Disjoint Programming in Computational Decision Analysis

Size: px
Start display at page:

Download "Disjoint Programming in Computational Decision Analysis"

Transcription

1 44444 Journal of Unceran Ssems Vol4, No, pp4-3, Onlne a: wwwusorgu Dson Programmng n Compuaonal Decson Analss aosong Dng,, Mas Danelson, Love Eenberg Inernaonal Busness School, Beng Foregn Sudes Unvers 89, Beng, People s Republc of Chna Deparmen of Compuer and Ssems Scences, Socholm Unvers and Roal Insue of Technolog Forum, Ksa, SE-64 4, Sweden Receved 8 December 8; Revsed Ma 9 Absrac Ths paper dscusses a seres of mprecse decson models and her correspondng compuaonal aspecs arsng n compuaonal decson analss The mprecse decson models rela he radonal pon esmaes no nervals and ncorporae varous pes of vague nformaon represened as lnear consrans from a decson-maer When he prncple of mamzng epeced ul s appled as he decson rule, he evaluaons of hese models become nonconve opmzaon problems and requre some global opmzaon sraeges Ths paper presens a class of global opmzaon algorhms for solvng such non-conve programs We ae advanage of polar cus and he dson srucural proper of he mprecse decson models o develop generalzed cung plane mehods ha are dfferen from he radonal class of branch and bound approaches World Academc Press, UK All rghs reserved Kewords: global opmzaon, dson programmng, cung plane, mprecse decson analss Inroducon Mos classcal decson analss approaches conss of a se of sraghforward decson rules appled o precse esmaes of weghs, probables, and/or ules no maer how unsure a decson-maer s of hs esmaes The requremen for numercall precse daa has been consdered unrealsc b an ncreasng number of researchers and decson-maers Technques allowng mprecson have been suggesed, among whch he nerval mehods have been wdel ulzed o srenghen oher decson models ha are bul upon he prncple of mamzng he epeced ul (PMEU), eg, [7, 8, 9, 3, 6, 7] When PMEU s appled, dson blnear programmng (DBLP) programs such as α or where pu αwu α or α are coeffcens, are obaned whle searchng for he bes alernaves f boh nerval probables or weghs and nerval ules are represened b varables When modelng mulcrera problems ha also have unceran evens, combnng hese wo nds of decson models gves rse o mulcrera probablsc models wh rlnear epressons such as α w p u, and elds compuaonall hard ( ) dson rlnear programmng (DTLP) programs ogeher wh consrans on weghs, probables, and ul values Moreover, for decson suaons ha possess dson probabl chan proper adopng PMEU ma even lead o dson mul-lnear programmng (DMLP) programs Alhough opmzaon echnques have been developed rapdl n recen ears, much of he area of opmzaon has been devoed o solvng large problems Neverheless, he parcular programs arsng n mprecse compuaonal decson analss requre approaches for solvng sequences of relaed smaller global opmzaon problems n neracve me because n mahemacal programmng, DBLP, DTLP and DMLP all ehb non-conve Therefore, some global opmzaon sraeges ha erae beween a global phase and a local phase n search of he global opmum are necessar Two man branches of deermnsc approaches for handlng such non-conve programs are cung plane mehods [,, 9,, ], and branch and bound mehods [,, 4, 4, 5, 8] Combned approaches have also been suggesed [3, ] Ths paper nends o dscuss a seres of mprecse decson models and her correspondng compuaonal aspecs A class of generalzed cung plane mehods s developed b ang advanage of polar cus and he dson srucural proper of he mprecse decson models The followng secon presens an mprecse decson framewor Correspondng auhor Emal: aosongdng@gmalcom ( Dng)

