Economic Decision Analy
Economic Decision Analy ISYE 6230
Popular in Course
Popular in Industrial Engineering
This 0 page Class Notes was uploaded by Maryse Thiel on Monday November 2, 2015. The Class Notes belongs to ISYE 6230 at Georgia Institute of Technology - Main Campus taught by Julie Swann in Fall. Since its upload, it has received 27 views. For similar materials see /class/234185/isye-6230-georgia-institute-of-technology-main-campus in Industrial Engineering at Georgia Institute of Technology - Main Campus.
Reviews for Economic Decision Analy
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 11/02/15
Game Theory and Military Decision Making Then vs Now ISYE 6230 Economic Decision Analysis April 15 2008 Lee Evans Outline Military Decision Making Process Doctrine minimax vs maximin Battle of the Bismarck Sea Applications for Personnel Recovery Conclusions Questions Military Decision Making Process Current Doeth Receive the Mis V 39ent 4 Strategies Mission Analysis Course of Action ex elop Course of Action Analysis Utility Course of Action Comparison lt 2Person Game Course of Action Approval Orders Production MinimaX vs Maximin Capabilities VS Intentions EX Pearl Harbor Doctrine of Decision is Conservative 0 MinimaXMaximin 7 ZeroSum Mixed Strategy clumn Firm Them Purim C D mu ell A 4 3 F3351 300 200 13 m 39 33m Solution B 1 2 1 35 100 500 x 33 l 6 117 039 7 Row player wants to maximize gains While column player wants to minimize the row player s gains Self Preservation Battle of the Bismarck Sea 0 Situation 7 Feb 1943 struggle for New Guinea 7 Japanese organizing resupply from Rabaul to Lae i 2 Japanese courses of action N or S routes 7 2 US courses of action aircraft N or S 7 Nature Poor Visibilityitoktlg North 39 N73 Battle of the Bismarck Sea mm 51 3 o y Battle of the Bismarck Sea l anumsr mum Minimum of row a r ya ul 2 km 2 Inyu 2 1m mhxluuv Kenm menu u l I a 5 am I thy Maximum ul cnlnmn What Happened 7 General Kenney focused the aircraft to the north 7 Japanese sailed north 7 Ship convoy sighted 1 day after it sailed and was defeated after 2 days of bombing 7 Not a Japanese error they did the same thing 2 months earlier with minor losses importance of intelligence 7 US had equipped airplanes for lowlevel bombing 7 Matched Strategies discovering other s strategy would not change your own 7 Zero Sum 2 mm 3 mm Iuiuimnx Current Game Theory Applications Utility Theory and Ordinal Values EX Cold War mm mm Aw 12mm Am 22 41 Pareto Optimal so EX Iraq Invas1on rewards and noncooperau39ve penalties Am 3 gt 7 is a way to ensure cooperation 1m Ecklund 2005 EX UN Resolutions 678 amp 1441 Cardinal vs Ordinal Values Ordinal Values may not accurately convey the relationships between preferences Umudstates Each player values similar things differently 3 quot With any game requiring a mixedstrategy solution cardinal values are essential to meaningful results Applications for Personnel Recovery quotil quot4 V 4 The aggregation 0 m1 ary civ1 an political efforts to recover captured detained evading isolated or missing personnel from uncertain or hostile environments and denied areas Applications for Personnel Recovery 0 Traditionally a recovery force was immediately assembled and launched to preclude the capture and exploitation of the isolated persons In the contemporary operating environment this is not always the best response Ordinal values both attempting offensive operations Developing a quantitative scale of military worth that is the same for both players is the greatest challenge in determining using a mixedstrategy solution Applications for Personnel Recovery Assumptions Rescue only conducted by a conventional force Adversaries fit into 7 categories First Rate Adversaries SecondRate Adversaries ThirdRate Adversary with poor popular support ThirdRate Adversary with strong popular support Locally Focused Insurgent Group Globally Focused Insurgent Group Criminal Element or Organization Locally Focused Insurgent Group Insurgency is a strategy adopted by groups too weak to attain their locally focused political objectives through conventional means Employ tactics designed for selfprotection Locally focused insurgents may attempt to replace the existing government revolutionary insurgency or they may have more limited aims such as my Avoid conventional battlespaces and focus in areas Where they can operate on footing that is to their advantage Insurgents try to postpone decisive action