The problem is about forcasting the drugs distributed on 30 stores, further predicting the time for each store high need either in normal states or epidemic periods such as COVID19