2 Journal of Unceran Ssems, Vol4, No, pp4-3, 5 ha can be readl ransformed no DBLP and he fundamenal echnque for solvng DBLP Thereafer, we eend he decson framewor no an mprecse probablsc mul-crera decson model, ransform no DTLP, and dscuss he correspondng global opmzaon algorhm n grea deal The fourh secon presens an mprecse mullevel probabl chan decson model ha can be ransformed no DMLP, and dscusses he global opmzaon algorhm eended from he hrd secon The fnal secon concludes hs paper and ndcaes furher research opcs Dson Blnear Programmng A Blnear Imprecse Decson Model In nerval decson analss, a decson-maer s ofen encouraged o be delberael mprecse n hs subecve udgemens Inpu senences n decson problems conssng of dscree alernaves wh evens and consequences, eg, nerval esmaes and qualave nformaon, are ranslaed no lnear consrans For nsance, he probabl (or ul of a consequence c beng beween he numbers a and b s epressed as p a, b ] (or u a, b ]) [ [ Relaons can be handled smlarl: a measure of c s greaer han a measure of c s epressed as p p (or analogousl u ) In hs wa each saemen s represened b one or more consrans u l Defnon : An nformaon frame represens a decson suaon wh m alernaves, each alernave havng m, =,, m consequences, as a srucure C, P, U = {{ c } }, P, U,,, m,,, m, where each denoes a m m consequence, P s a fne ls of lnear consrans over p -varables denong probables over he consequences c, and U s a fne ls of lnear consrans over u -varables denong ules over he consequences c l c l Fgure : A blnear mprecse decson model Suppose we have a decson suaon as shown n Fgure where D s a decson node, E and E are probabl nodes, represenng ndeermnsm, wh assocaed probabl dsrbuons, and he leaves are consequence nodes wh conve ses of assocaed value or ul funcons Ths decson suaon can be mapped no an nformaon frame C, P, U = {{ c } }, P U wh m =, m = and m 4 m m, Ths srucure s hen populaed wh user saemens represened as lnear nequales Snce a vecor n he polope can be consdered o represen a dsrbuon, a probabl base P can be nerpreed as consrans defnng he se of all possble probabl measures over he consequences Analogousl a ul base U consss of consrans defnng he se of all possble ul funcons over he consequences These wo bases ogeher wh a ree srucure consue he nformaon frame Defnon : Gven an nformaon frame {{ c } }, P U, he epeced ul of an alernave s m m, E ( A, where and are varables n P and U, respecvel; denoes he probabl of ) = p m u p u p occurrng gven ha alernave A s chosen; and u denoes he correspondng ul Usng precse numbers, evaluang he epeced ul of an alernave s raher sraghforward Neverheless, when numercall mprecse nformaon s nvolved, he epeced ul has o be evaluaed wh respec o P and U = A c

3 6 Dng, M Danelson and L Eenberg:Dson Programmng n Compuaonal Decson Analss B consderng all possble soluon vecors, a range of epeced ules for each alernave can be receved Once we have a well-defned nformaon frame and appl PMEU as he decson rule, we need o calculae he mnmal and mamal epeced ules for each alernave Frsl when mprecse nformaon represened as lnear consrans ess whn eher P or U, o evaluae he decson model s reduced o lnear programmng (LP) problems ha can be readl solved wh modern opmzaon echnques [8] Secondl as mprecse nformaon ess n boh P and U concernng ceran creron, he evaluaon of a decson model under hs creron based on PMEU resuls n a smplfed DBLP program n he sense ha onl conans cross produc erms such as pv bu no lnear erms For eample, f we need o evaluae he second 4 alernave n Fgure, we have o calculae ma/ mn = p u subec o dson lnear consrans concernng P and U, respecvel I should be noed ha for hs blnear mprecse decson model, we onl allow a decsonmaer o ssue hs mprecse saemens wh respec o each alernave such as p + p 3, bu eclude he cross lnear consrans among dfferen alernaves such as p + p 3 Then, hs decson model can be ransformed no DBLP ha s mahemacall saed as mn f (, = ( c + d )( c + d ) mn f (, = c + d + C () n s = { R n : A b, }, Y = { R : A b, } In (), A and represen he lnear consrans populaed n P and U, respecvel b A b Evaluaon of DBLP Local Opmzaon The mos mporan proper of DBLP s ha, even hough f (, ma no be quas-concave, here ess an ereme pon and an ereme pon Y such ha (, s an opmal soluon of DBLP [] Ths soluon proper and he srucure of DBLP self sugges an LP based vere followng algorhm ha converges o a Karush-Kuhn- Tucer pon [] Defnon 3: Consder P : mn f ( ) subec o S, where S s a compac polhedral se and f s non-conve A local sar mnmzer (LSM) of P s defned as a pon such ha f ( ) f ( ) for each N S (), where N S () denoes he se of ereme pons n S ha are adacen o For DBLP, an ereme pon s adacen o (, f and onl f s of he form eher (, or (, ), where N ( ) and N ( ) Y Defnon 4: An ereme pon (, s called a pseudo-global mnmzer (PGM) f f (, f (, for each B δ ( ) and for each, where Y B δ () s a δ neghborhood around Wh he above proper and defnon, we can have an LP based mounan clmbng procedure o oban a PGM of DBLP Algorhm : (a) Fnd a feasble ereme pon ~, where represens he reduced feasble regon n h eraon afer all cus have been added (b) () Solve: ( ~ mn{ f, Y} o eld an opmal ~ ; () Solve: mn{ (, ~ ) f } o eld an opmal ~ ; Se ~ ~ and repea sep (b) unl converges o an LSM (, (c) Suppose s non-degenerae and le ˆ N( ) be such ha f ˆ, ˆ) mn f ( ˆ, < mn f (, f (, ) ( = Y Y = Go o sep (b()) wh ~ ˆ (d) Termnae wh (, as a PGM Algorhm s relavel eas o mplemen and here are oher approaches concenrang on he locaon of dfferen pes of ereme pon [, 9] However, n hs paper, we wll no sep no he deals of oher local opmzaon mehods snce we focus on he ulzaon and generalzaon of such local algorhms n search of he global opmum The ncorporaon of oher approaches s supposed o be smlar

