So, as we mentioned, this problem is motivated by observing those huge amount of costs. So we are actually operating 12 facilities currently, but we have no idea whether that's the most efficient solution or not. In practice, if you are newly hired as a CEO, you may check with your vice presidents or some people. So there are some people that they work there for many years you ask them, well, what's the point of having twelve facilities in Taiwan? They may told you that this has been there for 10 years, 20 years we are always doing that way. That may not be a very good answer even though it appears so frequently in practice. People are used to some settings, some old settings, but that may not be good. Especially that may not be most efficient because the situation is always changing. Maybe 20 years ago, those office rents are very cheap, but today they are quite expensive. There's no reason to keep using the same setting. So as we mentioned, maybe you want to cut down some facilities, that's going to create a lower operating cost like office rents. But if you have more facilities, that's going to give you a low service cost because the traveling distance in average would be shorter. So the optimal solution really depends on the real situations, depends on the distance, depends on the fuel cost, depends on rents and so on and so on. What make this problem difficult is that whether to shut down the facility must be determined by considering all facilities, customers, and the engineers. For example if you feel that, hey, the Shenzhou facility is quite expensive I don't want to have the Shenzou facility, well, that's fine. But once you decide to cut down the Shenzou facility, you need to re allocate, reassign all those customers that were assigned to Shenzou need to reassign them to somewhere else. And that creates some other costs, okay? You also need to reallocate those engineers at Shenzou, somehow that creates other constraints other costs. So you need to just consider all these facilities, all the customers, all the engineers at once. And typically that problem is not so easy to be answered by intuition. Plus, the environments are always changing, right? There may be more or fewer customers in the next year in the next, next year. The required frequency for visiting customers may be higher or lower. Maybe in the future we will have higher advanced technologies. The machines are much more durable and we don't need to visit customers so much. If that's the case, obviously you will try to cut down facilities. So somehow you know you need to keep doing these decisions. This year, next year, next, next year, you need to keep an eye on your facility location decision. Somehow you need to continuously get some suggestions. You don't want to in this year hire a group of consultants, pay them a lot of money, and then in the next year you do it again. You somehow want to answer the following questions maybe by yourself. Maybe you want to find a balance between the operating and the service costs. Maybe you should be able to equip yourself as being able to generate new solutions after environment changes. So somehow all in all, that is telling us that a model can be helpful. We may somehow take a look at the situation, build a conceptual or mathematical model to describe, to incorporate the interaction between the operating cost and the service cost. And then once we input the data for the current situation is going to tell us what to do regarding locations, regarding customers and then regarding engineers. For the next year or next, next year, when we have a chance to modify our decisions, then we input the data for the next time. And then we do it again, and then we get a new suggestion. Once the environment changes, we input the new data and get new solutions, that's more hope. So let's make it more concrete. The objectives for this research is that we want to build a mathematical model to formulate the facility location problem. Okay, that's the heart for operations research, a mathematical model. That model is going to precisely describe the problem and be the foundation for any heuristic algorithms. That model itself may generate an optimal solution. That's a suggestion for us. But also, if we have a lot of customers a lot of facility locations, the model may be too complicated to be wrong and to get an optimal solution. Then we need a heuristic algorithm. So in this project, we also design a heuristic algorithm and we'll show you the concept. Once we have this tool, we aim to do two things. First is that we want to take a look at the current environment, which is different from 20 years ago, and then we want to do optimization for that. So we will really take a look at all those customers we have at this moment. All those engineers we have at this moment, and all those 12 facilities we have at this moment. We will ask ourselves, should we keep all those facilities? If not, which to shut down? And once we do that, how to reallocate engineers and then reassign customers? So that's some optimization thing we want to do for this moment, but we actually want more. We want that the optimization may be carried for this year, next year, next, next year and actually be carried for possible new business chances. So understanding from time to time you may see there are some new potential customers. For example, suppose currently you are serving 711. Maybe in the next year, Family Mart is going to try to find someone to help him. So if that's the case, you are going to try to acquire new customers. But getting new customers does not always mean profitable. The customer is going to pay you something, but then you need to pay additional cost to see whether you are going to be beneficial. Okay? You're going to earn some money, but that takes some cost. You need to do some evaluation to check whether we are really earning some money. So if you have this model then before you really get the customer, you are able to see whether any adjustments about engineer allocation or customer assignments is going to help you best balance everything you have. And then you will be able to estimate the cost for getting this new customer. Then you will be able to see how strong it is for you to try to win these customers. How should you try to face this customer? Do you really want to get this new customer or you should let go? You really need to have a nice way to estimate the cost for serving this new customer. And again this model is going to help you. So all in all, I hope you are convinced that we really need a model for this particular problem. If you don't have a model, then you need to really spend a lot of efforts to get one answer for current setting, but we don't want that. We want a model which is a solution. That model is going to help us generate the optimal setting, generate decisions suggestions for any possible new settings. That's why we want a model.