tim hw 4

Quantitative Decision Analysis (DA) to manage risk 1. Problem: how can DA Bbe used to model project risk by determining the project payoff($) Terminology (notation): Basic Building Blocks for DA: How do we calculate the payoff (emv) for a project Approach & Execution [use HW #3, problem #2: “umbrella problem”: should I buy an umbrella or not? ] Step 1:

choice 1

2 Decision Decision

choice Raw Block

1 2 probability

uncertain uncertain

event event 52awbability X

Raw Block Mobility

3 obtective

Maximize payoff or Expected neonetary chuelemu

plan orprocess forsowingall DecishnAnalysisproblems

yesx Rain

BUY an 21 NOumbrella 5 No

Gh

List all the building blocks relevant to the particular problem of interest, with all the associated branches

Step 2: Important remarks: when creating the ID sue only the “raw” building blocks Step 3: Covert the ID (from step 2) into a Decision Tree (DT) by including all the appropriate branches for the “decision” & “uncertain event” clocks from step . Include all relevant cost & probability information in the DT. For the umbrella problem, the nominal probability, p1 = 0.40 Step 4: Fold back the DT to obtain (calculate) the payoff associated with each of the choices in the decision block(s).

Time0 Timel Time2

present tomorrow future

BUYan Umbrella

Payoff minimize lose

Rain

Time 0 Tine 1 Time 2

yes Rain

25

25Buy.an No 4kgumbrella

pains 75

go

ii.im I 1 I

payoff yes I 2511 t 25114 cost 1512510.411 1525 10.6 5125

(Do NOT skip this step) Create an Influence diagram (ID), which is a high-level chronological view (from left to right) of the relation between the blocks in step 1

IIIIII I to 75

NOT 0 1 0.6

payoff No 7514 do k IT cost 75110.411 0110.6

7517 NahalMabuhay 04

SY Yes 25 9

040

By an T c L Yes t 25

umbrella No 75g C no 30

For the neutral case yes

buy an Umbrella1 Cost C payoff Notbuy 1 buy an KCLNo1 757

anunbtella umbrella 75

I

µ l y

25 ol l I i d l yti BS 1 oh Nobability

ofRain

Step 5: Choose the option (decision) which maximizes payoff. In this case, maximizing payoff is equivalent to minimizing cost for p1 = 0.40 (nominal) Cost (“yes”) = $25 Cost (“no”) = $75 p1 =$30 => Decision “yes” -> buy an umbrella Step 6: perform a sensitivity analysis to determine the “robustness” of your decision w.r.t. changes in the nominal value of the input parameters Pnom. Nominal probability => use the DT => decision 1 in the nominal probability, & then use in the DT to determine the corresponding decisions with these new probabilities. Example: for a +- 10% change in the nominal probabilities Robust Decision is one that does not change when the nominal probability is either increased or decreased by a certain percent (say 10%)

No Not buy an umbrella 975k

C ye Buy an Umbrella 25

at E75 1 25 7 4 0.33

Q f u40forthisMblem

makesmallcharges Lolo Isoko

front0.109mm 3 US he Decision2 a

0.4 10.111040

from 0.104nom w w coffle Declan30.4 0.110.40 0.36

2. HW #3 & project phase 1 Sign - up for a project review meeting next week phase 1 deliverables

Tech Companies

Existhycomsia.in n.orwseafoF7

medium Tim80cg

I smart

4250M

q 750

SlimHoon NB

V YourCompanyIs Youaregon a

medium sized

toanalyze Companymodeledafter www.tbignbz aMWhtkMk6mfuy Exishyldye Company HwSikhs

w mediumGIG Icrendbyyour team

D DD D x x x x

114 ExtryProducts technology whichYourmediumshed company Is armed Athy ii Seung to real Customers to obhhs an animalsewerrW V ofspoofsom newfrodwyserwce

idea to be

yawing

designed developed

mopsforyourConnydiesModulfsentas by Tomtommadeyouraway Wehrlyyourmentmke

Wen µ see antedduntfer te

list 4 tasks awardfor

thisIdea