4 Journal of Unceran Ssems, Vol4, No, pp4-3, 7 Generaon of a Cung Plane Gven a PGM (, locaed b Algorhm, we emplo polar cus o cu off local vere soluons Assume s a non-degenerae ereme pon of ; le, N, be he n nonbasc varables a, where N s he nde se for he nonbasc varables Then has precsel n dsnc edges ncden o Each half lne ξ = { : = a λ, λ }, N conans eacl one such edge [5] Defnon 5: The generalzed reverse polar of Y for a gven scalar α s gven b Y ( α) = { : f (, α} for all Y Le (, be a PGM, le he ras ξ be defned as above, le α be he curren bes obecve value (CBOV) of f (,, and le λ be defned b λ = ma{ λ : f ( a λ, α for all Y } f ξ Y ( α) () As for he case when ξ Y ( α), we smpl se λ o raher han emplong he negave eenson of polar cus [9] because n ha case, he followng approach o compue λ canno be used an longer The nequal N / λ deermnes a vald cung plane Each λ can be compued b an approach on he bass of LP dual heor ha coss onl one LP eraon Consder () and (), n whch we need o oban ma{ λ : f ( a λ, α A b, } = ma{ λ : mn [ c ( a λ ) + d + ( a λ ) C] α A b, } Usng LP dual heor he foregong can be rewren as ma{ λ : ma [ c ( a λ ) + b u] α A u d + C ( a λ ), u } = ma{ λ : mn [ c a λ b u] c α C a λ + A u C + d, u } In mar form, hs epresson ass for he mamal λ n ma λ Hence, we can oban Global Opmzaon u u λ, u c a b λ c α s, u C a A u C + d λ when b solvng us one LP program ξ Y ( α ) A hs sage, we are able o presen a global opmzaon algorhm for solvng DBLP b usng Algorhm and polar cus Algorhm : (a) Le CBOV, ob = + ; le he nal bes feasble soluon {( ˆ, ˆ )} = φ Se = and = (b) If = φ, ermnae wh ob as he global mnmum and ( ˆ, ˆ ) as s correspondng global mnmzer (c) Fnd a PGM (, ) b usng Algorhm, and se he correspondng ob mn{, (, )}, ( ˆ, ˆ = ob f ) = arg mn{ ob, f (, )} (d) Compue λ s b solvng LP programs f, and se ξ Y ( α ) λ s o f Generae a polar cu ξ Y ( α) + + and defne = H ( ) (e) If here ess no λ such ha ξ Y ( α ), ermnae wh ob as he global mnmum and as s ( ˆ, ˆ ) correspondng global mnmzer (f) Se +, and reurn o (b) Accordng o he soppng rules, Algorhm elds an eac global mnmum for DBLP (3)

