 seconds per unit. Finally, the customer is served and can happily leave the restaurant. Now, we've drawn this process flow diagram here, in PowerPoint to get rid of my ugly handwriting, just to iterate the symbols, the triangles, stands for flow units waiting, arrows capture the flow of the flow units and boxes capture the resources that are carrying out the activities in the process flow. Now let me at this point refer you to an important difference between project management and process management. Process management, the topic of this course is all about doing things repeatedly. You want to serve hundreds and hundreds of customers over the day, and at this point are primarily interested in computing the flow of customers through the process. It turns out that for the flow of customers through the process, it doesn't matter whether we work stations one, two, or three sequentially or in parallel. If all you have to do is make one sandwich, it would be lovely to, for example, ring out the customer at station 3, in parallel to actually making the sandwich at station 1 and 2. But, at the end of the day, every customer here in the flow has to go through station 1, 2, and 3. And so, we are not going to serve anymore customers by working in parallel. Now, we are ready for some definitions. We've already seen the concept of the processing time which captures how long a resource takes to serve a flow unit. For example, station 2 in our previous setting had a processing time of 46 seconds per customer. Next, we define the capacity of the resource as one over the processing time. In our case, one over 46, and now careful with the units, that is customers per second. Now in casual English you would typically not say that a worker has a capacity of serving a 46th of a customer per second. If you want to make this something easier to imagine, just multiply this with the 3,600 seconds that there are in an hour, and use see that the worker here at station two, is able to serve roughly some 78 customers per hour. Now this is a case which we have just one person working at the resource. If there are multiple person or multiple machines carrying out the same work, we define the capacity as m, the number of parallel resources and divide it by the processing time. Now the chain is only as strong as its weakest link, and if we ask our self how much capacity the entire restaurant has, we're going to look for the capacity of each individual step, and we will then pick the lowest capacity. This is the idea behind the concept of a bottleneck. The bottleneck is the step with the lowest capacity. Next we want to figure out the flow rate. We already defined the flow rate is the number of customers going through the process per unit of time. But it can never by definition be more flow through the process then we have capacity at the bottle neck. However, there might be a situation when even the bottle neck has some excess capacity. Those are situations where we have insufficient demand. In off hours, late at night, early morning, we might just not have the demand rate to keep the bottleneck busy. And so the flow rate is defined as the minimum between demand and process capacity. We can then compute the utilization of a resource. It's the ratio between the flow rate and the capacity. Now, remember the flow rate captures the flow, meaning it captures how much work a resource is currently doing, versus the capacity, which really measures how much work the resource could be doing if it worked all out. Just to reiterate, we already had defined in a previous session the flow time, the time it takes a flow unit to go through the process. And the inventory, the number of flow units in the system. Now let's jump to Excel and practice our new definitions. Let's start with the processing times. Station 1 has a processing time of 37 seconds, Station 2 of 46, Station 3 of 37 seconds, and all of this is expressed in seconds per unit. We then saw that the capacity is defined as one over the processing time. And this is now expressed in units per second. If you want to get to the capacity per hour. We simply multiply the previous number with 3600. And this is now expressed in units per hour. Next we define the process capacity as the minimum of these capacities above. Which in this case is driven here by 78.26. Now the flow rate is the minimum between the demand and capacity. Let's say for sake of argument that we have a demand through the process of 50 units per hour. With 50 customers per hour we can than compute a utilization of, remember the definition, we are dividing the flow rate, in this case that would be demand, flow rate divided by the capacity. Which gives us a utilization of 51% at the first station. You notice that the utilization is higher at the second station, the bottleneck, with 63%. And then, again, it's 51% at the last station. If there is more demand coming our way, so if we improve, here, the demand rate from 50 to 60, the utilization goes up. Notice that, according to our definition, though, the utilization can never exceed 100%. Because the flow rate is the minimum between demand and capacity. In this session I threw a lot of vocabulary. We saw how resources and workers have processing times, how we can use the processing times to compute capacity levels, and how the resource with the lowest capacity in the process is called the bottleneck. We also introduced the measure of utilization. Together these calculations help us to determine the flow rate of the process without actually observing the process in action. In the last sessions we just sat there at Subway, and counted the customers to compute the customers served per hour. Now we're actually able to just predict the flow rate by simply knowing the processing times, the staffing level, and the demand rate.
