Online Supplement for The Value of Bespoke : Demand Learning, Preference Learning, and Customer Behavior

Size: px
Start display at page:

Download "Online Supplement for The Value of Bespoke : Demand Learning, Preference Learning, and Customer Behavior"

Transcription

1 Online Supplemen for The Value of Bespoke : Demand Learning, Preferene Learning, and Cusomer Behavior Tingliang Huang Carroll Shool of Managemen, Boson College, Chesnu Hill, Massahuses 0467, inglianghuang@bedu Chao Liang Cheung Kong Graduae Shool of Business, Beijing 00738, China, liang@kgsbedun Jingqi Wang Fauly of Business and Eonomis, The Universiy of Hong Kong, Pokfulam, Hong Kong, jingqi@hkuhk This Online Supplemen inludes four seions Online Supplemen provides he ehnial proofs o he proposiions, heorems and orollaries in he paper All he oher seions inlude some supporing maerials and exra analysis for he paper Speifially, Online Appendix inludes supplemenal analysis for 5 of he paper in Online Supplemen 3, e analyze a varian of he basi model in 3 and 4 by alloing he firm o prie afer seeing he demand realizaion Online Supplemen 4 shos ha our model used in 5 of he paper an be operaionalized and suppored by he inerpreaion of produ line design or assormen planning Online Supplemen Proofs Proof of Proposiion Leing uθ, p p θ = 0, e have θ = p Noe 0 θ The demand is p D hen p, is 0 hen p >, and is D hen p < Le p 0, q 0 and π 0 denoe he opimal prie, opimal produion quaniy and opimal profi of he radiional sysem, respeively We mus have p 0 and p 0 d L q0 p 0 d H Noe ha for p and p d L q p hen D = d H, and πp, q = p min p d H, he firm s profi πp, q = p min p d H, q q = p q d L, q q = p p d L q hen D = d L By aking he expeaion over D, e have πp, q = p[q + p d L ] q Noie ha πp,q = p and q πp, q is linear in q For any p here i When q = p p p, he opimal q is eiher d L or p d H d L, e have πp, q = p d L p, hih is maximized a p = p L = + by solving dπ = 0, and he orresponding produion quaniy and profi are q = q dp L = d L and

2 Huang e al: The Value of Bespoke Arile submied o ; manusrip no πp L, q L = d L Noe p L due o + and > ii When q = p d H, e have πp, q = µ D d H µ D pp d H µ D For d H µ D, e have ha µ D pp d H µ D is maximized a p = p H = + by solving dπ = 0, and he orresponding produion dp quaniy and profi are q = q H = d H d H µ D and πp H, q H = µ D d H µ D Noe ha p L p H due o d H µ D and d H µ D For < d H µ D, e kno µ D pp d H µ D 0 hen p Noe ha πp L, q L πp H, q H if and only if d H d L µ D Combining ases i and ii, e d L µ D have he folloing If < d H µ D, hen p 0 = p L = +, q 0 = q L = d L, and π 0 = πp L, q L = d L If d H µ D, hen hen d H d L µ D, e have ha p d L µ 0 = p L = +, q 0 = q L = d L, D µ D and π0 = πp L, q L = d L ; oherise, e have ha p 0 = p H = +, q 0 = q H = d H d H µ D, and π0 = πp H, q H = µ D d H µ D Beause d H µ D < d H d L µ D, Proposiion follos d L µ D Proof of Proposiion Beause he firm deermines is produion quaniy q afer learning he realizaion of D, i alays ses q = p D for any prie p, here D = d L or d H depending on he demand realizaion The profi an be expressed as πp = p d H p + p d L p = µ D p p, hih is maximized a p = p L = +, and he orresponding opimal profi is µ D Proof of Proposiion 3 Beause he firm an ompleely remove he preferene-mismah os, i alays ses he opimal prie p p = and all he usomers ould like o buy a his prie The firm s profi is π, q = E[ mind, q] p q Similar o he analysis in Proposiion, e kno ha he opimal produion quaniy q p saisfies d L q p d H Afer aking he expeaion over µ D µ D D, e have π, q = [q + d L ] p q Noie ha π,q q d H µ D = p and π, q is linear in q We kno ha if p, hen q p = d H and he firm s opimal profi π P = µ D p d H ; oherise, q p = d L and π P = p d L Proof of Proposiion 4 In Bespoke sysem, he firm alays harges he prie sine i an ompleely remove he preferene-mismah, and i alays ses he produion quaniy equal o he demand realizaion d H or d L sine i an learn he demand realizaion before deermining he quaniy Therefore, is opimal profi πap = p d H + p d L = p µ D Proof of Lemma Reall κ = d H d L µ D We an prove ha κ dereases in by d L µ D shoing dκ d µ D < 0 Noe µ D = d H + d L We have µ D d L µ D = d H d L µ D + d L d L µ D d H d L µ D, and he resul κ follos Noie ha d L µ D d L+µ D and κ We have κ d H d L +µ D µ D d L +µ D Proof of Theorem Reall κ = d H = d H d L +d H µ D µ D d L = d H d L + d H d L d H d L µ D d L µ D = d L µ D We have shon ha κ is dereasing in, and κ > There exiss 0 > suh ha κ 0 = We disuss he folloing ases:

3 Huang e al: The Value of Bespoke Arile submied o ; manusrip no 3 i Suppose < min p, κ} Then by Proposiions -4 and simplifying, he suffiien and neessary ondiion π0 + πap πa + πp is equivalen o p Noe ha he ondiion < min p, d H d L µ D } is equivalen o < µ D d p and d H d L µ D > L µ D µ D d Beause κ = L µ D d H d L µ D and is dereasing in, e have ha d H d L µ D > µ D d L µ D µ D d holds if <, and hus L µ } D < min p, d H d L µ D is equivalen o < min d p, L µ 0}, here κ 0 = Hene, in his ase, D µ D he ondiion for omplemens is < min p, 0} and p ii Suppose min p, κ} max p, κ} This ondiion is equivalen o min p, 0} max p, 0} We disuss o subases: iia If 0 < p, hen he supposed ondiion beomes 0 p In his subase, e have p and κ By Proposiions -4 and sim- plifying, he suffiien and neessary ondiion π0 + πap πa + πp [ subase, e have p is equivalen o p ] + d H µ D iib If 0 p, hen he supposed ondiion beomes p 0 In his and κ By Proposiions -4 and simplifying, he suffiien and neessary ondiion π 0 + π AP π A + π p is equivalen o p p+ Nex, e sho p p+ alays holds in his subase Beause 0 in his subase and κ dereases in, i is suffiien o sho κ 0 κ p p+ Noe κ 0 =, κ p p+ p p+ Lemma, and ha p p+ p p+ p p+ is equivalen o + p hih is alays rue due o p > and + Hene, he ondiion in his subase is p 0 Noie ha have p p p p p+ an be rerien as p Sine p in his subase, e, hih is simplified o p Therefore, he ondiions in his subase an also be rien as p 0, and p iii Suppose > max p, κ}, hen by Proposiions -4 and simplifying, he suffiien and neessary ondiion π 0 + π AP π A + π p is equivalen o p + + d H µ D, hih alays holds due o +, p >, and d H µ D Hene, he ondiion in his ase is equivalen o > max p, 0} We are no ready o summarize all hese ases ogeher: If and only if one of he folloing o ondiions holds, demand learning and preferene learning are omplemens: 0 and p ; 0 and p C p, [ ] } here C p = max + d H µ D, Here is from ases i and iib, and is from ases iia and iii Proof of Corollary When 0 p, he suffiien and neessary ondiions an be rien as < p and p in ase i, p 0 in ase ii ie, ase iib, and > 0 in ase iii In order o ge he resul, e jus need o sho ha p is redundan in ase i hen 0 p Beause P 0, i is suffiien o sho 0, equivalenly,

4 Huang e al: The Value of Bespoke 4 Arile submied o ; manusrip no κ 0 κ Noe κ 0 =, κ Lemma, and ha is equivalen o hih is alays rue due o + Hene, hen 0 p, he suffiien and neessary ondiion in ase i an be simplified o < p, and he desired resul follos Proof of Lemma From equaions -5 in he paper, e an express, in erms of demand learning auray I, he firm s poserior probabiliies of he demand foreas afer observing signal s as P D = d H s = d H = β H = I, P D = d L s = d H = β H = I, and P D = d L s = d L = β L = I, P D = d H s = d L = β L = I d H Suppose ha he firm observes demand signal s = d H, hen is updaed demand Ds = d H = ih probabiliy H I β H = I and Ds = d H = d L = 0 ih probabiliy H I = I Based on he analysis for he radiional sysem, e kno ha e need o disuss he folloing o ases: If H I /, ie, I, hen i is opimal o produe zero quaniy, hih resuls in zero expeed profi, ie, π s=d H = 0 If H I /, ie, I, hen he firm s opimal prie p s=d H = +/ H I, opimal produion quaniy qs=d H = d H H I, and opimal profi π s=d H = H Id H H I Suppose ha he firm observes demand signal s = d L, hen is updaed demand Ds = d L = d H ih probabiliy L I β L = I and Ds = d L = d L = 0 ih probabiliy L I = I Based on he analysis for he radiional sysem, e kno ha e need o disuss he folloing o ases: If L I /, ie, I, hen i is opimal o produe zero quaniy, hih resuls in zero expeed profi, ie, π s=d L = 0 If L I /, ie, I hen he firm s opimal prie p s=d L = +/ LI, opimal produion quaniy q s=d L = d H and opimal profi π s=d L = LId H L I, L I, To rie he firm s expeed profi before observing he demand signal s, e need o ompare I and I I urns ou I I is equivalen o / Noe ha π AI, = π s=d H + π s=d L Then, e an ompue his expeed profi based on differen ombinaions of he ondiions on and I For example, if [0, ], e have I 0 I If e furher assume I [I, ], hen π AI, = I d H I Similarly, e obain he profi funions in oher ases

5 Huang e al: The Value of Bespoke Arile submied o ; manusrip no 5 Proof of Theorem We firs disuss all he possible ases Suppose We disuss he folloing o ases depending on he magniude of I: If I [I, ], hen π0 + πap πa + πp is equivalen o 0 + d H p H Id H H I + 0, hih is simplified o p I I If I [0, I ], hen π 0 = π A = π p = 0 and π AP = d H p I is lear ha i is alays ha demand learning and preferene learning are omplemens in his ase Suppose [, p ] Under his assumpion, e disuss he folloing ases depending on he magniude of I: If I [I, ], hen he ondiion for omplemens is d H +dh p, hih an be simplified o p I H Id H H I If I [0, I, hen he ondiion for omplemens is d H p H Id H H I + L Id H I + d H + L I, hih an be simplified o I p I + I I Suppose p We disuss he folloing ases depending on he magniude of I: If I [I, ], he ondiion for hem being omplemens is d H + dh p H Id H p H I + p d H, hih an be simplified o I I A- Noie ha his inequaliy holds ih he sri inequaliy hen I = as shon in he basi model, sine he RHS an be simplified o [ + ] < = p Hoever, e an sho ha he RHS is a srily inreasing funion a he poin I = To see his, denoe fi I I Then, e have he firs-order derivaive f I = [ ] A I =, e have f = I > 0 Hene, here exiss I 0, suh ha he RHS is equal o, ie, fi < p Hene, for I [maxi, I, ], his inequaliy alays holds If I [0, I ], hen he ondiion for hem being omplemens is d H + dh p HId H H I + p d H, + LId H hih an be simplified o I p I + I L I I A-

