actionrwd.reward

actionrwd.reward, function

def process actionrwd.reward(
    timeIndex:     number,
    baseline:      number,
    dispersion:    number,
    alpha:         number,
    stockOnHand:   number,
    leadTime?:     ranvar,
    stepOfReorder: number,
    stockOffHand?: zedfunc,
    arrivalTime?:  ranvar,
    stockOnOrder?: number,
    samples?:      number,
    seed?:         number,
    numLeadTime?:  number)  : (demand: ranvar, holdingTime: zedfunc)

Returns the uncovered demand demand within the responsibility window and the expected holding time holdingTime for the nth ordered unit, driven by the ISSM model defined by baseline, dispersion, and alpha.

• timeIndex: ordering of periods within each item.
• baseline: expected demand per period.
• dispersion: variance divided by mean for the demand model.
• alpha: update speed of the hidden level in the ISSM model.
• stockOnHand: stock available at order time.
• leadTime: lead time distribution; use with numLeadTime to shift deviates.
• stepOfReorder: interval between reorder opportunities, in periods.
• stockOffHand: optional downstream coverage (multi-echelon use case).
• arrivalTime: arrival time distribution for pending orders.
• stockOnOrder: quantity associated with arrivalTime (use both or neither).
• samples: number of simulated trajectories (default is 2500).
• seed: optional random seed for reproducibility.
• numLeadTime: fixed lead time; overrides leadTime when provided alone.

Examples

table Items = with
  [| as Id, as Dispersion, as Alpha, as StockOnHand, as LeadTime,
     as StepOfReorder |]
  [| "sku", 2.0, 0.2, 3, 1, 2 |]

table Periods = extend.range(6 into Items)
Periods.Baseline = 10

Items.Demand, Items.HoldingTime = actionrwd.reward(
  TimeIndex: Periods.N,
  Baseline: Periods.Baseline,
  Dispersion: Items.Dispersion,
  Alpha: Items.Alpha,
  StockOnHand: Items.StockOnHand,
  LeadTime: dirac(Items.LeadTime),
  StepOfReorder: Items.StepOfReorder)

Items.MeanDemand = mean(Items.Demand)
Items.HoldTime1 = valueAt(Items.HoldingTime, 1)

show table "Action reward" with
  Items.Id
  Items.MeanDemand
  Items.HoldTime1

This produces the following table:

Remarks

The demand output is a ranvar in the ordering space, while holdingTime reflects the expected number of periods the nth unit remains in stock. Use actionrwd.demand or actionrwd.segment for diagnostics of the same trajectory model.

See also

• How to compute the return of the next replenishment unit
• actionrwd.demand
• actionrwd.segment
• actionrwd.dampen