ranvar.periodicr

ranvar.periodicr(T.a ..) 🡒 T.r : ranvar, function

Converts time-series into ranvars by collecting observations over moving windows. Generalizes ranvar.segment with an horizon expressed as a ranvar and events at the extrema or at the middle of the chosen time interval are equally taken into account:

table Items[id] = with // 'id' is the primary dimension of 'Items'
  [| as id, as Start, as End, as LeadTime |]
  [| "A",  date(2023,1,3),  date(2023,1,14), 3 |]
  [| "B",  date(2023,1,10), date(2023,1,18), 2 |]
  [| "C",  date(2023,1,20), date(2023,1,25), 4 |]

table Orders = with
  [| as MyId, as MyDate, as Quantity |]
  [| "A",  date(2023,1,5), 5 |]
  [| "A",  date(2023,1,11), 2 |]
  [| "B",  date(2023,1,20), 1 |]

table Censored = with
  [| as MyId, as StockOutDate |]
  [| "B",  date(2023, 1, 3) |]
  [| "B", date(2023, 1, 5) |]

expect Orders.id = Orders.MyId // add 'id' as secondary dimension of 'Orders'
expect Censored.id = Censored.MyId // idem, for 'Censored'

Orders.D = ranvar.periodicr( // beware, it's not 'periodic' but 'periodicr'
  start: Items.Start,
  end: Items.End,
  horizon: dirac(Items.LeadTime),
  censoredDemandDate: Censored.StockOutDate, // optional
  date: Orders.MyDate,
  quantity: Orders.Quantity)

where id == "A" // 1 event displayed
  show table "Periodic Demand" a1d4 with
    id
    Orders.MyDate
    Orders.D

Indeed, ranvar.periodicr considers an infinite repetition of the input data and sums the event quantities over periods of all possible lengths contained in the horizon ranvar. As a consequence, if we consider the example above and we imagine an additional event of quantity 2 on Jan 7th, ranvar.periodicr with dirac(3) as the horizon would observe the following:

• Jan 1st - Jan 3rd: Q = 5
• Jan 2nd - Jan 4th: Q = 5
• Jan 3rd - Jan 5th: Q = 0
• Jan 4th - Jan 6th: Q = 0
• Jan 5th - Jan 7th: Q = 2
• Jan 6th - Jan 1st: Q = 2
• Jan 7th - Jan 2nd: Q = 7

and returns the ranvar ~28% Q = 0, ~28% Q = 2, ~28% Q = 5, ~14% Q = 7. For comparison, ranvar.segment with horizon 3 and step 1 would ignore the last two lines of the above list and return 40% Q = 0, 20% Q = 2, 40% Q = 5.