6 Huang e al: The Value of Bespoke 6 Arile submied o ; manusrip no Therefore, i beomes sraighforard o hek ha he saemen for a high I holds This omplees he proof Proof of Theorem 3 To pin don he profi funions for he four sysems, e need o disuss heher p + p p holds or no Suppose p + p p We have p p+ p Under his assumpion, e firs onsider he subase, ie, We have π 0 + π AP π A + π p equivalen o 0 + d H p p 0 + d H Simplifiaion of his equaion and subsiuing p = J yields p J J Suppose p + p p We have p p+ p Under his assumpion, e have o subases: i p + p and ii p + p We disuss eah of hese subases i Case ia: If p + p, hen he inequaliy π 0 + π AP π A + π p is equivalen o 0 + d H p p 0 + d H Simplifiaion of his equaion and subsiuing p = J yields p J J Case ib: If p + p, hen inequaliy π 0 + π AP π A + π p is equivalen o 0 + d H p p 0 + d H, simplifying hih yields p J J ii p + p If, e have π0 + πap πa + πp equivalen o 0 + d H p p 0 + d H Simplifiaion of his equaion and subsiuing p = J yields p J J To summarize he resul above, e have obained he resul in par proved: If and p J min J, } J, demand learning and imperfe prefer- ene learning are omplemens If pj pj+ J, e have p + p p If e furher have > folloing subases: a If p, e have π 0 + π AP π A + π p, e need o disuss he + equivalen o d H d H p p 0 + d H Simplifiaion of his equaion and subsiuing p = J [ ] yields p + J Noe ha in his ase, e also have he ondiion p p b If p + p, e have π 0 + π AP π A + π p equivalen o d H + d H p p d H p p + d H Simplifying his inequaliy yields p + We laim ha his inequaliy alays holds for he folloing reasons Sine p, i is suffiien o sho ha, and + + +, hih an be simplified o Bu +, e indeed have Hene, if p p, e alays have π 0 + π AP π A + π p Combining ase a and b, e obain he resul Proof of Proposiion 5 Suppose K AP 0 K A0 + K P 0, I 0 < p and K I 0 A K A Noe and K AP 0 K A0 K P 0 0 From Proposiion, and Proposiions A-3-A-5 in Online Appendix 4, e have he folloing resul: If p J AP 0 and K AP > K AP, p, hen he omplemenary ondiion π 0 +π AP π A +π P is equivalen o d H J AP 0 p K AP 0 K A0 K P 0, hih alays holds If p < J AP 0 and K AP > K AP, hen π 0 + π AP

7 Huang e al: The Value of Bespoke Arile submied o ; manusrip no 7 π A + π P if and only if d H [ J AP 0 p ] K AP 0 K A0 K P 0, hih alays holds due o p < J AP 0 3 If p J AP 0 and K AP K AP, p, hen π 0 + π AP π A + π P if and only if d H p K AP J AP 0 K AP 0 K A0 K P 0, hih alays holds due o K AP K AP, p and K AP, p = d H p J AP 0 [ p J AP 0 ] 4 If p < J AP 0 and K AP K AP, hen π 0 +π AP π A +π P if and only if d H p K AP J AP 0 K AP 0 K A0 K P 0, hih alays holds This is beause ih p < J AP 0, e have p > J AP 0 Wih K AP K AP, K AP = d H, e furher have d H p > K AP J AP 0 and hus d H p > K AP J AP 0 From -4, e kno ha demand learning and preferene learning are omplemens ih K AP 0 K A0 + K P 0, and K I 0 A > K A Beause K A = d H I 0, e kno ha K A > K A is equivalen o K A I 0 d H Noe ha I 0 0 is equivalen o I 0 Le = min, K A I 0 I 0 d H } and he desired resul follos Proof of Lemma 3 We solve for sub-game perfe equilibria by analyzing he firm s deision in he seond period firs Beause r s, he firm ill never hoose an iniial invenory level higher han he firsperiod sales in high demand ase Therefore, he firm ill have remaining invenory q in he seond period only if he demand is lo Given p, a usomer loaed a θ buys if r p θ 0 and θ θ Solving r p θ = 0, e have θ p = r p Then he period demand is d L θ p θ + = d L r p θ + And he firm s period profi maximizaion problem is max π p = min p r θ } r p d L θ, q p + s [ ] + r p q d L θ Solving his opimizaion problem, e have: if r θ > s, ie, θ < r s/, hen r θ +s if q r s θ dl, p q, θ = r θ q d L if q < r s θ dl, A-3 hereas if r θ s, ie, θ r s/, he firm does no an o sell in he seond period and all q unis are salvaged a s Subsiuing he opimal p bak, e have he opimal period- profi sq if θ r s π d q, θ = L r θ s + sq if r s q d L θ r s, q r θ q d L if θ < r s q d L A-4 We no onsider he firs period An individual onsumer does no kno he demand realizaion The uiliy of purhasing in period is p θ If θ r s/, hen he firm does no sell

8 Huang e al: The Value of Bespoke 8 Arile submied o ; manusrip no in he seond period, and θ p, p = p If θ < r s/, he expeed uiliy of purhasing in period is r p θ here p is onsumers raional expeaion of he period prie hen he demand is lo Though he opimal p depends on θ, for any individual onsumer, p does no depend on her on deision Solving p θ = r p θ, e have θ p, p = p r p Therefore, usomers opimal purhasing deision in he firs period is θ p, p = p r p p The firs-period demand is Dθ p, p if θ < r s/, if θ r s/ A-5 We nex analyze he firm s soking/invenory deision q in period Noe ha θ p, p does no hange ih respe o he firm s iniial invenory deision Hene, aking p and θ as given, he firm should hoose q o maximize is expeed profi Beause p is alays higher han, given θ, he opimal invenory level should be higher han θ d L In addiion, as r s, he opimal invenory level should never exeed θ d H When he demand is lo, q = q θ d L So q r s follos: θ dl is equivalen o q r s + θ dl The firm s iniial invenory deision is as max p q + p θ d L + π q θ d L, θ q, θ d L q θ d H here π q, θ is given in equaion A-4 If θ r s, he firm s opimal invenory problem is max p q + p θ d L q + sq θ d L θ d L q θ d H The firs order derivaive is π q = p + s Hene, he opimal iniial quaniy is q = θ d L if p < s, and q = θ d H if p s In his ase, he firm does no sell in he seond period, and θ = p Subsiuing θ = p In his equilibrium, if p < s, and if p s, π p = p p π p = p p d H + p p bak o θ r s, e have p r + s p d L + p d L, If θ < r s, he firm s opimal invenory/soking problem is d L + s p d H d L max π q; p, θ, θ d L q θ d H A-6

9 Huang e al: The Value of Bespoke Arile submied o ; manusrip no 9 here p q + if θ d L q < r s + θ dl, π q; p, θ = p q + p θ d L + d Lr θ s if r s + θ dl q θ d H p θ d L + q θ d L r θ q θ d L d L q q + sq θ d L When θ d L q < r s + θ dl, e have π q = p + A-7 r + θ q d L, π q; p, θ = /d L < 0, and he soluion o he firs-order ondiion π q; p, θ = 0 is q = r + θ dl + d L p When r s hen q r s + θ dl q θ d H, π q = p + s If p < θ r + + θ dl and π q 0 for r s < s, + θ dl q θ d H In addiion, e an verify ha q < d L θ Therefore, he opimal iniial invenory is q = d L θ and he firm does no sell in period Then, θ = p Hene, θ < r s is equivalen o p > r + s The firm s profi is π p Subsiuing θ bak o p < θ r +, e have p < + r If θ r + p < s, hen q r s + θ dl and π q 0 for r s + θ dl q θ d H In addiion, e an verify ha q > d L θ In his ase, he opimal invenory is q = q Then q = r θ dl + d L p and p = r θ q d L = r θ p Subsiuing p = r θ p ino equaion A-5 and solving for θ, e have θ = r +p + Then + θ < r s is equivalen o p > + r+s+ +r+, and q = d + p + L Subsiuing + θ and q bak o equaion A-7, e have he firm s profi as π 3 p Subsiuing θ bak o p θ r +, e have p + r If p + s > 0, ie, p > s, hen q > r s + θ dl and π q > 0 for r s + θ dl q θ d H In his ase, he opimal invenory is q = θ d H Then q = θ d H d L and p = r θ +s Subsiuing p = r θ +s ino equaion A-5 and solve for θ e have θ = r p + s + r p Then q = d + s H Subsiuing θ + = r p + s + bak o θ < r s, e have p > r + s Subsiuing θ and q bak ino equaion A-7, e have he firm s profi as π 4 p Comparing he hresholds of p, e have o ases depending on he primiive parameers: Case : When s r + s r + s, hen r+++s s + If p < s, hen beause p < s r ++s equilibrium θ r s Therefore, θ = p and q = p d L If s p < r + s, hen beause s Therefore, θ = p and q = p d H, and + r and p < s r + s, in r +s+, in equilibrium θ r s

10 Huang e al: The Value of Bespoke 0 Arile submied o ; manusrip no If p r + s, hen beause p > s, in equilibrium θ r s Therefore, θ = r p + s + and q = d H r p + s + To summarize, in his ase, he firm s profi as a funion of p is π p if p < s, π p = π p if s p < r + s, π 4 p if p r + s A-8 The profi π 4 p an be obained by subsiuing q = d H r p + s + and θ = r p + s + bak ino equaion A-7 he par ih r s + θ dl q θ d H Case : When s > r+s, hen + r > r+s, + r > r+++s + and + r < s If p < r + s, hen beause p < r + s < s θ = p and q = p d L, in equilibrium θ r s Therefore, If r + s < p < + r, hen in equilibrium θ < r s Therefore, θ = p q = p d L If + r p < s, hen beause p + r > r++s+, in + equilibrium θ < r s Therefore, θ = r +p + + and and q = d L +r+ + p + If p s, hen p > r + s, and hus in equilibrium θ r s Therefore, θ = r p + s + and q = d H r p + s + To summarize, in his ase, he firm s profi as a funion of p is π p if p < + r, π p = π 3 p if + r p < s π 4 p if p s, A-9 The profi π 3 p an be obained by subsiuing q = d L +r+ + p + and θ = r +p + + bak ino equaion A-7 he par ih θ d L q < r s + θ dl Proof of Lemma 4 In he preferene learning sysem, all usomers obain heir ideal produs sine here is no preferene mismah Thus, in he seond period, he firm s opimal prie is p = r and usomers buying in he seond period ill ge zero uiliy Therefore, he firm s opimal prie in he firs period is p = and all usomers an o buy in period The only remaining deision is he iniial invenory level I is lear ha he opimal invenory level is beeen d L and d H The firm s expeed profi is max dl q d H π q = q + d L + s q d L p q Then, π q = + s p Hene, he resuls follo Proof of Proposiion 6 Wih d L = 0, e an use Proposiion for he radiional sysem as: i if, he π 0 = 0; ii if >, he π 0 = d H Wih d L = 0, e an use Proposiion 4

11 Huang e al: The Value of Bespoke Arile submied o ; manusrip no for he Bespoke sysem as π AP = d H p Noe + J is equivalen o J We disuss he omplemenary ondiion in elve differen parameer ranges: Suppose J, I [0, I A, p ] and < Then, by Proposiions A-6 and A-7 in Online Appendix 4 and he profi expressions for he radiional and he Bespoke sysems lised above, he suffiien and neessary ondiion π 0 + π AP π A + π p an be rien as 0 + d H p d H J + 0, hih an be simplified as p Suppose J, I [0, I J A, p ] and < p Then, π 0 + π AP π A + π p if and only if p [ J I [0, I B, p ] and p Then, π 0 + π AP π A + π p if and only if p ] 3 Suppose J, [ J ] 4 Suppose J, I [I A, p, ] and < Then, π 0 + πap πa + πp if and only if J π A + π p I 5 Suppose J if and only if J and p Then, π 0 + π AP π A + π p + I, I [I A, p, ] and < p Then, π 0 + π AP 6 Suppose J, I [I B, p, ] if and only if J 7 Suppose + I J >, I [0, I A, p ] and < Then, π 0 + πap πa + πp if and only if p + J 8 Suppose J >, I [0, I A, p ] and < p Then, π 0 + πap πa + πp if and only if p + π 0 +π AP π A +π p if and only if p J + 9 Suppose J >, I [0, I B, p ] and p Then, [ J ] 0 Suppose J >, I [I A, p, ] and < Then, π 0 + π AP π A + π p if and only if J I Suppose J >, I [I A, p, ] and < p Then, π 0 + π AP π A + π p if and only if J I Suppose J >, I [I B, p, ] and p Then, π0 + πap πa + πp if and only if J I No e invesigae he pariular ase saed in his proposiion Suppose I [0, I B, p ] and J [0, ] From laim 3 analyzed above, he ondiions are [ J ] p < Nex, e prove ha here exiss Ĵ > 0 suh ha hen J Ĵ, p ] alays holds Noe ha [ J ] p > [ ] is equivalen o < p p >, < and +, e kno ha p > J=0 [ [ J = [ ] Beause + + hih is alays rue due o J ] alays holds Beause [ J ] is oninuous and inreasing a J, e an alays find suh an Furhermore, by solving p [ J=0 J ], e have Ĵ = p + Ĵ > 0 Suppose I [0, mini A, p, I B, p }] and J [0, min, p }] Then + from laims -3 of above, demand learning and preferene learning are omplemens if and only if one of he hree ondiions holds: and p J, < p p [ J ], and 3 p < Proposiion 6 follos by ombining ondiions and 3 and

