This action might not be possible to undo. Are you sure you want to continue?

1.155, 2.943, 3.577, 6.938, 10.816, 13.621, 16.862, 22.82, ESD.72

CBA 4.

Including Uncertainty

George E. Apostolakis Massachusetts Institute of Technology

Spring 2007

CBA 4. Including Uncertainty 1

Uncertainty

• Practically any CBA requires consideration of uncertainty. • Most methodologies in use are ad hoc, due to the intrinsic difficulty of the generalized problem.

CBA 4. Including Uncertainty

2

**Methods 1. Scenario analysis 2. Adjustments of interest rates 3. Decision Theory 4. Simplified probabilistic models
**

CBA 4. Including Uncertainty 3

CBA 4.Scenario Analysis • Preparation and analysis of scenarios: -.“Optimistic” or “most favorable estimate” -. Including Uncertainty 4 .“Pessimistic” or “least favorable estimate” • Interpretation is difficult without assignment of probabilities to scenarios. • Benefit: Brings additional information into the process.“Most likely” or “best estimate” or “fair estimate” -.

000 • What are we to do with such information? CBA 4. Annual number of units: 900 1.100 Savings per unit: $50 55 60 Operating costs: $2.200 etc. Fair Optim. PW: -$49.000 1. Including Uncertainty 5 .000 120. Pessim.Example • A new machine is to be purchased for producing units in a new manner.000 1.600 1.000 22.

Possibility of displacement bias. • We really don’t know how conservative (pessimistic) the final answer is. CBA 4. define the three values..Developing Scenarios • For each element in the problem. interest rate and costs. • People are bad processors of information. • Overconfidence. Including Uncertainty 6 . e.g. • Point estimates tend to cluster around the median value. • Extremes greater than the 75th or smaller than the 25th percentile are difficult to imagine.

“Assessment and propagation of model uncertainty.” Journal of the Royal Statist. • They used 10 leading econometric models under each of 12 scenarios embodying a variety of assumptions about inputs. 43 economists and energy experts forecast the price of oil from 1981 to 2020 to aid in policy planning. such as supply. B (1995) 45-97. Draper. Soc. Including Uncertainty 7 . and growth rates. • In 1980.. demand.Forecasting Oil Prices D. CBA 4.

” • The 10 models were applied to the plausible scenario. CBA 4. Including Uncertainty 8 .The Plausible Scenario • One scenario was termed as the “plausible median case.” It represented “the general trends to be expected. • Results for 1986: ¾ Actual price: $13 ¾ Range of predictions: $27 to $ 51.

000 (A/P.872 CBA 4.272 Annual operating cost $28.. Including Uncertainty 9 . • Example Total annual income: $55. 6yrs) = $30. 30%.000 – (30.000 Annual capital recovery with return: $80.000 Capital cost: $80.600 Net annual profit: 55.g. 30%. e.High Interest Rates • Justify choice of alternatives using a high interest rate.272 + 28.600) = = -$3.

• If the annual income were $60. Its choice is arbitrary.Example (cont’d) • The high rate of 30% is intended to cover uncertainty. then the net annual profit would be 60. CBA 4. • A high interest rate does not guarantee that all uncertainties are accounted for.128 and the venture would be accepted.000 .(30.600) = = $1.272 + 28. Including Uncertainty 10 .000.

Including Uncertainty 11 .Decision Theory: Manufacturing Example • Decision: To continue producing old product (O) or convert to a new product (N). The payoffs depend on the market conditions: s: strong market for the new product w: weak market for the new product CBA 4.

P(w2/w1) = 1.000/yr.8.2 = P(w1w2) P(s1) = P(s1s2) + P(s1w2) = 0.4 = P(s1s2) 0. new product and the market is strong. old product.000/yr. L2: $30.0 CBA 4. Including Uncertainty 12 . P(s2/s1) = 0.2.5. P(w1) = 0.000/yr.4 = P(s1w2) 0. new product and the market is weak • Demand and Probabilities: Period 2 (5 yrs) Period 1 (5 yrs) s1 s2 s1 w2 w1 w2 Probabilities 0.Manufacturing Example Payoffs · Earnings (payoffs): L1: $15. L3: -$10.