avoid defeat sustain themselves expand their support and over time change the power balance in their favor The best strategy is to attack the US s alleged center of gravitythe Will of its people Must appear strong While simultaneously attacking the US s legitimacy and credibility and the perception that the U S is in control of the situation for their information operations 10 campaigns With at least some level of passive public support and even minimal active support there is a reasonable expectation that the local populace Will give capturele to the insurgents Locally Focused Insurgent Group Insurgents have chance for propaganda Victory and chance to in ict casualties r39m39xedsmes I trong Kerover No Rerm er US unable to protect itself quot Capture 2 j 3 4 I Enemy 4 32 No Capture l Insurgents rely on populace for capture US Viewed as Weak strengthening IO campaign Populace perception Insurgents are Weak Locally Focused Insurgent Group 1pm E mm rims Outcome using maXimin Capture Recover Pareto Ef cient 0 Unique Nash Equilibrium 0 Highly Stable outcome I lxm my I 4 3 I What if Insurgents didn t use Maximin anaylsis FirstRate Adversan39es There wlll be no guarantee atleastlmually thatthe U s wlll have y or freedom ofmaneuver Adversarles ndfew reasons not to capture US personnel aganda value of a captured person would be less of a drawing factor beeause s eln ul not e are pnrnary focus of an enemy who as decldedto flghtthe U s on the eonyenuonal batde el Wouldrequlre oy rwhelmlng avla e non suppon packages to lncrease me eess to an acceptable level ses The nsk would probably be suen that reeoyery would be a feaslble or acceptable opuon m only the rarest ofcases FirstRate Adversaries l nun Shlu FirstRate Adversaries 0 Outcome using maximin Capture No Recover Pareto Efficient 0 Unique Nash Equilibrium 0 Highly Stable outcome dominant strategies 0 Can the Pareto Ef cient outcome be achieved FirstRate Adversaries The Normal Form game can be manipulated through the use of contracts ie the Third Geneva Convention 1949 Dictates treatment standards for Prisoners of War 115 countries signed the treaty Enforceable by each country or third parties such as the International Criminal Court The Hague C01 lCluSlOl lS The essential thought of game theory must be incorporated into a doctrine which permits commanders at all levels to make decisions based on their evaluations Hayward Jr 1954 predictability when Ordinal values can planning military op Determining a quantitative scale of military worth Specifi 39 0 R wgylsiheian gnormpus A breakthrough V Contracts can change the normal form game only if they are enforceable QUESTIONS G Humand Jr Jnlu39ml nf Dp exarinns sing 211121113 nMumsnmgirnerisinnixmtm and En 39nnm m Llarslull ErklumlU n mm 1m 12221 nn minus nf A 1949 Inmrnnrinnzl Cummings nf um Rm 1111 52 Ist 6230 Economic Decision Analysis Optimal Pricing Policy and Market Segmentation The case of wine Alejandro Mac Cawley April 17th 2008 l Alejandro Mac Cawley EB Geogegg e Ist 6230 Economic Decision Analysis Agenda Motivation and description of the product market and company Literature review Proposed models Single company no market interaction Single company with market interaction Two companies wo market interaction Discussion Conclusion Georgia Alejandro Mu Cawley Teohw Ist 6230 Economic Decision Analysis Motivation quotIn victory you deserve Champagne in defeat you need itquot quotRemember that half the grapes sell for above the median but half sell below Jim Verhey o Silversdo WineGrowers trying to encourage unrealistic expectations by some of the growers Wine gives us liberty love takes it away Wine makes us princes love makes us beggars Wycherly The Country Lile l Gear ia l Teath Alejandro Mac Cawley ist 6230 Economic Decision Analysis Product Market and Company Most brandeddifferentiated product you can ever think The only label in the table Cues for purchase RedWhite Country of origin even valley variety brand vintage Limited Markets are very complex Structure and regulations It s a drug Multiple price segments 2 U up to over 400 Multiple customers Countries Connoisseurs and non regular buyers Pricing and market decisions l Georgia ll Tech Alejandro Mac Cawley Ist 6230 Economic Decision Analysis Product Market and Company Company Prefers to stay con dential Sells a portion How much ofhis production to open market High prices and signaling and the rest to corporations as gilts Product is observed