12 Huang e al: The Value of Bespoke Arile submied o ; manusrip no Online Supplemen Supplemenal Maerials for 5 In his seion, e provide supplemenal maerials for 5: Exensions o he Basi Model Online Supplemen Analysis for 5 Imperfe Preferene Learning We firs analyze he preferene learning sysem For analyial onveniene, e also normalize d L o zero in his seion We ill abuse noaion by riing p = p J given ha J is fixed The proposiion belo provides he firm s opimal expeed profi for an imperfe preferene learning sysem J: Proposiion A- In an imperfe preferene learning sysem ih he level of usomizaion i If [ p + J, +, he firm opimal expeed profi π pj, = [ J p]d H ii If [ p, p + J], he firm opimal expeed profi π pj, = iii If [0, p ], he firm opimal expeed profi π pj, = 0 d H J p Proof of Proposiion A- To find he onsumers ho are indifferen o purhasing he produ versus no purhasing loaed a θp for a given prie p, e le uθ, p p θ p = 0 We have θp = p p If p [ p, ], e have θp [0, ] so ha he number of onsumers ho purhase he produ is D p p = pd p If p [0, p], all he onsumers ill purhase he produ, ie, D p p = D If p >, D p p = 0 The firm hooses is prie p and produion quaniy q o maximize he expeed profi πp, q = pe mind p p, q p q p mind H, q p q, if p [0, p] πp, q = p min pd H p, q p q, if p [ p, ] p q, if p [, + Le us disuss differen ases depending on he range of prie p For p [0, p], he opimal prie has o be p sine πp, q srily inreases in p Then, he opimal quaniy q = d H if p p and q = 0 if p < p Hene, in his region, he maximum expeed profi πp, q = [ p p ]d H p < p if p p and πp, q = 0 if For p [ p, ], e have πp, q = p min pd H p, q p q p p q and q pd H p Therefore, πp, q = p p pd H p Firs-order ondiion ih respe o prie p yields p = + p Then, e obain he orresponding quaniy q = d H p p Finally, e need o make sure ha he prie is ihin his region, ie, p = + p p, hih is simplified o p p + p Under his ondiion, he maximized expeed profi πp, q = d H p p If his ondiion does no hold, e disuss o subases: If > p + p, hen he opimal prie should

13 Huang e al: The Value of Bespoke Arile submied o ; manusrip no 3 be on he boundary, p = p and he opimal quaniy q = d H The opimal expeed profi is πp, q = [ p p ]d H ; If < p, hen he opimal prie p =, opimal quaniy q = 0, and expeed profi is zero sine no onsumers purhase a his prie To sum, he opimal expeed profi is [ p p ]d H, if [ p + p, + π d p = H p p, if [ p, p + p] 0, if [0, p ] Sine p = J, e an express he profi funion in erms of J Nex, e analyze he Bespoke sysem here preferene learning is imperfe ih he level of usomizaion J and demand learning is perfe Proposiion A- In a Bespoke sysem of imperfe preferene learning ih he level of usomizaion J and perfe demand learning: i If p + J, he firm opimal expeed profi π AP J, = d H J p ii If p + J, he firm opimal expeed profi π AP J, = d H J p Proof of Proposiion A- In his sysem, he firm an observe he demand realizaion D before making is quaniy deision Hene, he firm an perfely mah supply ih demand Suppose D = d H, hen he number of onsumers ho ill purhase he produ is d H, if p [0, p] pd D AP d H = H p, if p [ p, ] 0, if p [, + We hen disuss under eah of he hree prie ranges, ha he opimal prie should be If p [0, p], he firm profi is π dh p, q = p p d H The opimal prie p d H = p and opimal quaniy q d H = d H if p p Hene, π d H = p p d H if p + p If p [ p, ], hen q d H = pd H p The firm profi π dh p, q = p p pd H p Firs-order ondiion ih respe o prie p yields p p p = 0 Hene, p = +p For i o be opimal, e require p = +p [ p, ], hih is simplified o [ p, p + p ] Hene, under his ondiion, e have p d H = +p, q d H = d H p, and he opimal profi πd H = d H p p If his ondiion p does no hold, p + p, e have p d H = p, q d H = d H, and he opimal profi π d H = p p d H The opimal expeed profi is πap = πd H Hene, e obain dh π AP = p p, if p + p d H p p, if p + p Sine p = J, e an express he profi funion in erms of J

14 Huang e al: The Value of Bespoke 4 Arile submied o ; manusrip no Online Supplemen Analysis for 53: Invesmen Coss and Endogenous Learning Levels In his seion, e provide he deailed analysis for he hree sysems in he linear-os seing The resul for radiional sysem remains he same as in he basi model Proposiions A-3-A-5 belo summarize he resuls for demand learning, preferene learning and Bespoke sysems, respeively We firs inrodue some noaions for exposiional onveniene Le K A K A d H I 0 [ I 0 ][ [ I 0 ][ I 0 ] I 0 [ I 0 ] }, and K A3 I 0 ] } Proposiion A-3 In a demand learning sysem ih linear oss, e have: d H I 0, d H I 0 i The opimal learning level I = I 0, if and only if one of he folloing hree ondiions is saisfied: and K I 0 A > K A, < I 0 and K I 0 A > K A, 3 < and K I 0 A > K A3 The orresponding opimal profis are πa = K A0, π A = [ I 0 ]d H [ I 0 ] K A0 I 0 d H [ I 0 ] K A0 and π A = [ I 0 ]d H [ I 0 ] + ii Oherise, he opimal learning level I =, and π A = d H K A I 0 K A0 Proof of Proposiion A-3 We prove he resul in o ases and > Case : When, from Proposiion 5 and CAI = K A0 + K A I I 0, e have: π π A I = A I K A0 K A I I 0 if I [I 0, I ] πai [ I ]d H [ I ] K A0 K A I I 0 if I [maxi 0, I }, ], here I = Beause dπ A I = K di A, d πa I = 0, dπ di A I = d H [ ] di [ I ] K A, and d πa I > 0, e kno ha he opimal learning level I is eiher I di 0 or Noe ha I 0 I is equivalen o For For For I 0 and ha I 0 < πai = I 0 πai if I [I 0, I ] πai if I [I, ] <, π I 0 AI = πai for I [I 0, ] alays holds We have he folloing: I 0, by omparing π AI 0 = K A0 and π A = d H K A0 K A I 0, e kno ha if K A > K A, hen I = I 0 and πa = K A0 ; oherise, I =, and πa = d H K A I 0 K A0 For I 0 <, by omparing π AI 0 = [ I 0 ]d H [ I 0 ] K A0 and π A = d H K A0 K A I 0, e kno ha if K A > K A, hen I = I 0 and π A = [ I 0 ]d H [ I 0 ] K A0 ; oherise, I =, and πa = d H K A0 K A I 0

15 Huang e al: The Value of Bespoke Arile submied o ; manusrip no 5 Case : When >, from Proposiion 5 and CAI = K A0 + K A I I 0, e have: π 3 A I [ I ]d H [ I π AI = ] + I d H [ I ] K A0 K AI I 0 if I [I 0, I ] πa 4 I [ I ]d H [ I ] K A0 K AI I 0 if I [maxi 0, I }, ], here I = Noe ha π4 AI = πai in ase Beause dπ3 A I = d H [ di [ I] ] K [ I ] A, d πa 3 I > 0, dπ4 di A I = d H [ ] K di [ I ] A, and d πa 4 I > 0, e di kno ha he opimal learning level I is eiher I 0 or Noe ha I 0 I is equivalen o I 0 and ha For < For I 0 > alays holds We have:, π I 0 AI = πai 4 for I [I 0, ] πai = I 0 πai 3 if I [I 0, I ] πai 4 if I [I, ] For < I 0, by omparing π AI 0 = [ I 0 ]d H [ I 0 ] K A0 π A = d H K A0 K A I 0, e kno ha if K A > K A, hen I = I 0 and π A = [ I 0 ]d H [ I 0 ] K A0 ; oherise, I =, and πa = d H K A0 For and K A I 0 I 0 <, by omparing π AI 0 = [ I 0 ]d H [ I 0 ] + I 0 d H [ I 0 ] K A0 and π A = d H K A0 K A I 0, e kno ha if K A > K A3, hen I = I 0 and π A = [ I 0 ]d H [ I 0 ] + I 0 d H [ I 0 ] K A0 ; oherise, I =, and πa = d H K A I 0 K A0 Proposiion A-3 shos ha for any given, he opimal demand learning level is high hen he invesmen os K A is small, hih is onsisen ih our inuiion ha he opimal learning auray level is higher hen learning is less expensive We also observe ha demand learning level is high hen is inermediae When is small, he average demand d H is small, hih implies a high learning os per uni When is large, he demand unerainy is lo he oeffiien of variaion is lose o zero and hus he firm has less inenive o learn he demand Le K P, p d H J P 0 [ p p J P 0 ] and K P d H In preferene learning and Bespoke sysems, e may abuse noaion by riing J p = J and J AP = J henever no onfusion arises Proposiion A-4 In a preferene learning sysem ih linear oss, e have: i The opimal learning level J = J P 0, if and only if one of he folloing hree ondiions is saisfied: p, p < p and K J P 0 P > K P, p, 3 > p and K J P 0 P > K P The orresponding opimal profis are π P = K P 0, π P = π P = [ J P 0] p }d H K P 0 d H J P 0 p K P 0 ii Oherise, he opimal learning level J =, and π P = p d H K P 0 K P J P 0 and

16 Huang e al: The Value of Bespoke 6 Arile submied o ; manusrip no Proof of Proposiion A-4 From Proposiion A-, e have πp J K P 0 K P J J P 0 if < p π P J = πp J d H J p K P 0 K P J J P 0 if p p + J πp 3 J [ J] p}d H K P 0 K P J J P 0 if > p + J Beause dπ P J dj = K P, d π P J dj = 0, dπ P J dj = d H p J K P, d π P J dj = d H p J 3 > 0, dπ3 P J dj = d H K P, and d πp 3 J = 0, he opimal learning level J is eiher J dj P 0 or due o d πp i J 0 dj i =,, 3 Noe ha p o + J is equivalen o J p p J P 0 We have he folloing: If p, hen π P J = π P J for J P 0 J If p < If > p J P 0, hen π P J = πp J πp 3 J p J P 0, hen π P J = π 3 P J for J P 0 J By omparing π P J P 0 and π P in p, p < desired resuls if J P 0 J p if p < J and ha p J P 0 is equivalen p and > p J P 0 J P 0, e have he Nex, e presen he resul for a Bespoke sysem ih linear oss Le K AP, p d H p J AP 0 [ p J AP 0 ] and K AP d H Proposiion A-5 In a Bespoke sysem ih linear oss, e have: i The opimal learning level J = J AP 0, if and only if one of he folloing o ondiions is saisfied: p J AP 0 and K AP > K AP, p, p < J AP 0 and K AP > K AP The orresponding opimal profis are π AP = J AP 0 p ] K AP 0 d H J AP 0 p K AP 0 and π AP = d H [ ii Oherise, he opimal learning level J =, and π AP = d H p K AP 0 K AP J AP 0 Proof of Proposiion A-5 From Proposiion A-, e have π π AP J = AP J d H J p K AP 0 K AP J J AP 0 if p + J πap J d H [ J p] K AP 0 K AP J J AP 0 if > p + J Sine dπ AP J dj d π AP J dj = d H p J K AP, d π AP J dj = d H p J 3 > 0, dπ AP J dj = d H K AP, and = 0, he opimal learning level J is eiher J AP 0 or due o d π i AP J dj 0 i =, Noe ha p + J is equivalen o J p o p J AP 0 We have he folloing: and ha J AP 0 p is equivalen

