 Hello everyone, my name is Icaro Bustinho and I'm a PhD student from Federal University of Santa Catarina, Brazil. And today I represent the research online schedule and approach for Distributed-Archive Manufacturing Systems. This research is part of the FASTIN project, flexible and autonomous manufacturing systems for customer design and products that product receiving funds for European Commission Horizon 2020. So, just to have a brief introduction on Distributed Manufacturing Systems represents a new product in the industrial context. So, in the past, nowadays, we have globalized mass production and we move the production to low-cost countries and do the production on a higher scale. And in this case, we have long delivery times along with the global supply chains. This increases the logistics and CO2 emissions. So, now the idea is to have Distributed Production in small production centers close to the final client. And we can do that to achieve manufacturing on an undermined small scale close to the final client. And this kind of system is very complex to control. So, the idea of this research is proposed and implemented in R, of course, a new conceptual model for dynamic allocation of production orders in these DM systems considering data provided by Internet of Think Technologies. And the proposed approach will be tested in a user case using real data. And the idea is to support a first-come pilot application. So, that's the proposed model. The idea is to decide where each production model will be executed and then for that, you consider four main variables, setup times, skill times, production times and transportation times. And you need to, for each PO, minimize this objective function that is the sum of these variables. So, for each PO request, we use an external application called Open Street Maps. It's an open source API. We can use this API to have a real-time transportation time between multiple plays. And then for skill time, production times and setup times, we get this information from IoT platform. For production times, we constitute the historical time series of prediction. We estimate a set of forecasted models, then do the forecasts and select the best one. And with all this information together, we can apply a search algorithm to find the best option combining these data and then minimize the objective function and do the better allocation. So, about the model implementation, this is the basic architecture. We use some of our libraries as deployed as online, ATTR, forecast and plumber. And then we build a REST API using HTTP protocol and the Swagger framework. And about the implementation and integration, this model has a graphical interface, a web graphical interface, so the user can put a PO into the optimization tool. And then use the information of this PO, we can consult this external application and then calculate the transportation time between the defined delivery place and all the production centers, get the historical data and real-time data for the IoT platform, and then run our model that is a Swagger API and then give back the information to the system, notify the users and control the production order location. So, we did a simulation-based experiment to evaluate our model and then we simulated two scenarios, each one with 90 simulation runs and for each scenario, you consider 1000 of POs. So, we have efficient gains with the reduction of 77% of the average waiting time of the system. So, as you can see in the graphs on the left, you have the allocation by distance, which means we always allocate the production in the closest place and then the allocation of our model that considers real-time data and the estimated production time using forecast models. So, you have a big difference between these two approaches. So, as conclusion, the model was able to deal with fisticastic characteristics in a dynamic environment, considering internal and external information. The model was implemented in R, that is an open-source language and can be easily applied in distributed manufacturing contests considering real-time data provided by IoT technology. In the future, the idea is to have reactive scheduling and adaptive capabilities into the model and incorporate these capabilities and evaluate this model in a more complex distributed network and then support a 4th-con pilot application. So, thanks. That's my mail. You can send me a message if you have any doubts or suggestions.