Including Uncertainty 13 .26K $96.61K CBA 4.09K O $121.Decision Tree Decision Options States of Nature P1 P2 s2 s1 w2 w2 PW of Payoffs $243.94K N w1 -$81.

5 ) = (1 + 0.04 .74 = 243 .04 ) (1 + 0 .26 K 10 PWw1 = −133.04 ) 5 − 1 1 = 30 x x = 109 .0. 0 .09K CBA 4.58 = −81.52 + 109 .04 ) 5 − 1 PWs1 = 30 x ( P / A .74 x 10 = − 36 . Including Uncertainty 14 PW w 2 = − 109 .52 K 5 0.04 ) PWs 2 = 30 x ( P / A .52 − 36.52x = −44. 0 .51 − 36.04 . 5 ) x ( P / F .04 x (1 + 0.04 .04 ) PWs1 s 2 = 133 .94K PWw1w 2 = −44.58 K 30 .04 x (1 + 0.Calculation of the Payoffs (1 + 0.74 K 5 5 0 .58 = 96.51K 30 PWs1w 2 = 133.5 ) = 30 x = 133 .

26x0.4 -81.61K PWO = 15 x ( P / A .04 ) New Product EMVN = 243.04 x (1 + 0.10 ) = 15 x 10 0.0.04 .09x0. Including Uncertainty 15 .94x0.4 + 96.86K Decision Stay with the old product? CBA 4.04 )10 − 1 = 121 .2 = 119.Calculation of the EMV Old Product (1 + 0.

2 = 0.637 New Product EU(N) = 0. U(-2)= 0] U(243.665. Including Uncertainty 16 . U(-81.590 Old Product EU(O) = 0.6368 Decision Stay with the old product? CBA 4.29 -2 ≤ x ≤ 2 (x in $M) [U(2) = 1.09) = 0.18 ln(x+5) .26) = 0.4 + 0. U(121.665x0.1.61) = 0.637.632x0. U(96.632.4 + 0.590x0.Calculation using Utilities Let the utility of payoffs be U(x) = 1.94) = 0.

• Note that (1) is of the form: Y = a0 X 0 + a1X 1 + a2 X 2 + .v..v. + aT X T CBA 4..Probabilistic Models • We have: XT X1 X2 PW[ X(T)] = X 0 + + + . Including Uncertainty 17 .. + 2 (1 + i ) (1 + i ) (1 + i )T (1) where X j ≡ B j − C j are the net benefits in year j. ⇒ PW[X(T)] is a r. • All Xj are r.’s...

CBA 4. • Try to compute the quantities E[PW]. the expected value of PW.e. Including Uncertainty 18 . and 2 σ PW i. the variance of PW.Analysis • Computing the probability density function (pdf) of PW is usually difficult in practice..

a and b constants) ⇒ ⇒ E[Z] = a E[W] + b 2 σZ 2 2 = a σW CBA 4. Including Uncertainty 19 .’s.v.Fundamental Relationships from Probability Theory (1) • Let Z = aW+b (Z and W are r.

CBA 4.’s.’s) E [ Z ] = E [W1 ] + E [W2 ] 2 2 2 σZ = σW + σ W2 1 Note: Extends to any number of mutually independent r. Including Uncertainty 20 .v. W2 : independent r.v.Fundamental Relationships from Probability Theory (2) • Let Z = W1 + W2 ⇒ ⇒ (W1 .

CBA 4. then Z is also normal. We often assume that Z is normal even if W1 and W2 are not.Fundamental Relationships from Probability Theory (3) • ⇒ Let Z = aW1 + bW2 + c E[Z] = aE[ W1 ] + bE[ W2 ] + c 2 2 2 2 σ2 = a σ + b σ W2 Z W1 • • If W1 and W2 are normal. Including Uncertainty 21 .