17 Huang e al: The Value of Bespoke Arile submied o ; manusrip no 7 If p + J AP 0, hen π AP J = πap J πap J if J AP 0 J p if p < J If > p + J AP 0, hen π AP J = π AP J for J AP 0 J By omparing π AP J AP 0 and π AP in p + J AP 0 and > p + J AP 0, e have he desired resul Proposiions A-4-A-5 sho ha he opimal preferene learning auray levels in boh preferene and Bespoke sysems end o be high hen he invesmen oss K P and K AP are small We also observe ha he firm ends o hoose a high preferene learning auray level hen he expeed demand is high ie, a large and/or d H or hen p lo average invesmen os per-uni produ A small p is small This is beause a high expeed demand implies a relaively o learn he preferene Our numerial sudy also shos ha hen K P = K AP means a high margin and hus more inenive and J P 0 = J AP 0, he firm alays hooses a higher preferene learning auray level in he Bespoke sysem ompared o he preferene learning sysem A higher J means a higher invesmen os The firm s learning os may no pay off in he preferene learning sysem hen he realized demand is less han is produion quaniy, hile he firm never faes his senario in he Bespoke sysem beause i an exaly mah supply and demand Based on he resuls for all he four sysems, e are ready o invesigae he inerrelaionship beeen demand learning and preferene learning We are able o provide all he neessary and suffiien ondiions for demand learning and preferene learning being omplemens For breviy, hey are no presened here bu available from he auhors Online Supplemen 3 Analysis for 54: Impa of Salvage Value In his seion, e provide a represenaive numerial sudy o demonsrae he impa of salvage value s in he onex of sraegi usomers and bargain-huning onsumers From our numerial analysis, e observe ha as he salvage value s inreases, demand learning and preferene learning are more likely o be omplemenary regardless ho sraegi usomers are; see Figure A- for represenaive examples Sine here is no lefover invenory in boh demand learning and Bespoke sysems ie, heir profis are invarian ih respe o s, his observaion implies ha he radiional sysem benefis more from high salvage value ompared o he preferene learning sysem The explanaion is as follos Wih a higher salvage value, he firm ends o order a higher quaniy in boh radiional and preferene learning sysems In he preferene learning sysem, he firm has reahed he highes quaniy d H hen p is small and he highes prie i an harge even ih zero salvage value, hile i does no in he radiional sysem This implies ha he firm has more flexibiliy o adjus he prie and order quaniy o beer leverage he benefi from a higher salvage value in he radiional sysem Furhermore, ih he presene of sraegi behavior in he radiional sysem, a higher prie in period resuling from a higher salvage value redues usomers aiing inenives and hus inreases he firm s profi

18 Huang e al: The Value of Bespoke 8 Arile submied o ; manusrip no Figure A- The Impa of Salvage Value d H = 60, d L = 0, = 36, = 8, = 8 Online Supplemen 4 Analysis for 55: Correlaion Beeen Demand and Preferene Learnings Proposiion A-6 belo summarizes he resul for he demand learning sysem in hih onsumer preferene is also imperfely learned We follo he frameork in Seion 5 o model he imperfe preferene learning Wihou losing managerial insighs, e normalize d L o zero for analyial raabiliy in his seion J: Proposiion A-6 In a perfe demand learning sysem ih imperfe preferene learning of auray i If + J, he firm s opimal expeed profi π A = d H [ J ] ii If < + J, he firm s opimal expeed profi π A = d H J Proof of Proposiion A-6 In his sysem, he firm an observe he demand realizaion D before making is quaniy deision Hene, he firm an perfely mah supply ih demand Suppose D = d H, hen he number of onsumers ho ill purhase he produ is d H, if p [0, p] pd D A d H = H p, if p [ p, ] 0, if p [, + We hen disuss under eah of he hree prie ranges, ha he opimal prie should be If p [0, p], he firm profi is π dh p, q = p d H The opimal prie p d H = p and opimal quaniy q d H = d H p Hene, π d H = p d H if p + If p [ p, ], hen q d H = pd H p The firm profi π dh p, q = p pd H p Firs-order ondiion ih respe o prie p yields p p = 0 Hene, p = + For i being opimal, e require p = + [ p, ], hih is simplified o [, + p ] Hene, under his ondiion, e have p d H = +, q d H = d H p, and he opimal profi π d H = d H p If his ondiion does no hold, ie, > + p, e if

19 Huang e al: The Value of Bespoke Arile submied o ; manusrip no 9 have p d H = p, q d H = d H, and he opimal profi π d H = p d H The opimal expeed profi is π A = π d H Hene, e obain π A = dh p, if + p d H p, if + p Sine p = J, e an express he profi funion in erms of J I is expeed ha he firm s profi inreases as he preferene learning auray J inreases Proposiion A-7 belo summarizes he resul for he preferene learning sysem in hih demand is also imperfely learned We follo he frameork in Seion 5 o model he imperfe demand learning For exposiional onveniene, e inrodue he noaions: I A, p p and I B, p p Proposiion A-7 In a perfe preferene learning sysem ih imperfe demand learning of auray I: i If [0, p ] and I [I A, p, ], he firm s opimal expeed profi π p = d H [ I ] p } ii If [0, p ] and I [0, I A, p ], he firm s opimal expeed profi π p = 0 i If [ p, ] and I [I B, p, ], he firm s opimal expeed profi π p = d H [ I ] p } i If [ p, ] and I [0, I B, p ], he firm s opimal expeed profi π p = p d H Proof of Proposiion A-7 Wih perfe preferene learning, he firm also ses p = and all usomers ould like o buy a his prie The firm s expeed profi an be expressed as π p q = E[minD, q] p q Suppose ha he firm observes demand signal s = d H, hen is updaed demand Ds = d H = d H probabiliy H I I and Ds = d H = d L = 0 ih probabiliy H I = I So he profi π s=dh = H I p q When I I A, p ie, H I p 0, q s=d H = d H and π s=d H = H I p d H ; hen I < I A, p, q s=d H = 0 and π s=d H = 0 Suppose ha he firm observes demand signal s = d L, hen is updaed demand Ds = d L = d H probabiliy L I I and Ds = d L = d L = 0 ih probabiliy L I = I So he profi π s=dl = L I p q When I I B, p ie, L I p 0, q s=d L = d H and π s=d H = L I p d H ; hen I > I B, p, q s=d L = 0 and π s=d L = 0 To rie he firm s expeed profi before observing he demand signal s, e need o ompare I A, p p and I B, p p I urns ou I A, p I B, p is equivalen o p / Noe ha π pi, a = π s=d H + π s=d L Then, e an ompue his expeed profi based on differen ombinaions of he ondiions on and I For example, if [0, p ], e have I A, p 0 I B, p If e furher assume I [I A, p, ], hen π pi, = d H [ I ] p } Similarly, e obain he profi funions in oher ases ih ih One may argue ha alhough he firm imperfely learns he onsumer preferene ihou spending speifi effors in he demand learning sysem, he produion os migh sill inrease above We keep he produion os unhanged as in his seion in order o isolae he impa of he learning-mehanism orrelaion

20 Huang e al: The Value of Bespoke 0 Arile submied o ; manusrip no Online Supplemen 3 Responsive Priing In he basi model, e assumed ha he firm deermines he prie before unerain demand is realized or learned This assumpion is onsisen ih he exensive operaions lieraure see, eg, Peruzzi and Dada 999 and referenes herein This fis he senarios here he firm observes he demand afer selling he produ, or he firm needs o use he prie for adverisemen before observing he demand, as i is ypially he ase for many produs Hoever, in some seings, he firm may be able o se he prie afer he demand is realized or learned In his seion, e invesigae ho responsive/oningen priing may affe our main resul in he basi model We denoe q 0 and π 0 as he firm s opimal produion quaniy and opimal expeed profi respeively For he radiional sysem, he folloing proposiion haraerizes he equilibrium ouome Proposiion A-8 In a radiional sysem ih responsive priing: i When, e have: ia For, d L dh q 0 = and d + π 0 = ; H d + L d H d L ib For >, d L dh q 0 = dh and π 0 = d H + d L ; i For any given demand realizaion D and produion quaniy q, he opimal prie pq, D = and pq, D = q D oherise ii When >, le + d L iia For min,, q 0 = d H + d L d + H, + d L d H, and e have: and π 0 = d L iib If <, hen for < <, q0 = d L and π0 = d L d L d H + d L ; ii If >, hen for < <, π 0 = max orresponding q 0 = d + H d L iid For > max,, q 0 = d H or q 0 = d H ; ; + d H d L d + H d L D if q, d H + d L, and he and π 0 = d H + d L; iie For any given demand realizaion D and produion quaniy q, he opimal prie pq, D = if q D, and pq, D = q D oherise Proof of Proposiion A-8 We ork bakards: firs solve for he opimal prie pq, D given produion quaniy q and demand realizaion D, and hen solve for q 0 Clearly, for he opimal prie p e have p, and he demand is p D The firm profi an be expressed as πp; q, D = p minq, p D} q When p q p, e have q D and πp; q, D = p q In his ase, πp; q, D is maximized a D pq, D = q D q q and he resuling profi is πq, D = q When p > D D and πp; q, D = D p p q In his ase, πp; q, D is maximized a pq, D = is πq, D = D q D p, e have q > D and he resuling profi q Noe ha q is equivalen o D q, is equivalen o, D is equivalen o D q, q > q is equivalen o >, q D < and < Nex, e disuss pq, D in differen regions When >, e have q > q a For D < q, e have q < < q < Beause <, e kno ha for q p, e have p > and D D D hus πp; q, D = D p p q The opimal prie pq, D = and he resuling profi πq, D = D b For q D q, e have < q D < < Combining he opimal prie pq, D for p q

21 Huang e al: The Value of Bespoke Arile submied o ; manusrip no q q and for p, e have ha for he hole range p, he opimal prie pq, D = D D and he resuling profi πq, D = D he opimal prie pq, D for p q D for q p, he opimal prie pq, D = D q For D > q, e have < < q < Combining D q and for p, e have ha for he hole range D q and he resuling profi πq, D = q D Combining resuls a-, e have he laim i in Proposiion A-8 as q q D ih πq, D = q if q D D pq, D =, ih πq, D = D q if q > D A-0 Finally, e solve for he opimal quaniy q 0 Noe D is eiher d H ih probabiliy, or d L ih probabiliy We disuss hree possible regions: q d L, d L < q d H and q > d H Based on equaion A-0, e ompue π q = E[πq, D] by aking he expeaion of πq, D over D: π q = q q d H + d L if q d L, π q = π q = q d H q + d L q if d L < q d H, π 3 q = µ D q if q > d H Noie ha π q is onave and maximized a q = d H, and π 3 q dereases in q Beause equivalen o >, and dh d L dh d + H d L d + H d L, π q is onave and maximized a q = d L is equivalen o d L, dh dh > d L d H is alays rue, e have he folloing resul If d L dh, hen q0 = By subsiuing d + q 0 = ino π H d + q, e have π0 = If >, L d H d + L d H d d L L dh hen q0 = d H By subsiuing q 0 = d H ino π q, e have π0 = d H + d L We have finished he proof for laim i in Proposiion A-8 In a similar ay, e an derive he resul for laim ii > Wih responsive priing, he profi in radiional sysem should be higher han ha in he basi model ihou responsive priing The prie influenes he number of usomers ho an o buy he produ Wih responsive priing, he firm an se he prie based on he realized demand o beer mah he number of usomers ho an o buy he produ ih he produion quaniy and hus improves is profi Oher han he radiional sysem, he opimal sraegy and profi do no hange in he oher hree sysems: Regardless of he demand realizaion, he firm alays ses p = in preferene learning and Bespoke sysems, and alays ses p = + in demand learning sysem The equilibrium resul for he demand learning sysem as saed in Proposiion A-9 belo is he same as ha in he basi model bu he proof is slighly differen due o responsive priing is + Proposiion A-9 In a demand learning sysem ih responsive priing: The firm s opimal prie p A =, opimal produion quaniy q AD = D for a demand realizaion D, and opimal profi π A = µ D Proof of Proposiion A-9 The firm deermines is prie p and produion quaniy q afer learning he realizaion of D Obviously, p, and in his ase he number of usomers ho an o buy he