5 8 Dng, M Danelson and L Eenberg:Dson Programmng n Compuaonal Decson Analss Convergence Proof: Frs, noe ha Algorhm s fne so sep (c) n Algorhm elds eac soluons Consder he sequence of PGMs {(, )} generaed and le H ( ) be he cung plane ha elmnaes h In sep (d) of eraon, he algorhm s ermnaed as he consequence of nroducng polar cus, and he algorhm s ermnaed f + H ( ) = φ Oherwse, he cu ( + H ) s appled and a new PGM ( +, ) s found where + + H ( ) and + H ( ) I s possble for he process no o ermnae b an of he soppng rules n Algorhm An nfne sequence would hen be generaed, and we need o show ha he sequence { } has a lm pon such ha + lm H ( ) = φ Snce s a compac se, here ess a lm pon such ha for a gven ε and a posve neger ν, ε for nfnel man ν If + l l H ( ) φ for all ν, hen all subsequen PGMs (, ) generaed wll sasf he condon l H + ( ν ) for all l > ν From he defnon of a PGM, B ( ) and l B ( ) for some δ > Hence, l δ for all l > ν Ths conradcs he saemen ha s a lm pon δ Therefore, + lm H ( ) = φ and he cung plane algorhm s ermnaed 3 Dson Trlnear Programmng 3 A Probablsc Mul-Crera Decson Model The mprecse blnear decson model handles a decson suaon wh respec o onl one creron However, n man decson suaons s necessar o sud a decson problem from more han one perspecve usng mulple crera There are several mul-crera models, rangng from value funcon based mehods over goal and reference pon mehods o ouranng mehods For hs PMEU-based model, an addve mul-arbue ul heor (MAUT) approach s emploed, n whch he epeced ules of each alernave are weghed ogeher accordng o mporance weghs Consder n nformaon frames under n crera and suppose ha he mprecse saemens concernng mporance weghs are colleced n a se W wh w -varables Defnon 6: An eended nformaon frame represens a decson suaon wh m alernaves Each alernave s assessed under n crera,, n, and each creron has an nformaon frame I = {{ c} m } m, P U, as a δ, srucure W,{ I} where W s a fne ls of lnear consrans over w -varables denong weghs over he crera Snce he addve model s PMEU-based, evaluaon rules are based on an addve eenson of he epeced ul Defnon 7: Gven an eended nformaon frame W, { I },,, n, he weghed epeced ul of an n m alernave s G A ) = w ( p u ), where p and u are varables n P and U, respecvel; A ( = = c denoes he probabl of occurrng gven ha A s aen; and u denoes he correspondng ul valued under creron When mprecse nformaon ess n weghs, probables, and ules, he evaluaon of weghed epeced ul n an eended nformaon frame resuls n DTLP ha s mahemacall saed as: mn f (, = ( c + d )( c + d )( c z + d ) s = { R n Y = { R n n3 Z = { z R : A3 z b3, z } In (4), A and represen he lnear consrans populaed n P, U and W, respecvel To b, A b A3 z b3 receve a range of he weghed epeced ules for each alernave under n crera, we need o calculae boh mnmal and mamal values When he receved ranges for wo alernaves overlap, a decson-maer s suggesed o provde more nformaon for he ranges o become separaed, and hus ndcae a preference order : A b, } : A b, } z z p (4)