as a scarce product Structure of the process Company sets one price to both markets Markets demand a quantity How much to sell to each market Questions Optimal pricing Quantity to sell to each market Growth and marketing Competition l Gear ia l Teh Alejandro Mac Cawley ist 6230 Economic Decision Analysis Literature Review Mne Most researcn done in a marketing perspective Descriptive e Luckshin L SpavvtnnA and Macintnsn mus Usinvq Frodud brand andpurcnasinginvoivement ror retaii segmentation J Retaiiing and Cust Service a 4 Nn e Malurgiu3 Camazi and Grazia c 2n 7 Effectivenessofappellalionsoforigin in internationai wine markets Pmceeedings ins Cunference Eurupean Assnciatinn nrAgncuiturai Ecnnnmics Luca a mmcellllle r J n d Rndriguez i992 A Price discrimnation in ma kets numai cit Pniiticai Price discrimination and Revenue Management 7 We i984 Market segmentation sefrseleclion andpradud iine design Marketing Science a u e Cheri F Market Zuni Segmentatinn Advanced Demand inrormation and Supply cnain Penormance Man ampSer 0 Mana ementvni3 Nu i F39hlips L The Ecnnnmics er Price Discriminatinn NEWYurk Cambridge University Press i983 Mceiii J and van R yeri G 1899 Revenue Management Research overview andprospeds Transpunatiun Science Val 33 issue 2 Markets 7 Airiines Elmens tein Airiine lndusl M vni mun 3 e Hutels DunaghgK McMannnu and MchvveiiD lnternatiunalJuurnalufHuspitality Management vnii2issue2 i as General Eckel C and Smith i992 Price Discriminatinn thn Cnrreiated Demands Suuthem Ecnnnmianurnaivni 59 l Georgia ii Tech Alejandro Mac sawiey Ist 6230 Economic Decision Analysis Models 3 Models Single company 2 Markets No Interaction Single company 2 Markets Interaction Multiple companies 2 Markets Interaction Questions One price for both markets Optimal price Quantity to sell to each market Growth and marketing Competition l Georgia ii Alejandro Mac Cawley Tech Ist 6230 Economic Decision Analysis Model 1 F Price F Price Price is equal to both markets a Open Market 0 Corporae We sum across the quantities to obtain the total demand QT Q Q a a apb b a b I Gear 393 l Tegclhw39 Alejandro Mac Cawley Ist 6230 Economic Decision Analysis Model 1 o o c o Hwpq ptfcq Hf 2 t 5 NW 0 c 5 272b 17 lt0 qo f SQT 5p 0 gtlt 7a a cb b ia cb i n p 5 p i T T F W Tb WW Hwpq tfcq q pQ HQ a C F W E FOC 2a bTib aTch m 5H 9 Zb Tq Monopolisnc 5 r a a 72b b p cb b 0 T p of 975315 ia zfcb b i 7 2 T p n MOW 7 mm uzdgbrimr br 7 Tiq Monopolism b c aTb 7 T T Alejandro Mac Cawley I Gear 39 Tesg k z Alejandro Mac Cawley Example Q q q Q ISVE 6230 Economic Decision Analysis Model 2 Interaction between markets n 3 g i D D a a E Price is equal to both markets a c i 0 a Open Market QC 0 Corporate HP 7 HP QT QOQC a ac7pbabcigcypa a 1 Alejandro Mac Cawley ISYE 6230 Economic Decision Analysis Model 2 anw 2p27020gtp2c 55 UV 4quot p q cq q 3 an S PM p p p a a cb b rsw bquot n P 2babif i 25739 Pmam HWF4 4c4 4FQ CQ W P me 5 a a rlb b rspi pcb b rspivpg0 q p 7aiacbibgis yvp 391 p 21 1 75 mm SOC q a 7ltb M 5217 72bquotb r lt0 ltbquotb 5 5 5 Mbquot 5F 2 2M s q a 2 77 ZH MF 25 W W Georgia i Tech J Alejandro Mac Cawley Example Price Alejandro Mac Cawlev E 5 En gu 5 E1 5 Ist 6230 Economic Decision Analysis Example gqcvpa 705 0 05 15 As cross price elasticity goes up Optimal market price goes up Open market quantity goes down Corporate market quantity depends Value of e and b Profit is increased i Gear 392 i Teath Alejandro Mac Cawiey w ISVE 6230 Economic Decision Analysis Model 3 2 Companies P Price n P Price m a HP i Q Ci Open Market QC a Open Corporate P 7b air an 7 q 7a ibP q 7a qcvpa QT Qf Qz QHQJ a ap 0 b 5 m 11 t i Alejandro Mac Cawley I Getrgi ii ec V ISYE 6230 Economic Decision Analysis Model 3 Is a Bertrand equilibrium Bemanqu 0 ifR gt13 ci lta acgtepltb Wham ifP lt1 11 by 11 a 7pb b ist 1 ti 2 my If price under maximum quantity PQ1Q220 Then both companies produce their max q o If price under maximum quantity PQ1Q2ltC The price is set to cost and profit is 0 i Alejandro Mac Cawley l q Geargi ri ec V Ist 6230 Economic Decision Analysis Discussion Under single company no market relation Company aggregates markets and behaves as a monopoly Sets monopolistic prices and quantities Growth is determined by the monopoly quantity without quantity constrains Under single company with market relation om ggregates markets and behaves as a monopoly Sets monopolistic prices and quantities Quantities are affected by the cross price elasticity As