22 Huang e al: The Value of Bespoke Arile submied o ; manusrip no Figure A- The Inerrelaionship of Demand Learning and Preferene Learning ih Responsive Priing d H = 60, d L = 0, = 30, = 9, = produ is p D The profi an be expressed as πp, q; D = p minq, p D} q I is easy o sho ha he firm alays ses q = p D for any p and D Then, he profi an be expressed as πp; D = p min p D, p D} p D = D pp hih is maximized a p = +, and he orresponding opimal quaniy is q AD = opimal profi given D is D and he expeed opimal profi is µ D + D = D The We expe ha demand learning and preferene learning are more likely o be omplemenary Numerial resuls onfirm our inuiion, as shon in Figure A- here e use he same se of parameers as in Figure for he basi model Furhermore, onsisen ih our basi model, demand learning and preferene learning are omplemenary hen he os of usomizaion is lo and he probabiliy of having high demand is large Online Supplemen 4 Conneion o Produ Line Design The model seup follos he basi model in he main body of he paper Suppose usomers/onsumers are uniformly disribued on a Hoelling line [ /, /] There is a monopolis firm selling n horizonally differeniaed produs, hih are posiioned a loaion θ i, i,,, n} on he Hoelling line, ih orresponding pries p i, i,,, n} Wihou loss of generaliy, e assume ha θ θ θ n Consumers ge uiliy from he produ, and inur he ravelling os and pay he prie The per uni ravelling os is denoed as If a onsumer a loaion θ buys produ i, her ne uiliy is θ θ i p i A onsumer buys he produ ha gives her he highes non-negaive ne uiliy Consumers may subsiue for anoher produ only if he produ also gives her a non-negaive ne uiliy if her favorable produ is ou of sok The firm s per uni produion os is For any given number of produs n, he firm needs o deermine boh pries and soking levels for all he produs

23 Huang e al: The Value of Bespoke Arile submied o ; manusrip no 3 Lemma A- In any opimal sraegy, all produs are sold a he same prie, no onsumer ges srily posiive uiliy from muliple produs, and onsumers on he boundaries -/ and / ge non-posiive uiliy from purhasing Proof of Lemma A- Define he marke overage of a produ as he onsumers ho ge non-negaive uiliies from purhasing his produ Clearly, he marke overage of any produ is oniguous in ha if usomers a boh loaions buy produ i, hen any usomer loaed beeen hese o loaions buys produ i Firs, here is no produ overage overlapping in any opimal sraegy Oherise, here are some usomers ho ge posiive uiliies from muliple produs Then he firm an inrease he produ prie or shif he produ loaion o inrease profi To see his, suppose produ i and j have overlapping overage, hen if one produ ihou loss of generaliy, assuming i is i only overlaps on one side ih he oher one, hen e an inrease p i and shif he posiion of produ i aay from produ j o inrease he profi If boh produs has overage overlapping on boh sides, hen e an repea he argumen and sho ha all produs have overlapping on boh sides, and usomers a boh ends ge srily posiive uiliy from purhasing Then, e an inrease all pries by he same amoun ihou hanging demands for all produs, hih ill inrease he oal profi Therefore, in any opimal sraegy, here should be no produ overage overlapping Seond, a he opimal sraegy, usomers a boh ends -/ and / ge non-posiive uiliy from purhasing Suppose no, hen a usomer a one end ihou loss of generaliy, assuming loaion 0 a srily posiive uiliy from purhasing produ produ is loses o loaion 0, hen he firm an inrease p o p + ɛ here ɛ is an suffiienly small amoun and inrease θ o θ + ɛ Then, he demands are he same and he profi is higher Given here is no overage overlapping and usomers a boh ends ge non-posiive uiliy from purhasing, L i define as L i = p i / is he overage of produ i and DL i is he demand for produ i Then, he invenory problem for produ i is max q i min d H L i, q i } p i + min d L L i, q i } p i q i Solving his nesvendor problem e have he opimal invenory d H L i if p i q i = d L L i if p i < Subsiuing q i bak, e have he opimal profi d H + d L p i p p i d i H if p i π i p i = p d i L p i if p i < Third, all produs mus have he same prie Suppose no, hen here exis o adjaen produs ih differen pries Tha is, here exiss a produ i, suh ha p i p i+ Wihou loss of generaliy, assume p i > p i+ Then L i < L i+ and he oal overage is L i + L i+ = p i p i+ / As long as e keep he

24 Huang e al: The Value of Bespoke 4 Arile submied o ; manusrip no oal prie p i + p i+ he same, e an keep he same oal overage by he o produs L i + L i+ e may need o shif he loaions and do no affe oher produs We an verify ha π i p i is onave in he o domains [, and [/,, so if p i > p i+ / or p i+ < p i < /, e an replae p i and p i+ ih heir average p i + p i+ / and reloae he o produs so ha he oal overage remains he same and e ge a higher profi If p i > / and p i+ < /, hen π p i+ = d L + p i+, and π p i = µ D sho ha π p i+ = π p i 0 an never be he ase if p i > / and p i+ < / Solving π p i+ = π p i gives us p i+ = d H d L d L + µ D d L p i p i + d H We Simplifying π p i 0, e have p i µ D+d H µ D Then p i+ = d H d L d L + µ D d L p i d H d L d L + µ D +d H d L = Therefore, π p i+ = d L + p i+ d L + = 4d L > 0, hih onra- dis ih π p i+ 0 Hene, π p i+ = π p i 0 anno be he ase given p i > / and p i+ < / Then, here remain o possibiliies: π p i+ π p i or π p i+ = π p i > 0 We nex onsider hese o ases one by one Case : π p i+ π p i If π p i+ > π p i, e an inrease p i+ by ɛ and redue p i by ɛ o inrease he profis from he o produs ihou affeing oher produs Similarly, if π p i+ < π p i, e an redue p i+ by ɛ and inrease p i by ɛ o inrease he profis from he o produs ihou affeing oher produs Case : π p i+ = π p i > 0 We an hen inrease boh p i and p i+ o inrease he profis from he o produs Wih he higher pries, he overage of he o produs ill be smaller, and hus on affe oher produs Therefore, in any opimal sraegy, i mus be ha all pries are equal Then all produs should have equal overage ih no overlapping and hus equal demand, and heir opimal invenory levels are he same Then e only need o deide one prie for all produs, and he oal invenory level for all produs Beause here should be no overlapping, eah produ overs a mos /n on he Hoelling line Therefore, solving p i / /n, he opimal prie mus saisfy p i /n The nex proposiion shos ha a he opimal sraegy, offering n produs is equivalen o reduing he ravelling os o /n Proposiion A-0 Given any fixed and, denoe he oal profi fp, n, as a funion of prie p, number of produs n, per-uni raveling os under he opimal produ posiioning and invenory sraegy Then, e have fp, n, = fp,, n Proof of Proposiion A-0 In he opimal sraegy, all n produs have he same prie and he same soking level Denoe he opimal oal soking level given prie p as q p Sine all produs should have he same soking level, he opimal soking level for eah produ is q p /n The demand for eah produ is D p / Then, he oal expeed profi from all produs is [ D p f p, n, = ne min, q p }] p q p n

25 Huang e al: The Value of Bespoke Arile submied o ; manusrip no 5 [ }] D p = E min, q p p q p /n Subsiuing n = and = /n e have fp,, }] D p [min n = E, q p p q p /n Clearly, fp, n, = fp,, n Mapping he number of produs n o he level of usomizaion J Proposiion A-0 allos us o esablish a one-o-one mapping of he number of produs n in his seion o he level of usomizaion/personalizaion J in he main body of he paper When he firm offers n produs, he redued ravelling os is p = /n Then he level of usomizaion is J = p / = /n In he exreme, hen n =, he level of usomizaion is zero; hereas as n goes o infiniy, he level of usomizaion approahes one Due o his equivalene, hoosing he number of produs n is he same as deermining he level of usomizaion J Therefore, he produ line design inerpreaion jusifies our sylized model of using J for preferene learning and personalizaion

Amit Mehra. Indian School of Business, Hyderabad, INDIA Vijay Mookerjee

Amit Mehra. Indian School of Business, Hyderabad, INDIA Vijay Mookerjee RESEARCH ARTICLE HUMAN CAPITAL DEVELOPMENT FOR PROGRAMMERS USING OPEN SOURCE SOFTWARE Ami Mehra Indian Shool of Business, Hyderabad, INDIA {Ami_Mehra@isb.edu} Vijay Mookerjee Shool of Managemen, Uniersiy

More information

Economics 202 (Section 05) Macroeconomic Theory Practice Problem Set 7 Suggested Solutions Professor Sanjay Chugh Fall 2013

Economics 202 (Section 05) Macroeconomic Theory Practice Problem Set 7 Suggested Solutions Professor Sanjay Chugh Fall 2013 Deparmen of Eonomis Boson College Eonomis 0 (Seion 05) Maroeonomi Theory Praie Problem Se 7 Suggesed Soluions Professor Sanjay Chugh Fall 03. Lags in Labor Hiring. Raher han supposing ha he represenaive

More information

Mathematical Foundations -1- Choice over Time. Choice over time. A. The model 2. B. Analysis of period 1 and period 2 3

Mathematical Foundations -1- Choice over Time. Choice over time. A. The model 2. B. Analysis of period 1 and period 2 3 Mahemaial Foundaions -- Choie over Time Choie over ime A. The model B. Analysis of period and period 3 C. Analysis of period and period + 6 D. The wealh equaion 0 E. The soluion for large T 5 F. Fuure

More information

Online Appendix for "Customer Recognition in. Experience versus Inspection Good Markets"

Online Appendix for Customer Recognition in. Experience versus Inspection Good Markets Online Appendix for "Cusomer Recogniion in Experience versus Inspecion Good Markes" Bing Jing Cheong Kong Graduae School of Business Beijing, 0078, People s Republic of China, bjing@ckgsbeducn November

More information

Linear Quadratic Regulator (LQR) - State Feedback Design

Linear Quadratic Regulator (LQR) - State Feedback Design Linear Quadrai Regulaor (LQR) - Sae Feedbak Design A sysem is expressed in sae variable form as x = Ax + Bu n m wih x( ) R, u( ) R and he iniial ondiion x() = x A he sabilizaion problem using sae variable

More information

An Inventory Model for Weibull Time-Dependence. Demand Rate with Completely Backlogged. Shortages

An Inventory Model for Weibull Time-Dependence. Demand Rate with Completely Backlogged. Shortages Inernaional Mahemaial Forum, 5, 00, no. 5, 675-687 An Invenory Model for Weibull Time-Dependene Demand Rae wih Compleely Baklogged Shorages C. K. Tripahy and U. Mishra Deparmen of Saisis, Sambalpur Universiy

More information

Problem Set 9 Due December, 7

Problem Set 9 Due December, 7 EE226: Random Proesses in Sysems Leurer: Jean C. Walrand Problem Se 9 Due Deember, 7 Fall 6 GSI: Assane Gueye his problem se essenially reviews Convergene and Renewal proesses. No all exerises are o be

More information

Expert Advice for Amateurs

Expert Advice for Amateurs Exper Advice for Amaeurs Ernes K. Lai Online Appendix - Exisence of Equilibria The analysis in his secion is performed under more general payoff funcions. Wihou aking an explici form, he payoffs of he