6 Journal of Unceran Ssems, Vol4, No, pp4-3, 9 3 Evaluaon of DTLP 3 Local Opmzaon For DTLP, because of he dson consran ses concernng P, U and W, respecvel he soluon proper of DBLP can be eended Theorem : If,Y and Z are nonemp and bounded, hen DTLP has an opmal soluon n whch (,, z,), and z are basc feasble soluons of,y and Z, respecvel Proof: Le, z be an opmal soluon Frs, consder he LP problem concernng, mn{ f (, }, and le be s opmal basc soluon Then we have f (, f (, Ne, consder he LP problem concernng mn{ f (, Y}, and le be s opmal basc soluon B he same argumen, we have f (,, f (, Fnall consder he LP problem concernng z, mn{ f (,, z Z }, and le z be s opmal basc soluon Then we have f (,, z ) f (,, Therefore,, and z are basc feasble soluons of,y and Z, respecvel Based on hs soluon proper and he nowledge n DBLP, we can have he followng local opmzaon algorhm o locae a PGM of DTLP Algorhm 3: (a) Fnd feasble ereme pons ~ and ~ Y (b) ()Solve: ( ~ ~ mn{ f,, z Z}, o eld an opmal ~ z ; () Solve: ~ ~ mn{ f (,, z ) }, o eld an opmal ~ ; (3)Solve: ( ~,, ~ mn{ f z ) Y}, o eld an opmal ~ ; Se ~ ~, ~ ~ and repea (b) unl converges o an LSM (, (c) Suppose s non-degenerae and le ˆ N( ) be such ha f ( ˆ, ˆ, = mn Y f ( ˆ, < f (, Go o (b()) wh ~ ˆ, ~ z z (d) Suppose s non-degenerae and le ˆ N( ) be such ha f ( ˆ, zˆ) = mn z Z f ( ˆ, < f (, Go o (b()) wh ~ ~ z zˆ (e) Termnae wh (, as a PGM 3 Global Opmzaon Gven a PGM (, of DTLP, we need o develop a generalzed cung plane mehod o oban he global opmum As a resul of he ndependenc beween,y and Z, we can develop a generalzed cung plane mehod solvng DTLP Algorhm 4: (a) Le CBOV, ob = +, le he nal bes feasble soluon {( ˆ, ˆ, zˆ )} = φ Se = and = (b) If = φ, ermnae wh ob as he global mnmum and he soluon ( ˆ, ˆ, zˆ ) as s correspondng global mnmzer (c) Fnd a PGM n b usng Algorhm 3, and se he correspondng ob mn{, (,, )},( ˆ, ˆ = ob f z, zˆ ) = arg mn{ ob, f (,, z )} (d) Compue λ concernng Y, and compue λ concernng Z for all N Y Z (e) If eher here ess no λ such ha ξ Y ( α ) or λ such ha ξ Z ( α), ermnae wh ob as he global mnmum and ( ˆ, ˆ, zˆ ) as s correspondng global mnmzer Y + + (f) Le λ = mn{ λ, λ } for all N, generae a polar cu, and le = H ( ) Y Z Z

7 Dng, M Danelson and L Eenberg:Dson Programmng n Compuaonal Decson Analss and (g) Se = +, and reurn o (b) The fundamenal dea n Algorhm 4 s ha a a local mnmzer n Z, respecvel and choose he smaller one as, we compue λ Y and λ concernng Z λ for each edge Ths wll guaranee he global opmal as wha we have done n DBLP Convergence of Algorhm 4 can be proved analogousl o he prevous wor Proposon: Gven an eended nformaon frame, ma G ( A and can be compued b DTLP ) mn G( A ) opmzaon Ths wa weghed epeced ul ranges can be obaned for each alernave Ths follows from he observaon ha he problem srucure of evaluang PMEU wh respec o an eended nformaon frame concdes wh he formulaon of (4) 4 Dson Mul-Lnear Programmng 4 A Mullevel Decson Model Wh approprae modfcaons o an nformaon frame, a well defned decson srucure as shown n Fgure can be obaned, n whch we have a sequence of chance nodes represenng unceran Neverheless, P s now broen no several dson pars, eg,, P and, wh respec o each level P P 3 Y Fgure : A mul-lnear decson model To calculae he epeced ul we have E( A ) = p( p( p3u + p3u) + p( p33u3 + p34u4)) + p( p3u5 + p4u6)) (5) = p p p3u + p p p3u + p p p33u3 + p p p34u4 + p p3u5 + p p4u6 As mprecse nformaon prevals, we ranslae no lnear consrans whn each level I should be noed ha currenl we onl allow consrans from he same level raher han dfferen levels For eample, he nerval saemen p 3 s consdered proper, whle he nerval saemen + p3 p + p3 3 s consdered mproper snce p and are from level and level, respecvel and hs wll desro he dson srucural p3 proper In oher words, mprecse saemens are supposed o be confned whn each level To evaluae he alernave b usng PMEU wh mprecse nformaon, we have o compue wo ereme values, e, mamum and mnmum, of (5) subec o dson lnear consrans n order o receve a range of epeced ul for he alernave Ths wll resul n a specal case of DMLP ha s nrnscall hard o solve If we ae, and, hen he mprecse mullevel decson = p = ( p,, pn ), =,, n n+ = ( u,, un ) u model (5) can be ransformed no he followng DMLP model as T = mn f (,, n+ ) = J n s = { R : A b, }, =,, n + (6) In (6), J denoes he nde se, and we can have a mos one decson varable from for each Tang (5) J as an eample, here ess a mos one decson varable from each level whn each produc erm Ths proper of he obecve funcon and he dson lnear consran ses, s, demonsrae he dson proper of DMLP, and herefore, s soluon proper s smlar o ha of DTLP