elasticity goes up Increase in optimal price Reduce quantity on open market Cor orate market is uncertain Pro t is increased by the price elasticity Marketing Under multiple company withwithout market relation Is a Bertrand Price is set according the maximum production of both m anies or co Pro t is reduced for both companies 39 39 the market Establish barriers l Georgia il Tech Alejandro Mac Cawley Ist 6230 Economic Decision Analysis Conclusion Example of 2 markets in which a company sells their product Monopolistic behavior is expected Under cross market effects optimal prices and quantities are affected Optimal growth and marketing strategies have to be determined A new company entering the market will destroy the attractive of the market Build barriers Branding Collusion More research can be clone Effect of brands Variability on the production and signaling ect of marketing and reviews i Al 39 cl M c l Gear is l ejzn ro ac away Tesahw Ist 6230 Economic Decision Analysis I am falser than vows made in wine William Shakespeare 1564 1616 AS You Like 1599 1600 act III SC V I 73 1 Georgia L Alejandro Mac Cawley Tech Airfor Sale US Broadband Spectrum Auction a iVIatt Riley Agenda Properties of Auctions Types of Auctions Solution Strategies Development of an Auction Broadband Spectrum License Auction Lessons Learned Simultaneous Ascending Auction Properties of Auctions Open vs Closed Open Auctions Bids are made such that everyone knows about them Closed Auctions Bids are private Sealed Bid First price vs Second price Highest bidder wins Firstprice Winner pays value bid Secondprice Winner pays value on bid of runnerup Properties of Auctions Private vs Common value Private Value Everyone has different valuations of the object Usually used for their own purpose and not for resale Keepsakes memorabilia etc Common Value Has a true value market value and each bidder attempts to guess the value based on their own incomplete information Oil fields broadband licenses etc These are the auctions generally studied l AWE s M Classical Auctions First Price Second Price English Auction Dutch Auction First Price Auctions Each player submits a sealed bid The highest bidder wins and pays the price on their bid Properties In general FirstPrice Simultaneous Sealedbid Common or Privatevalued Usually incomplete information We have seen the classical model in class Example Incomplete Information 1950 s auction of offshore drilling sites to oil companies No reliable way to tell real value of site Let s say a particular field is worth 12 million Oil companies may value it between 520 million Highest bidder wins gt 19 million paid Total profit is l Wadi r M Winner s Curse A bidder tends to overestimate the value of an item and bids too much Solution Scale valuation by a factor 6601 Inexperienced bidders typically fall victim to Winner39s curse Experienced bidders anticipate winner s curse and scale their bids Solution Strategy Each bidder bids less than their valuation Second Price Auctions Vickrey Auction Each player submits a sealed bid The highest bidder wins and pays the price on the second largest bid Properties SecondPrice Simultaneous Sealedbid Common or Privatevalued Usually incomplete information SecondPrice Auctions Reduces effect of winner s curse since winners don t pay highest bid Solution Technique Bid valuation weakly dominant strategy l s English Auction An auction silent yelling buzzer where bids are indicated Via signals and the highest bidder wins Properties Sequential Common or Privatevalued Usually incomplete information Examples Art auction Property Realestate auction Strategically equivalent to the SecondPrice Auction Dutch Auction Similar to Chicken Game Price on a display At start price starts decreasing as time goes on First to buzz in wins and pays price on the display Properties Sequential Open Common or Privatevalued Usually incomplete information Strategically equivalent to FirstPrice Auction Revenue Equivalence Th eo re m m I Assumptions Risk neutral Independent private valuations Note tnat priv te valuations doem t mean it s a privatervalue auction it means ha each player svaluatiohs is not known by other players Symmetric strategy Payment is based on bids alone Also assume the valuation distribution F is continuous In equilibrium The bidder with the highest bid wihs The expecte payment ofthe bidder ofthe lowest type is 0 1Lecture belowl FirstPrice SecondPrice English and Dutch auctions