More information

Problem 1 / 25 Problem 2 / 10 Problem 3 / 15 Problem 4 / 30 Problem 5 / 20 TOTAL / 100

Problem 1 / 25 Problem 2 / 10 Problem 3 / 15 Problem 4 / 30 Problem 5 / 20 TOTAL / 100 Deparmen of Applied Eonomis Johns Hopkins Universiy Eonomis 60 Maroeonomi Theory and Poliy Miderm Exam Suggesed Soluions Professor Sanjay Chugh Summer 0 NAME: The Exam has a oal of five (5) problems and

More information

AN INVENTORY MODEL FOR DETERIORATING ITEMS WITH EXPONENTIAL DECLINING DEMAND AND PARTIAL BACKLOGGING

AN INVENTORY MODEL FOR DETERIORATING ITEMS WITH EXPONENTIAL DECLINING DEMAND AND PARTIAL BACKLOGGING Yugoslav Journal of Operaions Researh 5 (005) Number 77-88 AN INVENTORY MODEL FOR DETERIORATING ITEMS WITH EXPONENTIAL DECLINING DEMAND AND PARTIAL BACKLOGGING Liang-Yuh OUYANG Deparmen of Managemen Sienes

More information

mywbut.com Lesson 11 Study of DC transients in R-L-C Circuits

mywbut.com Lesson 11 Study of DC transients in R-L-C Circuits mywbu.om esson Sudy of DC ransiens in R--C Ciruis mywbu.om Objeives Be able o wrie differenial equaion for a d iruis onaining wo sorage elemens in presene of a resisane. To develop a horough undersanding

More information

SIMULATION STUDY OF STOCHASTIC CHANNEL REDISTRIBUTION

SIMULATION STUDY OF STOCHASTIC CHANNEL REDISTRIBUTION Developmens in Business Simulaion and Experienial Learning, Volume 3, 3 SIMULATIO STUDY OF STOCHASTIC CHAEL REDISTRIBUTIO Yao Dong-Qing Towson Universiy dyao@owson.edu ABSTRACT In his paper, we invesigae

More information

Solutions to Exercises in Chapter 5

Solutions to Exercises in Chapter 5 in 5. (a) The required inerval is b ± se( ) b where b = 4.768, =.4 and se( b ) =.39. Tha is 4.768 ±.4.39 = ( 4.4, 88.57) We esimae ha β lies beween 4.4 and 85.57. In repeaed samples 95% of similarly onsrued

More information

New Oscillation Criteria For Second Order Nonlinear Differential Equations

New Oscillation Criteria For Second Order Nonlinear Differential Equations Researh Inveny: Inernaional Journal Of Engineering And Siene Issn: 78-47, Vol, Issue 4 (Feruary 03), Pp 36-4 WwwResearhinvenyCom New Osillaion Crieria For Seond Order Nonlinear Differenial Equaions Xhevair

More information

The Role of Money: Credible Asset or Numeraire? Masayuki Otaki (Institute of Social Science, University of Tokyo)

The Role of Money: Credible Asset or Numeraire? Masayuki Otaki (Institute of Social Science, University of Tokyo) DBJ Disussion Paper Series, No.04 The Role of Money: Credible Asse or Numeraire? Masayuki Oaki (Insiue of Soial Siene, Universiy of Tokyo) January 0 Disussion Papers are a series of preliminary maerials

More information

Math 2142 Exam 1 Review Problems. x 2 + f (0) 3! for the 3rd Taylor polynomial at x = 0. To calculate the various quantities:

Math 2142 Exam 1 Review Problems. x 2 + f (0) 3! for the 3rd Taylor polynomial at x = 0. To calculate the various quantities: Mah 4 Eam Review Problems Problem. Calculae he 3rd Taylor polynomial for arcsin a =. Soluion. Le f() = arcsin. For his problem, we use he formula f() + f () + f ()! + f () 3! for he 3rd Taylor polynomial

More information

The Relativistic Field of a Rotating Body

The Relativistic Field of a Rotating Body The Relaivisi Field of a Roaing Body Panelis M. Pehlivanides Alani IKE, Ahens 57, Greee ppexl@eemail.gr Absra Based on he pahs of signals emanaing from a roaing poin body, e find he equaions and properies

More information

Seminar 4: Hotelling 2

Seminar 4: Hotelling 2 Seminar 4: Hoelling 2 November 3, 211 1 Exercise Par 1 Iso-elasic demand A non renewable resource of a known sock S can be exraced a zero cos. Demand for he resource is of he form: D(p ) = p ε ε > A a

More information

Generalized electromagnetic energy-momentum tensor and scalar curvature of space at the location of charged particle

Generalized electromagnetic energy-momentum tensor and scalar curvature of space at the location of charged particle Generalized eleromagnei energy-momenum ensor and salar urvaure of spae a he loaion of harged parile A.L. Kholmeskii 1, O.V. Missevih and T. Yarman 3 1 Belarus Sae Universiy, Nezavisimosi Avenue, 0030 Minsk,

More information

HOTELLING LOCATION MODEL

HOTELLING LOCATION MODEL HOTELLING LOCATION MODEL THE LINEAR CITY MODEL The Example of Choosing only Locaion wihou Price Compeiion Le a be he locaion of rm and b is he locaion of rm. Assume he linear ransporaion cos equal o d,

More information

T L. t=1. Proof of Lemma 1. Using the marginal cost accounting in Equation(4) and standard arguments. t )+Π RB. t )+K 1(Q RB

T L. t=1. Proof of Lemma 1. Using the marginal cost accounting in Equation(4) and standard arguments. t )+Π RB. t )+K 1(Q RB Elecronic Companion EC.1. Proofs of Technical Lemmas and Theorems LEMMA 1. Le C(RB) be he oal cos incurred by he RB policy. Then we have, T L E[C(RB)] 3 E[Z RB ]. (EC.1) Proof of Lemma 1. Using he marginal

More information

Lecture Notes 3: Quantitative Analysis in DSGE Models: New Keynesian Model

Lecture Notes 3: Quantitative Analysis in DSGE Models: New Keynesian Model Lecure Noes 3: Quaniaive Analysis in DSGE Models: New Keynesian Model Zhiwei Xu, Email: xuzhiwei@sju.edu.cn The moneary policy plays lile role in he basic moneary model wihou price sickiness. We now urn

More information

SOLUTIONS TO ECE 3084

SOLUTIONS TO ECE 3084 SOLUTIONS TO ECE 384 PROBLEM 2.. For each sysem below, specify wheher or no i is: (i) memoryless; (ii) causal; (iii) inverible; (iv) linear; (v) ime invarian; Explain your reasoning. If he propery is no

More information

CHERNOFF DISTANCE AND AFFINITY FOR TRUNCATED DISTRIBUTIONS *

CHERNOFF DISTANCE AND AFFINITY FOR TRUNCATED DISTRIBUTIONS * haper 5 HERNOFF DISTANE AND AFFINITY FOR TRUNATED DISTRIBUTIONS * 5. Inroducion In he case of disribuions ha saisfy he regulariy condiions, he ramer- Rao inequaliy holds and he maximum likelihood esimaor

More information

Errata (1 st Edition)

Errata (1 st Edition) P Sandborn, os Analysis of Elecronic Sysems, s Ediion, orld Scienific, Singapore, 03 Erraa ( s Ediion) S K 05D Page 8 Equaion (7) should be, E 05D E Nu e S K he L appearing in he equaion in he book does

More information

Second-Order Boundary Value Problems of Singular Type

Second-Order Boundary Value Problems of Singular Type JOURNAL OF MATEMATICAL ANALYSIS AND APPLICATIONS 226, 4443 998 ARTICLE NO. AY98688 Seond-Order Boundary Value Probles of Singular Type Ravi P. Agarwal Deparen of Maheais, Naional Uniersiy of Singapore,

More information

Teacher Quality Policy When Supply Matters: Online Appendix

Teacher Quality Policy When Supply Matters: Online Appendix Teaher Qualiy Poliy When Supply Maers: Online Appendix Jesse Rohsein July 24, 24 A Searh model Eah eaher draws a single ouside job offer eah year. If she aeps he offer, she exis eahing forever. The ouside

More information

Problem Set 5. Graduate Macro II, Spring 2017 The University of Notre Dame Professor Sims

Problem Set 5. Graduate Macro II, Spring 2017 The University of Notre Dame Professor Sims Problem Se 5 Graduae Macro II, Spring 2017 The Universiy of Nore Dame Professor Sims Insrucions: You may consul wih oher members of he class, bu please make sure o urn in your own work. Where applicable,

More information

(Radiation Dominated) Last Update: 21 June 2006

(Radiation Dominated) Last Update: 21 June 2006 Chaper Rik s Cosmology uorial: he ime-emperaure Relaionship in he Early Universe Chaper he ime-emperaure Relaionship in he Early Universe (Radiaion Dominaed) Las Updae: 1 June 006 1. Inroduion n In Chaper

More information

5. An economic understanding of optimal control as explained by Dorfman (1969) AGEC

5. An economic understanding of optimal control as explained by Dorfman (1969) AGEC This doumen was generaed a 1:27 PM, 09/17/15 Copyrigh 2015 Rihard T Woodward 5 An eonomi undersanding of opimal onrol as explained by Dorfman (1969) AGEC 642-2015 The purpose of his leure and he nex is

More information

T. J. HOLMES AND T. J. KEHOE INTERNATIONAL TRADE AND PAYMENTS THEORY FALL 2011 EXAMINATION

T. J. HOLMES AND T. J. KEHOE INTERNATIONAL TRADE AND PAYMENTS THEORY FALL 2011 EXAMINATION ECON 841 T. J. HOLMES AND T. J. KEHOE INTERNATIONAL TRADE AND PAYMENTS THEORY FALL 211 EXAMINATION This exam has wo pars. Each par has wo quesions. Please answer one of he wo quesions in each par for a

More information

κt π = (5) T surrface k BASELINE CASE

κt π = (5) T surrface k BASELINE CASE II. BASELINE CASE PRACICAL CONSIDERAIONS FOR HERMAL SRESSES INDUCED BY SURFACE HEAING James P. Blanhard Universi of Wisonsin Madison 15 Engineering Dr. Madison, WI 5376-169 68-63-391 blanhard@engr.is.edu

More information

Inventory Analysis and Management. Multi-Period Stochastic Models: Optimality of (s, S) Policy for K-Convex Objective Functions

Inventory Analysis and Management. Multi-Period Stochastic Models: Optimality of (s, S) Policy for K-Convex Objective Functions Muli-Period Sochasic Models: Opimali of (s, S) Polic for -Convex Objecive Funcions Consider a seing similar o he N-sage newsvendor problem excep ha now here is a fixed re-ordering cos (> 0) for each (re-)order.