Example: Reliability Physics • (RPRA 3. slide 30) A capacitor is placed across a power source. CBA 4. Including Uncertainty 22 . The breakdown voltage of the capacitor is 130 volts. Assume that surge voltages occur on the line at a rate of one per month and they are normally distributed with a mean value of 100 volts and a standard deviation of 15 volts. • Suppose that the breakdown voltage is also normally distributed with standard deviation of 15 volts.

C.21 CBA 4.Example (2) • The capacitor fails when the surge voltage. σC = 15 volts • Define a new rv D ≡ C – S = aC + bS • Then. S. σS = 15 volts • C: rv with E[C] = 130. E[D] = 130 – 100 = 30 volts and σD = σC + σS = 21. Including Uncertainty 2 2 volts 23 . • S: rv with E[S] = 100. is greater than the capacity.

08 The uncertainty in the breakdown voltage increased the failure probability. shows that Pd/sv = conditional probability of damage given a surge voltage = P(surge voltage>130 volts/surge voltage) = P(Z > 130 − 100 ) = P(Z > 2) = 15 = 1 − P ( Z < 2 ) = 1 − 0 .08 • RPRA 3.41) = = P(Z > 1. Including Uncertainty 24 .Example (3) • D is also normally distributed. page 31.41) = 0.21)) = P(Z < . CBA 4. therefore • Pd/sv(D < 0) = P( Z < -(30/21.5 – 0.42 = 0.1. 0228 < 0 . 9772 = 0 .

..T) are mutually independent 2 r. CBA 4..Assuming Independence of Xj X1 X2 XT PWInd [ X(T)] = X 0 + + + . Including Uncertainty ...v.Ind = T ∑ (1 + i ) j j= 0 T j E[ X j ] ∑ (1 + i )2 j j= 0 25 σ2 j • Note: Gaussian approximation for pdf of Y may work well in this case... 1...’s with known E [ X j ] and σ X (= σ 2 j) E[PWInd { X(T)}] = E[YInd ] = 2 σ2 = σ PW ...Ind Y .. + 2 (1 + i ) (1 + i ) (1 + i )T • Xj ( j = 0.

00 Net Benefits t=1 t=2 t=3 $3K $4K $5K $6K $7K $3K $4K $5K $6K $7K $3K $4K $5K $6K $7K 26 CBA 4.Example: Project A T = 3 yrs. p 0.25 0. i = 8%.10 1. Including Uncertainty .10 0.30 0.. Initial Cost = $10K Probability.25 0.

do not have to be identically distributed or symmetric or discrete. and X3. these choices are made just to keep the example simple. Then: PWInd(A) = YInd = -10K + X1/(1. 2 and 3.08) + X2/(1. respectively. the net benefits of A in years 1.Example (2) • Denote with X1. X2. • Denote with Y the present worth of A. PW(A). and X3. X2.08)2 +X3/(1.08)3 CBA 4. Including Uncertainty 27 . X1. • Note: Net benefits.

CBA 4. Including Uncertainty 28 . and X3 that will actually materialize • Corresponding a priori probability of any value of Y is equal to probability of that particular combination of X1. X2 and X3. X2.Observations • Y is a random variable (takes more than one value with different probabilities for any given implementation of project A) • Value of Y will be determined by the values of the combination of X1.

σ X 1 ≈ $1. Including Uncertainty 29 . ) 2 = 0.25x4+0.25x6+0. we have E[X2] = $5K and E[X3] = $5K.1x(3-5)2 + 0. X2.3x(5-5)2 + 0.1x(7-5)2 = 1.3x5 +0. and X3. separately: E[X1] = 0.000 ≈ (1.000 (Similarly.140 = σ X 2 = σ X 3 CBA 4.Expectation and Variance of Annual Net Benefits • It is easy to determine the expected value and variance of each of X1.1x7 = $5.140)2 or.25x(4-5)2 + 0.25x(6-5)2 + σX 1 + 0.1x3+0.300.