all produce the same expected payof f for each bidder and forthe seller For detailed proof see Lecture Note Lots of real analysis in this presentation ofthe proof Broadband Spectrum Auction eel US government decided to auction off licenses to use broadband spectrum in 1993 Previously done under administrative decision Change due to inefficiency First examine other countries39 broadband spectrum auctions Lesson Learned s M New Zealand 1990 Second Price simultaneous auction Result Very low revenue NZ36 million of project NZ24O million Winning bids in excess of NZ100000 Second highest bid N256 Or highest NZ7 million second highest NZ50001 University student bid NZ1 on small city license1 No other bidders Who said you can39t get a free lunch 1McMillan 1994 Major Problem with Auction Rules No reserve price Absolute Auction Reserve price is used in place of many bidders to ensure price is acceptable and competitive Chose to use First Price Auctions l w 1 Australia 1993 Firstprice sealedbid auction for two satellite television licenses Results Winner s Curse caused major companies to bid low quotDark Horsequot companies submit as many as 20 bids in 5 million increments Withdraw all bids until the lowest winner I Two top bids AS212 and AS177 million Price paid AS117 and AS77 a decrease of AS195 million less than originally bid1 Sells bids for 21 million profit 1McMillan 1994 Major Problems with Auction Rules No deposit required Anyone can bid regardless of how serious they are in winning No limit on the number of bids Absence of default penalty Bids mean nothing Incentive to deviate from bid Get additional information from the bids and can default without penalty l E a 14 Lessons Learned Game theory needed to ensure adequate functioning of auction Deposit required to bid Reserve price if few bidders Bid penalty associated with default or withdrawn bid Background US Broadband License Auction Hired Game Theorists to create a new type of auction 51 major trading areas MTAs 2 licenses each 30 MHZ 492 subsets of MTAs is basic trading areas BTAs Five icensesOne 30MHZ four lOMHz Companies need to aggregate license to run efficiently Nationalstate coverage US Broadband Auction I Objectives quotefficient and intensive use of electromagnetic spectrumquot Rapid deployment of new technologies Preventing excessive concentration of licenses Ensure some licenses go to minorityowned women owned companies small businesses and rural telephone companies I But still need to ensure some aggregation for companies with national coverage I Revenue low on objective list Once other objectives met maximize revenue Simultaneous Ascending Auctions I 130 pages of rules Basic outline is Open and sequential auctions I Reduces Winner39s Curse Entrance deposit I Only serious bidders allowed All licenses open at the same time I There are interchangeable licenses Bidding continues until no more bids made Penalty on withdrawn bids I If highest bid withdrawn at any point in the auction must pay difference between withdrawn bid and final highest bid Activity Requirement see next page Activity Requirement I Requires each firms to indicate and submit deposits on the number of licenses they wish to obtain Three stages in which at each stage the firm must be the top bid or submit a bid over the minimum for a certain percentage of the number of licenses indicated above I Stage 1 13 I Stage 2 23 I Stage 3 100 If it does not meet these levels then the number of licenses able to be obtain decreases proportionally l meet 1 m Results of US Auction I Successful aggregation of licenses to multiple firms I Able to fulfill objectives of minority owned woman owned rural and small firms I Gave special privileges to a few firms who developed technology I Revenue higher than expected quotI3 M1 Summary Classical Auctions First and Second Price English Dutch Revenue Equivalence Theorem generalized US Broadband Auctions new auction type Bidder and seller receive same expected revenue regardless of classical auction type Can be further Game theory applied to successfully develop a References Journal ofEconomic Perspectives 101 159717 McMillan John 1994 Selling Spectrum Rights In Journal of Economic Perspectives 831457162 WisconsinrMad ison Accessed 4112008 McAfee R Preston and John Mcmillan quotAnalyzingthe AirwavesAuction ln Quint Daniel quotEcon 805 Advanced Micro Theory 1 Lecture 3 University of nnl
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'