liveBook · Manning

Backorder

As explained in Applied Reinforcement Learning

A backorder is an unfulfilled customer demand recorded for later fulfillment because available inventory was insufficient.

A backorder records customer demand that available inventory cannot satisfy. In production planning, it captures the shortfall between demand and fulfilled units, and carries a penalty. section 2.5

Factory demand

If scooter or bike demand exceeds available inventory, the difference becomes a backorder. section 2.5

Listing 2.6 Transition Dynamic based on stochastic demand
def transition(state, action, demand, params: FactoryProblemDefinition):
    day, inv_scooter, inv_bike, current_setup = state #A
    demand_scooter, demand_bike = demand              #A
 
    produced_scooter = params.prod_capacity_scooter if action == 0 else 0 #B
    produced_bike = params.prod_capacity_bike if action == 1 else 0 #B
 
    inv_scooter += produced_scooter #B
    inv_bike += produced_bike #B
 
    satisfied_scooter = min(inv_scooter, demand_scooter) #C
    satisfied_bike = min(inv_bike, demand_bike) #C
 
    inv_scooter -= satisfied_scooter #D
    inv_bike -= satisfied_bike #D
 
    backorder_scooter = demand_scooter - satisfied_scooter #E
    backorder_bike = demand_bike - satisfied_bike #E
 
    next_setup = action if action in (0, 1) else current_setup #F
    next_day = day + 1 #G
    next_state = (next_day, inv_scooter, inv_bike, next_setup) #G
 
    return next_state, produced_scooter, produced_bike, backorder_scooter, backorder_bike

The transition listing calculates each product’s backorder as demand minus units satisfied from inventory. section 2.5

How Applied Reinforcement Learning covers it

How deep it goes
Calculates backorders as demand minus fulfilled units and applies a penalty, without detailing subsequent fulfillment.
Examples
Uses scooter and bike demand exceeding available inventory as factory backorder examples.
What you can do after
Understand how production decisions leave demand unmet and how the shortfall affects reward.
How it connects
Connects inventory availability and demand fulfillment with Markov decision processes, transition functions, and reward functions.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Applied Reinforcement Learning. The section links point to the parts of the book it draws on.