X3 = 4) = P(X1 =3)·P(X2 = 6)·P(X3 =4) = = (0.00625.08) + 6K/(1. Y. the r. X2 = 6. with probability 0.. the PW of Project A. Including Uncertainty 30 . X3 = 4} is equal to P(X1 = 3.e. 2 and 3 are determined independently of one another. i.1)(0.25) = 0.v.08)2 + 4K/(1. will take on the value PWInd = YInd = -10K + 3K/(1. X2 = 6.14 CBA 4.08)3 ≈ $1.097.00625 that is. • This means the probability of the combination {X1 = 3.25)(0.Independence: Calculations • Assume that the net benefits obtained from Project A in years 1.

08)3]2 = [1/1.08 + 1/(1.5771) ≈ $2.08.08] σY X1 X2 X3 .140)2·[1/(1.140)2·(2.885 2 2 2 2+ σ 2 2]2+ σ σ [1/(1. Including Uncertainty 31 .700 CBA 4.2225)1/2 ≈ $1.08)2 + 1/(1.08)2+E[X3]/(1.Independence: Calculations (2) • Note that Y can take on a total of 125 (= 5·5·5) different values of independent outcomes in years 1.08)6] = = (1. 3) ≈ -10K + (5K)(2.Ind = (1.08)4 + + 1/(1. • Using the expressions on slide 21 we get E[YInd]= -10K+E[X1]/(1.08) [1/(1.Ind = (1.2225).140)·(2.08)3] = -10K+ (5K)·(P/A. σ Y . 2 and 3.08)3 = = -10K + (5K)·[1/1. or.08)+E[X2]/(1.08)2 + 1/(1. 0.

"on average.Conclusion • Project A will. CBA 4.700 around that average.885 and a standard deviation of approximately $1. Including Uncertainty 32 ." have a net present value equal to about $2.

What can we do with this information? Alternative 1 2 3 4 E[Y] = E[PW] 20 5 15 17 σY 15 7 12 16 CBA 4. Including Uncertainty 33 .

A1 and A4 may be indistinguishable. Depending on the uncertainties. • Assume normal distributions and find P(PW < 0). Including Uncertainty 34 . • Minimize the probability of loss. • Note how close A4 is and how large the standard deviations are.Possible Decision Criteria • Choose the alternative with the highest mean value (A1). CBA 4.

09 “best” alternative • A2: P(PW < 0) = P(Z < -(5/7)) = P(Z < -0.71) = = 0. Including Uncertainty 35 .25) = = 0.24 • A3: P(PW < 0) = P(Z < -(15/12)) = P(Z < -1.14 CBA 4.Probability of Loss • A1: P(PW < 0) = P(Z < -(20/15)) = P(Z < -1.10 • A4: P(PW < 0) = P(Z < -(17/16)) = P(Z < -1.33) = = 0.06) = = 0.

Probability of Loss for Project A • PInd(PWInd < 0) = P(Z < -(2885/1700)) = P(Z < -1.7) = 0.04 • The fundamental assumption is that of independence of the annual benefits. Including Uncertainty 36 . CBA 4.

25 0.30 0. for year 1 are known. Probability.10 0. X1. p 0. Including Uncertainty .Complete Dependence of Xj • Once the net benefits.10 1. we shall also know exactly the net benefits for years 2 and 3.00 Net Benefits t=1 t=2 t=3 $3K $4K $5K $6K $7K $3K $4K $5K $6K $7K $3K $4K $5K $6K $7K 37 CBA 4.25 0.

5771) ≈ $2.Mean and Variance (Dependence) ⎡ 1 1 1 ⎤ + + .885.140)2·[2. 8.Dep= (1.. Including Uncertainty . 3) ≈ $2.Dep X [(P/A. 8.08)2 +1/(1. as before or 2 2 = σ σ Y .140)·(2.. 3)]2 =(1..5771]2 σ Y . 3) • From slide 15 we get • E[YDep] = -10K + E[X]·(P/A.08) + 1/(1. 8. + PWDep [ X (T )] = X 0 + X ⎢ 2 T⎥ + ( i ) 1 ( i ) ( i ) + + 1 1 ⎣ ⎦ • PWDep (A) = YDep = -10K + X[1/(1.938 38 CBA 4.08)3] = -10K + X(P/A.