More information

Competitive and Cooperative Inventory Policies in a Two-Stage Supply-Chain

Competitive and Cooperative Inventory Policies in a Two-Stage Supply-Chain Compeiive and Cooperaive Invenory Policies in a Two-Sage Supply-Chain (G. P. Cachon and P. H. Zipkin) Presened by Shruivandana Sharma IOE 64, Supply Chain Managemen, Winer 2009 Universiy of Michigan, Ann

More information

Mass Transfer Coefficients (MTC) and Correlations I

Mass Transfer Coefficients (MTC) and Correlations I Mass Transfer Mass Transfer Coeffiiens (MTC) and Correlaions I 7- Mass Transfer Coeffiiens and Correlaions I Diffusion an be desribed in wo ways:. Deailed physial desripion based on Fik s laws and he diffusion

More information

ECON Lecture 4 (OB), Sept. 14, 2010

ECON Lecture 4 (OB), Sept. 14, 2010 ECON4925 21 Leure 4 (OB), Sep. 14, 21 Exraion under imperfe ompeiion: monopoly, oligopoly and he arel-fringe model Perman e al. (23), Ch. 15.6; Salan (1976) 2 MONOPOLISTIC EXPLOITATION OF A NATURAL RESOURCE

More information

Hybrid probabilistic interval dynamic analysis of vehicle-bridge interaction system with uncertainties

Hybrid probabilistic interval dynamic analysis of vehicle-bridge interaction system with uncertainties 1 APCOM & SCM 11-14 h Deember, 13, Singapore Hybrid probabilisi inerval dynami analysis of vehile-bridge ineraion sysem wih unerainies Nengguang iu 1, * Wei Gao 1, Chongmin Song 1 and Nong Zhang 1 Shool

More information

Notes for Lecture 17-18

Notes for Lecture 17-18 U.C. Berkeley CS278: Compuaional Complexiy Handou N7-8 Professor Luca Trevisan April 3-8, 2008 Noes for Lecure 7-8 In hese wo lecures we prove he firs half of he PCP Theorem, he Amplificaion Lemma, up

More information

0.1 MAXIMUM LIKELIHOOD ESTIMATION EXPLAINED

0.1 MAXIMUM LIKELIHOOD ESTIMATION EXPLAINED 0.1 MAXIMUM LIKELIHOOD ESTIMATIO EXPLAIED Maximum likelihood esimaion is a bes-fi saisical mehod for he esimaion of he values of he parameers of a sysem, based on a se of observaions of a random variable

More information

Essential Microeconomics : OPTIMAL CONTROL 1. Consider the following class of optimization problems

Essential Microeconomics : OPTIMAL CONTROL 1. Consider the following class of optimization problems Essenial Microeconomics -- 6.5: OPIMAL CONROL Consider he following class of opimizaion problems Max{ U( k, x) + U+ ( k+ ) k+ k F( k, x)}. { x, k+ } = In he language of conrol heory, he vecor k is he vecor

More information

Supplement for Stochastic Convex Optimization: Faster Local Growth Implies Faster Global Convergence

Supplement for Stochastic Convex Optimization: Faster Local Growth Implies Faster Global Convergence Supplemen for Sochasic Convex Opimizaion: Faser Local Growh Implies Faser Global Convergence Yi Xu Qihang Lin ianbao Yang Proof of heorem heorem Suppose Assumpion holds and F (w) obeys he LGC (6) Given

More information

Nevertheless, there are well defined (and potentially useful) distributions for which σ 2

Nevertheless, there are well defined (and potentially useful) distributions for which σ 2 M. Meseron-Gibbons: Bioalulus, Leure, Page. The variane. More on improper inegrals In general, knowing only he mean of a isribuion is no as useful as also knowing wheher he isribuion is lumpe near he mean

More information

Chapter 2. First Order Scalar Equations

Chapter 2. First Order Scalar Equations Chaper. Firs Order Scalar Equaions We sar our sudy of differenial equaions in he same way he pioneers in his field did. We show paricular echniques o solve paricular ypes of firs order differenial equaions.

More information

Optimal Transform: The Karhunen-Loeve Transform (KLT)

Optimal Transform: The Karhunen-Loeve Transform (KLT) Opimal ransform: he Karhunen-Loeve ransform (KL) Reall: We are ineresed in uniary ransforms beause of heir nie properies: energy onservaion, energy ompaion, deorrelaion oivaion: τ (D ransform; assume separable)

More information

Section 3.5 Nonhomogeneous Equations; Method of Undetermined Coefficients

Section 3.5 Nonhomogeneous Equations; Method of Undetermined Coefficients Secion 3.5 Nonhomogeneous Equaions; Mehod of Undeermined Coefficiens Key Terms/Ideas: Linear Differenial operaor Nonlinear operaor Second order homogeneous DE Second order nonhomogeneous DE Soluion o homogeneous

More information

3. Differential Equations

3. Differential Equations 3. Differenial Equaions 3.. inear Differenial Equaions of Firs rder A firs order differenial equaion is an equaion of he form d() d ( ) = F ( (),) (3.) As noed above, here will in general be a whole la

More information

Lecture 20: Riccati Equations and Least Squares Feedback Control

Lecture 20: Riccati Equations and Least Squares Feedback Control 34-5 LINEAR SYSTEMS Lecure : Riccai Equaions and Leas Squares Feedback Conrol 5.6.4 Sae Feedback via Riccai Equaions A recursive approach in generaing he marix-valued funcion W ( ) equaion for i for he

More information

A state space approach to calculating the Beveridge Nelson decomposition

A state space approach to calculating the Beveridge Nelson decomposition Eonomis Leers 75 (00) 3 7 www.elsevier.om/ loae/ eonbase A sae spae approah o alulaing he Beveridge Nelson deomposiion James C. Morley* Deparmen of Eonomis, Washingon Universiy, Campus Box 08, Brookings

More information

Chapter 7: Solving Trig Equations

Chapter 7: Solving Trig Equations Haberman MTH Secion I: The Trigonomeric Funcions Chaper 7: Solving Trig Equaions Le s sar by solving a couple of equaions ha involve he sine funcion EXAMPLE a: Solve he equaion sin( ) The inverse funcions

More information

The Asymptotic Behavior of Nonoscillatory Solutions of Some Nonlinear Dynamic Equations on Time Scales

The Asymptotic Behavior of Nonoscillatory Solutions of Some Nonlinear Dynamic Equations on Time Scales Advances in Dynamical Sysems and Applicaions. ISSN 0973-5321 Volume 1 Number 1 (2006, pp. 103 112 c Research India Publicaions hp://www.ripublicaion.com/adsa.hm The Asympoic Behavior of Nonoscillaory Soluions

More information

Examples of Dynamic Programming Problems

Examples of Dynamic Programming Problems M.I.T. 5.450-Fall 00 Sloan School of Managemen Professor Leonid Kogan Examples of Dynamic Programming Problems Problem A given quaniy X of a single resource is o be allocaed opimally among N producion

More information

Economic Growth & Development: Part 4 Vertical Innovation Models. By Kiminori Matsuyama. Updated on , 11:01:54 AM

Economic Growth & Development: Part 4 Vertical Innovation Models. By Kiminori Matsuyama. Updated on , 11:01:54 AM Economic Growh & Developmen: Par 4 Verical Innovaion Models By Kiminori Masuyama Updaed on 20-04-4 :0:54 AM Page of 7 Inroducion In he previous models R&D develops producs ha are new ie imperfec subsiues

More information

Solutions Problem Set 3 Macro II (14.452)

Solutions Problem Set 3 Macro II (14.452) Soluions Problem Se 3 Macro II (14.452) Francisco A. Gallego 04/27/2005 1 Q heory of invesmen in coninuous ime and no uncerainy Consider he in nie horizon model of a rm facing adjusmen coss o invesmen.

More information

Two Coupled Oscillators / Normal Modes

Two Coupled Oscillators / Normal Modes Lecure 3 Phys 3750 Two Coupled Oscillaors / Normal Modes Overview and Moivaion: Today we ake a small, bu significan, sep owards wave moion. We will no ye observe waves, bu his sep is imporan in is own

More information

Math 334 Fall 2011 Homework 11 Solutions

Math 334 Fall 2011 Homework 11 Solutions Dec. 2, 2 Mah 334 Fall 2 Homework Soluions Basic Problem. Transform he following iniial value problem ino an iniial value problem for a sysem: u + p()u + q() u g(), u() u, u () v. () Soluion. Le v u. Then

More information

Online Appendix to Solution Methods for Models with Rare Disasters

Online Appendix to Solution Methods for Models with Rare Disasters Online Appendix o Soluion Mehods for Models wih Rare Disasers Jesús Fernández-Villaverde and Oren Levinal In his Online Appendix, we presen he Euler condiions of he model, we develop he pricing Calvo block,

More information

Macroeconomic Theory Ph.D. Qualifying Examination Fall 2005 ANSWER EACH PART IN A SEPARATE BLUE BOOK. PART ONE: ANSWER IN BOOK 1 WEIGHT 1/3

Macroeconomic Theory Ph.D. Qualifying Examination Fall 2005 ANSWER EACH PART IN A SEPARATE BLUE BOOK. PART ONE: ANSWER IN BOOK 1 WEIGHT 1/3 Macroeconomic Theory Ph.D. Qualifying Examinaion Fall 2005 Comprehensive Examinaion UCLA Dep. of Economics You have 4 hours o complee he exam. There are hree pars o he exam. Answer all pars. Each par has

More information

Boyce/DiPrima 9 th ed, Ch 6.1: Definition of. Laplace Transform. In this chapter we use the Laplace transform to convert a

Boyce/DiPrima 9 th ed, Ch 6.1: Definition of. Laplace Transform. In this chapter we use the Laplace transform to convert a Boye/DiPrima 9 h ed, Ch 6.: Definiion of Laplae Transform Elemenary Differenial Equaions and Boundary Value Problems, 9 h ediion, by William E. Boye and Rihard C. DiPrima, 2009 by John Wiley & Sons, In.

More information

1 Consumption and Risky Assets

1 Consumption and Risky Assets Soluions o Problem Se 8 Econ 0A - nd Half - Fall 011 Prof David Romer, GSI: Vicoria Vanasco 1 Consumpion and Risky Asses Consumer's lifeime uiliy: U = u(c 1 )+E[u(c )] Income: Y 1 = Ȳ cerain and Y F (

More information

Reserves measures have an economic component eg. what could be extracted at current prices?

Reserves measures have an economic component eg. what could be extracted at current prices? 3.2 Non-renewable esources A. Are socks of non-renewable resources fixed? eserves measures have an economic componen eg. wha could be exraced a curren prices? - Locaion and quaniies of reserves of resources

More information

Product differentiation

Product differentiation differeniaion Horizonal differeniaion Deparmen of Economics, Universiy of Oslo ECON480 Spring 010 Las modified: 010.0.16 The exen of he marke Differen producs or differeniaed varians of he same produc

More information

The Trade-off between Intra- and Intergenerational Equity in Climate Policy

The Trade-off between Intra- and Intergenerational Equity in Climate Policy The Trade-off beween Inra- and Inergeneraional Equiy in Climae Poliy Kverndokk S. E. Nævdal and L. Nøsbakken Posprin version This is a pos-peer-review pre-opyedi version of an arile published in: European

More information

WEEK-3 Recitation PHYS 131. of the projectile s velocity remains constant throughout the motion, since the acceleration a x

WEEK-3 Recitation PHYS 131. of the projectile s velocity remains constant throughout the motion, since the acceleration a x WEEK-3 Reciaion PHYS 131 Ch. 3: FOC 1, 3, 4, 6, 14. Problems 9, 37, 41 & 71 and Ch. 4: FOC 1, 3, 5, 8. Problems 3, 5 & 16. Feb 8, 018 Ch. 3: FOC 1, 3, 4, 6, 14. 1. (a) The horizonal componen of he projecile

More information

3.1.3 INTRODUCTION TO DYNAMIC OPTIMIZATION: DISCRETE TIME PROBLEMS. A. The Hamiltonian and First-Order Conditions in a Finite Time Horizon

3.1.3 INTRODUCTION TO DYNAMIC OPTIMIZATION: DISCRETE TIME PROBLEMS. A. The Hamiltonian and First-Order Conditions in a Finite Time Horizon 3..3 INRODUCION O DYNAMIC OPIMIZAION: DISCREE IME PROBLEMS A. he Hamilonian and Firs-Order Condiions in a Finie ime Horizon Define a new funcion, he Hamilonian funcion, H. H he change in he oal value of

More information

ANSWERS TO EVEN NUMBERED EXERCISES IN CHAPTER 2

ANSWERS TO EVEN NUMBERED EXERCISES IN CHAPTER 2 ANSWERS TO EVEN NUMBERED EXERCISES IN CHAPTER Seion Eerise -: Coninuiy of he uiliy funion Le λ ( ) be he monooni uiliy funion defined in he proof of eisene of uiliy funion If his funion is oninuous y hen

More information

15. Vector Valued Functions

15. Vector Valued Functions 1. Vecor Valued Funcions Up o his poin, we have presened vecors wih consan componens, for example, 1, and,,4. However, we can allow he componens of a vecor o be funcions of a common variable. For example,

More information

1 Review of Zero-Sum Games

1 Review of Zero-Sum Games COS 5: heoreical Machine Learning Lecurer: Rob Schapire Lecure #23 Scribe: Eugene Brevdo April 30, 2008 Review of Zero-Sum Games Las ime we inroduced a mahemaical model for wo player zero-sum games. Any

More information

Lecture 2-1 Kinematics in One Dimension Displacement, Velocity and Acceleration Everything in the world is moving. Nothing stays still.

