Tutorial Three - The Newsvendor Model PDF

Preview

Summary

Week 8 - The Newsvendor Model...

Description

Week 8 – The Newsvendor Model -

-

-

The Ford Model T: Standardisation “any customer can have a car painted in any colour he wants as long as it is black” One product, produced by one company locally, had a very long life cycle (~15 years) Today, business is a global, team sport – Apple iPhone supply chain

Risk Information Risk: Decisions made with poor information Alignment Risk: Decisions made with self-interest, as opposed to group’s interest.

-

Managing Risk: The Newsvendor Model Order newspapers overnight to sell tomorrow in the face of uncertain demand. Sunny -> High Demand Rainy -> Low Demand Key features of model – Place a bet in the face of uncertainty, no recall to your decision, the product is perishable.

-

Capacity Planning Example – Executive Committee at British Petroleum Processing facilities are needed for a new oil field Maximum sustainable output from the oil-field ranges from 40,000 barrels per day to 90,000 barrels per day. BP will know potential output only after production begins These capital investments are measured in billions of $ How big should the processing facility be??

-

Quantify the Uncertainty: An uncertain outcome or a random quantity (also called stochastic variable) is a quantity that takes several possible values which are outcomes of a random experiment. Uncertain outcomes and quantities have a likelihood to happen Eg. “it is 60% likely to rain today”

-

Quantity the Uncertainty – Forecast: Weather forecasts, demand forecasts, economic forecasts A forecast is a distribution, not a single number Forecasts have an accuracy associated with them Forecast accuracy cannot be judged by one outcome A distribution captures all these features.

-

Dealing with Uncertainty: Given a forecast (three stocks, say), which stock would you invest in?

-

-

-

-

-

-

-

-

The human mind is particularly bad at understanding something uncertain; most people will simply choose the most likely outcome. This could be wrong. In the face of uncertainty, you do not just choose what is most likely to happen. EG: Doctor’s decision – how much of Justin Trudeau’s blood to carry while he is on a trip Most likely outcome is that no blood is needed (99.9999%), unlikely that a Litre of blood will be needed (0.0001%) How much blood do we think they carry? 10L If blood is not needed, there is no issue. If it is needed and they do not have it, he may die. The weight of the risk here is what matters. EG 2 – Bookstore’s decision – How many copies of one specific book to keep? Most likely that no one person will look for specific book Will probably stock none, as it is unlikely and if it were to happen, the outcome is not that bad. In the face of uncertainty, you tilt your bet AWAY from the most likely value in the direction where the consequences are less severe. You don’t simply want to be correct “most often”. You want to “wrong in the right way most often”. (When consequences are least severe)

Buying T Shirts for a theme party (EG) Random demand described by a Normal distribution with mean 300, and standard deviation 100. Economic parameters: Buy shirts at c = £7 Sell shirts at r = £10 Salvage shirts at s = £2 The “too much – too little problem”: Order too much and there is loss due to unsold shirts (salvage) Order too little and you lose potential sales (and profits) Co = Overage cost: The consequences of ordering one more unit than what you would have ordered had you known demand Suppose you had left-over inventory (you over ordered). Co is the increase in profit you would have enjoyed had you ordered one fewer unit For the theme party: Co = Cost – Salvage Value = C – S = 7 – 2 = 5 Cu = Underage cost: The consequences of ordering one fewer unit than what you would have ordered had you known demand Suppose you had lost sales (you under ordered). Cu is the increase in profit you would have enjoyed had you ordered one more unit. Cu = Price – Cost = R – C = 10 – 7 = 3 Balancing the Risk and Benefit of Ordering a Unit: Ordering one more unit increases the chance of overage Expected loss on the Qth Unit = Co . Pr(D /= Q) = Cu . (1 – Pr(D...