... + 2 (1 + i ) (1 + i ) (1 + i )T ⎡ 1 1 1 ⎤ PW + + .700). • The standard deviation in the dependent case ($2.16 • Compare with PInd(PW < 0) = 0. CBA 4.885). + Dep[ X(T)] = X0 + X⎢ 2 T⎥ ( 1 i ) + (1 + i) (1 + i) ⎦ ⎣ • In both the independent and dependent cases the mean values are the same ($2.Comparison X1 X2 XT PWInd [ X(T)] = X 0 + + + .98) = 0.04 (slide 32) • These two cases are considered as bounding the problem. Including Uncertainty 39 .(2885/2938)) = P(Z < -0... • PDep(PW < 0) = P(Z < ..938) is 73% larger than that of the independent case ($1.

- Moeller Datasheet MCS4 019711
- Adina Popescu-Manual de Spaga 09
- Adina Maria Secara Mihai Stoian Secret-Aloe Vera 09
- A. E. Van Vogt-Odiseea Navei Space Beagle 2.1
- Damian_Stanoiu-Pocainta_staretului_0.3_08__
- Damian Stanoiu-Nuvele Si Romane 0.3 07
- Damian Stanoiu-Necazurile Parintelui Ghedeon 1.0 10
- Damian Stanoiu-Cum Petrec Calugarii 0.3 07
- Damian Stanoiu-Casino Victoria 0.9 06
- Damian_Stanoiu-Camere_mobilate_0.3_07__
- Damian Stanoiu-Alegere de Stareta 0.3 10
- Politica Fiscala in Uniunea Europena
- No1Art18
- Strategia in Aprovizionarea Materiala
- EXPERIMENTUL LITERAR
- Performante Si Limite Umane
- CUNOASTEREA PLANORULUI
- Proiect La Calitatea Si Expertiza Marfurilor
- Floarea
- Determinarea Pozitiei Obiectelor in Spatiu
- Introducere in Sintaxa Si Semantica Logicii Propozitiilor
- Consistenta Si Consecinte
- Conectori logici
- Relieful Fundului Oceanic
- Aplicatie Practica Privind Impozitul Pe Venitul Din Drepturile de Proprietate Intelectuala

Sign up to vote on this title

UsefulNot useful- Collision Avoidance Minutes 03 14-Nov-07
- Uncertain
- PE_PSA_Presenting Uncertainty Around Cost-effectivenessestimates to Decision Makers How Sure Are We
- An Improved Method for Computing Wellbore Position Uncertainty and Its
- Chapter-7
- Group 2
- Managerial Economics
- Gottleib Format Uncertainty 2007 240 6
- Uncertainty
- Doc1
- Balancing Uncertainty in Structural Decision
- 7 International Probabilistic Workshop
- Project Planning_Improved Approach Incorporating Uncertainty
- Uncertainty Analysis of Raw Data and Calculated Data From Organic Rankine Cycle System
- Risk Analysis - Technical Note
- Imp Info RM Course 09
- Flight Safety Analysis Handbook Final 9 2011v1
- Bayesian EBDO Savannah Mourelatos Feb2008 V2
- Risk Uncertainty and Expected Value
- Aven_2012
- ADAM.pdf
- Engeco Chap 11 Breakeven and Sensitivity Analysis
- ch1
- 7 robustMPC
- As 3778.2.4-2007 Measurement of Water Flow in Open Channels General - Estimation of Uncertainty of a Flow Rat
- Uncertainty Calcium
- 027_Chp.7
- # Risk and Uncertainty (2009)
- Aven Perspectives on Risk in a Decision Making Context Review and Discussion 2009
- 8.4 Expected Value 5.7
- cba4