Lecture 2-1 Kinematics in One Dimension Displacement, Velocity and Acceleration Everything in the world is moving. Nothing stays still. Lecure - Kinemaics in One Dimension Displacemen, Velociy and Acceleraion Everyhing in he world is moving. Nohing says sill. Moion occurs a all scales of he universe, saring from he moion of elecrons in

More information

Appendix 14.1 The optimal control problem and its solution using

Appendix 14.1 The optimal control problem and its solution using 1 Appendix 14.1 he opimal conrol problem and is soluion using he maximum principle NOE: Many occurrences of f, x, u, and in his file (in equaions or as whole words in ex) are purposefully in bold in order

More information

Math Week 14 April 16-20: sections first order systems of linear differential equations; 7.4 mass-spring systems.

Math Week 14 April 16-20: sections first order systems of linear differential equations; 7.4 mass-spring systems. Mah 2250-004 Week 4 April 6-20 secions 7.-7.3 firs order sysems of linear differenial equaions; 7.4 mass-spring sysems. Mon Apr 6 7.-7.2 Sysems of differenial equaions (7.), and he vecor Calculus we need

More information

Neoclassical Growth Model

Neoclassical Growth Model Neolaial Growh Model I. Inroduion As disued in he las haper, here are wo sandard ways o analyze he onsumpion-savings deision. They are. The long bu finie-lived people who leave heir hildren no beque. 2.

More information

EE40 Summer 2005: Lecture 2 Instructor: Octavian Florescu 1. Measuring Voltages and Currents

EE40 Summer 2005: Lecture 2 Instructor: Octavian Florescu 1. Measuring Voltages and Currents Announemens HW # Due oday a 6pm. HW # posed online oday and due nex Tuesday a 6pm. Due o sheduling onflis wih some sudens, lasses will resume normally his week and nex. Miderm enaively 7/. EE4 Summer 5:

More information

KINEMATICS IN ONE DIMENSION

KINEMATICS IN ONE DIMENSION KINEMATICS IN ONE DIMENSION PREVIEW Kinemaics is he sudy of how hings move how far (disance and displacemen), how fas (speed and velociy), and how fas ha how fas changes (acceleraion). We say ha an objec

More information

Course Notes for EE227C (Spring 2018): Convex Optimization and Approximation

Course Notes for EE227C (Spring 2018): Convex Optimization and Approximation Course Noes for EE7C Spring 018: Convex Opimizaion and Approximaion Insrucor: Moriz Hard Email: hard+ee7c@berkeley.edu Graduae Insrucor: Max Simchowiz Email: msimchow+ee7c@berkeley.edu Ocober 15, 018 3

More information

A First Course on Kinetics and Reaction Engineering. Class 19 on Unit 18

A First Course on Kinetics and Reaction Engineering. Class 19 on Unit 18 A Firs ourse on Kineics and Reacion Engineering lass 19 on Uni 18 Par I - hemical Reacions Par II - hemical Reacion Kineics Where We re Going Par III - hemical Reacion Engineering A. Ideal Reacors B. Perfecly

More information

SMT 2014 Calculus Test Solutions February 15, 2014 = 3 5 = 15.

SMT 2014 Calculus Test Solutions February 15, 2014 = 3 5 = 15. SMT Calculus Tes Soluions February 5,. Le f() = and le g() =. Compue f ()g (). Answer: 5 Soluion: We noe ha f () = and g () = 6. Then f ()g () =. Plugging in = we ge f ()g () = 6 = 3 5 = 5.. There is a

More information

A New Formulation of Electrodynamics

A New Formulation of Electrodynamics . Eleromagnei Analysis & Appliaions 1 457-461 doi:1.436/jemaa.1.86 Published Online Augus 1 hp://www.sirp.org/journal/jemaa A New Formulaion of Elerodynamis Arbab I. Arbab 1 Faisal A. Yassein 1 Deparmen

More information

Robust estimation based on the first- and third-moment restrictions of the power transformation model

Robust estimation based on the first- and third-moment restrictions of the power transformation model h Inernaional Congress on Modelling and Simulaion, Adelaide, Ausralia, 6 December 3 www.mssanz.org.au/modsim3 Robus esimaion based on he firs- and hird-momen resricions of he power ransformaion Nawaa,

More information

Some Basic Information about M-S-D Systems

Some Basic Information about M-S-D Systems Some Basic Informaion abou M-S-D Sysems 1 Inroducion We wan o give some summary of he facs concerning unforced (homogeneous) and forced (non-homogeneous) models for linear oscillaors governed by second-order,

More information

t is a basis for the solution space to this system, then the matrix having these solutions as columns, t x 1 t, x 2 t,... x n t x 2 t...

t is a basis for the solution space to this system, then the matrix having these solutions as columns, t x 1 t, x 2 t,... x n t x 2 t... Mah 228- Fri Mar 24 5.6 Marix exponenials and linear sysems: The analogy beween firs order sysems of linear differenial equaions (Chaper 5) and scalar linear differenial equaions (Chaper ) is much sronger

More information

E β t log (C t ) + M t M t 1. = Y t + B t 1 P t. B t 0 (3) v t = P tc t M t Question 1. Find the FOC s for an optimum in the agent s problem.

E β t log (C t ) + M t M t 1. = Y t + B t 1 P t. B t 0 (3) v t = P tc t M t Question 1. Find the FOC s for an optimum in the agent s problem. Noes, M. Krause.. Problem Se 9: Exercise on FTPL Same model as in paper and lecure, only ha one-period govenmen bonds are replaced by consols, which are bonds ha pay one dollar forever. I has curren marke

More information

Explaining Total Factor Productivity. Ulrich Kohli University of Geneva December 2015

Explaining Total Factor Productivity. Ulrich Kohli University of Geneva December 2015 Explaining Toal Facor Produciviy Ulrich Kohli Universiy of Geneva December 2015 Needed: A Theory of Toal Facor Produciviy Edward C. Presco (1998) 2 1. Inroducion Toal Facor Produciviy (TFP) has become

More information

20. Applications of the Genetic-Drift Model

20. Applications of the Genetic-Drift Model 0. Applicaions of he Geneic-Drif Model 1) Deermining he probabiliy of forming any paricular combinaion of genoypes in he nex generaion: Example: If he parenal allele frequencies are p 0 = 0.35 and q 0

More information

Differential Equations

Differential Equations Mah 21 (Fall 29) Differenial Equaions Soluion #3 1. Find he paricular soluion of he following differenial equaion by variaion of parameer (a) y + y = csc (b) 2 y + y y = ln, > Soluion: (a) The corresponding

More information

Chapter Floating Point Representation

Chapter Floating Point Representation Chaper 01.05 Floaing Poin Represenaion Afer reading his chaper, you should be able o: 1. conver a base- number o a binary floaing poin represenaion,. conver a binary floaing poin number o is equivalen

More information

Lecture Notes 5: Investment

Lecture Notes 5: Investment Lecure Noes 5: Invesmen Zhiwei Xu (xuzhiwei@sju.edu.cn) Invesmen decisions made by rms are one of he mos imporan behaviors in he economy. As he invesmen deermines how he capials accumulae along he ime,

More information

Predator - Prey Model Trajectories and the nonlinear conservation law

Predator - Prey Model Trajectories and the nonlinear conservation law Predaor - Prey Model Trajecories and he nonlinear conservaion law James K. Peerson Deparmen of Biological Sciences and Deparmen of Mahemaical Sciences Clemson Universiy Ocober 28, 213 Ouline Drawing Trajecories

More information

COMPETITIVE GROWTH MODEL

COMPETITIVE GROWTH MODEL COMPETITIVE GROWTH MODEL I Assumpions We are going o now solve he compeiive version of he opimal growh moel. Alhough he allocaions are he same as in he social planning problem, i will be useful o compare

More information

The primal versus the dual approach to the optimal Ramsey tax problem

The primal versus the dual approach to the optimal Ramsey tax problem The primal versus he dual approah o he opimal Ramsey ax prolem y George Eonomides a, Aposolis Philippopoulos,, and Vangelis Vassilaos a Deparmen of Inernaional and European Eonomi Sudies, Ahens Universiy

More information

Solutions to Assignment 1

Solutions to Assignment 1 MA 2326 Differenial Equaions Insrucor: Peronela Radu Friday, February 8, 203 Soluions o Assignmen. Find he general soluions of he following ODEs: (a) 2 x = an x Soluion: I is a separable equaion as we

More information

EECE251. Circuit Analysis I. Set 4: Capacitors, Inductors, and First-Order Linear Circuits

EECE251. Circuit Analysis I. Set 4: Capacitors, Inductors, and First-Order Linear Circuits EEE25 ircui Analysis I Se 4: apaciors, Inducors, and Firs-Order inear ircuis Shahriar Mirabbasi Deparmen of Elecrical and ompuer Engineering Universiy of Briish olumbia shahriar@ece.ubc.ca Overview Passive

More information

Vehicle Arrival Models : Headway

Vehicle Arrival Models : Headway Chaper 12 Vehicle Arrival Models : Headway 12.1 Inroducion Modelling arrival of vehicle a secion of road is an imporan sep in raffic flow modelling. I has imporan applicaion in raffic flow simulaion where

More information

Matlab and Python programming: how to get started

Matlab and Python programming: how to get started Malab and Pyhon programming: how o ge sared Equipping readers he skills o wrie programs o explore complex sysems and discover ineresing paerns from big daa is one of he main goals of his book. In his chaper,

More information

Guest Lectures for Dr. MacFarlane s EE3350 Part Deux

Guest Lectures for Dr. MacFarlane s EE3350 Part Deux Gues Lecures for Dr. MacFarlane s EE3350 Par Deux Michael Plane Mon., 08-30-2010 Wrie name in corner. Poin ou his is a review, so I will go faser. Remind hem o go lisen o online lecure abou geing an A

More information

Time to Decide: Information Search and Revelation in Groups

Time to Decide: Information Search and Revelation in Groups Time o Deide: Informaion Searh and Revelaion in Groups Arhur Campbell y Yale Florian Ederer z UCLA Augus 3, 0 Johannes Spinnewijn x LSE and CEPR Absra We analyze osly informaion aquisiion and informaion

More information

CHAPTER 2: Mathematics for Microeconomics

CHAPTER 2: Mathematics for Microeconomics CHAPTER : Mahemaics for Microeconomics The problems in his chaper are primarily mahemaical. They are inended o give sudens some pracice wih he conceps inroduced in Chaper, bu he problems in hemselves offer

More information

Comments on Window-Constrained Scheduling

Comments on Window-Constrained Scheduling Commens on Window-Consrained Scheduling Richard Wes Member, IEEE and Yuing Zhang Absrac This shor repor clarifies he behavior of DWCS wih respec o Theorem 3 in our previously published paper [1], and describes

More information

5.2 Design for Shear (Part I)

5.2 Design for Shear (Part I) 5. Design or Shear (Par I) This seion overs he ollowing opis. General Commens Limi Sae o Collapse or Shear 5..1 General Commens Calulaion o Shear Demand The objeive o design is o provide ulimae resisane

More information

Linear Response Theory: The connection between QFT and experiments

Linear Response Theory: The connection between QFT and experiments Phys540.nb 39 3 Linear Response Theory: The connecion beween QFT and experimens 3.1. Basic conceps and ideas Q: How do we measure he conduciviy of a meal? A: we firs inroduce a weak elecric field E, and

More information

An Introduction to Backward Stochastic Differential Equations (BSDEs) PIMS Summer School 2016 in Mathematical Finance.

An Introduction to Backward Stochastic Differential Equations (BSDEs) PIMS Summer School 2016 in Mathematical Finance. 1 An Inroducion o Backward Sochasic Differenial Equaions (BSDEs) PIMS Summer School 2016 in Mahemaical Finance June 25, 2016 Chrisoph Frei cfrei@ualbera.ca This inroducion is based on Touzi [14], Bouchard

More information