___________________________________________________________________________ Operations Research Jan Fábry Probabilistic Inventory Models
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Probabilistic Inventory Models How much to order? When to order? How much to store in safety stock?
___________________________________________________________________________ Operations Research Jan Fábry Model with Continuous Demand Probabilistic Inventory Models
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Probabilistic Model with Continuous Demand Assumptions Single item Probabilistic distribution of demand (stationary demand) Deterministic lead time (constant) Continuous (but not uniform) depletion of the inventory
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Probabilistic Model with Continuous Demand Assumptions Purchasing cost is independent of the OQ Unit holding cost is independent of the OQ No additional cost in case of shortage Replenishment - exactly on the point when the shipment arrives
Time Inventory Level 0 ___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Probabilistic Model with Continuous Demand d r d q Cycle ICycle II Placing Order Shortage
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Probabilistic Model with Continuous Demand Probability distribution of demand μQμQ Demand μ Q + σ Q μ Q – σ Q Mean of demand μ Q Standard deviation σ Q
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Brewery Probabilistic Model with Continuous Demand Estimation of annual demand = cases Standard deviation of annual demand = cases Annual holding cost per case = 20 CZK Ordering cost – transportation = CZK per order – other = CZK per order Lead time = ½ of month Objective: minimize total annual cost
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Brewery Probabilistic Model with Continuous Demand Mean of annual demand Standard deviation of annual demand μ Q = cases σ Q = cases Annual holding cost Ordering cost Lead time c 1 = 20 CZK per case c 2 = CZK per order d = 1/2 of month = 1/24 of year
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Brewery Optimum order quantity Probabilistic Model with Continuous Demand
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Brewery Mean of demand within the LEAD TIME = = Optimum reorder point Probabilistic Model with Continuous Demand Standard deviation of demand within the LEAD TIME
5 000 Lead-Time Demand ___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Probabilistic Model with Continuous Demand Example – Brewery Mean μ d = Standard deviation σ d = 500
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Brewery Deterministic model – planned shortages Probabilistic Model with Continuous Demand Probabilistic model – random occurance of shortages building of SAFETY STOCK
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Brewery Service level - definition Probabilistic Model with Continuous Demand 1. Service Level is the PROBABILITY with which DEMAND will be MET within the inventory cycle. 2. Service Level is the PROBABILITY with which SHORTAGE WILL NOT OCCUR within the inventory cycle. 3. Service Level is the PERCENTAGE of TIME that all DEMAND is MET.
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Brewery Implemented reorder point (for the given service level p) Probabilistic Model with Continuous Demand Optimum reorder point Safety stock level
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Probabilistic Model with Continuous Demand d Time Inventory Level 0 d r *
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Probabilistic Model with Continuous Demand w 0 Time Inventory Level d r * d r p
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Brewery Mean of total cost Probabilistic Model with Continuous Demand Holding cost of safety stock Objective: find such SAFETY STOCK level w that satisfies the given SERVICE LEVEL p and minimizes MEAN of TOTAL COST TC
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Brewery Determination of optimum SAFETY STOCK level Probabilistic Model with Continuous Demand SERVICE LEVEL Real LEAD-TIME DEMAND Implemented REORDER POINT
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Brewery Determination of optimum SAFETY STOCK level Probabilistic Model with Continuous Demand Real LEAD-TIME DEMAND Q d ~ N (r *, σ d ) ~ N (0, 1) Transformation
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Brewery Determination of optimum SAFETY STOCK level Probabilistic Model with Continuous Demand
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Brewery Determination of optimum SAFETY STOCK level Probabilistic Model with Continuous Demand
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Brewery Optimum SAFETY STOCK level Probabilistic Model with Continuous Demand p = 0.95 p = 0.99
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Brewery Optimum mean of total annual cost Probabilistic Model with Continuous Demand p = 0.95 p = 0.99
___________________________________________________________________________ Operations Research Jan Fábry Single-Period Decision Model Probabilistic Inventory Models
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Single-Period Decision Model Assumptions Only one order in time period Probabilistic distribution of demand (continuous or discrete) End of time period - surplus - surplus - stockout - stockout penalty !!!
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Single-Period Decision Model Seasonal or perishable items Newspapers – „Newsboy problem“ Seasonal clothing Christmas trees Halloween pumpkins Bread Flowers Fruits
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Happyland Bakery department – optimize everyday order of rolls Purchase price = 1 CZK per roll Single-Period Decision Model Selling price = 2 CZK per roll Remaining rolls are changed into crumbs 20 rolls in 1 sack of crumbs Selling price of crumbs = 12 CZK per sack
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Happyland Daily demand – normal probabilistic distribution = rolls Single-Period Decision Model = 500 rolls Objective: determine optimum order quantity
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Happyland Single-Period Decision Model Real daily demand for rolls – Q Daily quantity of ordered rolls - q Evening Q < q Q > q Q = q
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Happyland Single-Period Decision Model Q < q Marginal loss per 1 roll ( q – Q ) rolls remain crumbs ML = purchase price – salvage value
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Happyland Single-Period Decision Model Q > q Marginal profit loss per 1 roll shortage of ( Q – q ) rolls MPL = selling price – purchase price Q = q No loss
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Happyland Single-Period Decision Model No stockout probability p Stockout probability (1 – p) Expected ML = p(ML) Expected MPL = (1-p)MPL
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Example – Happyland Single-Period Decision Model Optimum expected cost Probability with which no stockout occurs (optimum service level)
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models SERVICE LEVEL REAL DEMAND ORDER QUANTITY Determination of optimum order quantity Single-Period Decision Model Example – Happyland
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Determination of optimum order quantity Real demand Q ~ N ( , ) ~ N (0, 1) Transformation Single-Period Decision Model Example – Happyland
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Determination of optimum order quantity Single-Period Decision Model Example – Happyland
___________________________________________________________________________ Operations Research Jan Fábry Inventory Models Optimum order quantity Single-Period Decision Model Example – Happyland