8 Journal of Unceran Ssems, Vol4, No, pp4-3, 4 Evaluaon of DMLP 4 Local Opmzaon Based on he soluon proper of DMLP and he nowledge n DBLP and DTLP, we can have he followng local h opmzaon algorhm o locae a PGM of DMLP, n whch denoes he feasble regon of n eraon; and N ( n+ ) denoes he se of ereme pons n ha are adacen o n Algorhm 5: ~ (a) Fnd feasble ereme pons, =,, n (b) () Solve: ( ~ ~ mn{ f,, n, n+ ) n+ n+ } o eld ~ n+ () Solve: ~ ~ mn{ f (,,, n+,) } o eld ~ (3) Solve: ( ~ ~,,,, ~ mn{ f 3 n+,) } o eld ~ n + n+ + n+ M (n+) Solve: ( ~ ~,,,, ~ mn{ f n n n+,) n n} o eld ~ n Se ~ ~, =,, n, and repea (b) unl he soluon converges o an LSM (,, n + ) (c) Suppose s non-degenerae, and for each n+, =,, n, le ˆn+ N( n+ ) be such ha f,,, ˆ,,,, ˆ ) mn f (,,,,,,, ˆ ) f (,, Go o (b()) wh ( + n n+ = + n n+ < n+ ~, =,,, +,, n and ~ ˆ (,, n as a PGM + (d) Termnae wh ) 4 Global Opmzaon Now we can develop he generalzed cung plane mehod for solvng DMLP b ang advanage of s dson proper As before, for he generalzed reverse polar, f ξ (α ), =,, n, N, we smpl se λ = oher han emplong he negave eenson of polar cus Algorhm 6: (a) Le CBOV, ob = +, le he nal bes feasble soluon {( ˆ,, ˆ )} = φ, and se n+ = (b) If = φ, ermnae wh as he global mnmum and as s correspondng global n+ ob ( ˆ,, ˆ n+ ) soluon (c) Fnd a PGM n b usng Algorhm 5, and se he correspondng n + ob mn{, (,, )},( ˆ,, ˆ = ob f n+ n+ ) = arg mn{ ob, f (,, n+ )} (d) Compue λ wh respec o for all N, =,, n (e) If here ess no λ such ha ξ ( ob ), =,, n, ermnae wh ob as he global mnmum and ˆ,, ˆ ( n+ ) as s correspondng global mnmzer + + (f) Le λ = mn{ } for all N, =,, n, generae a polar cu, and le = H ) λ ) n+ n+ ( n+ (g) Se = +, and reurn o (b) The basc dea n Algorhm 6 s smlar o hose eplaned n Algorhm 4 and s convergence can be proved analogousl o he prevous wor 5 Conclusons and Furher Research In hs paper, we have dscussed hree pes of mprecse decson models arsng n compuaonal decson analss, he correspondng soluon proper of DBLP, DTLP and DMLP, and her global opmzaon algorhms Beng based on onl LP operaons, he do no conan radonal nonlnear compuaonal elemens Even hough we have ganed some compuaonal eperence n Algorhm ; see [3, 4,, ], Algorhm 4 and Algorhm 6 sll need furher nvesgaons and modfcaons wh respec o her compuaonal performance benchmar agans he general-purpose desgned branch and bound mehods [4, 5]

9 Dng, M Danelson and L Eenberg:Dson Programmng n Compuaonal Decson Analss As for he opmzaon algorhms hemselves, here are several aspecs for furher developmen and mprovemen Frsl s possble o ncorporae ceran lower boundng echnques n branch and bound procedures o fahom he res of he feasble regon and acheve faser convergence [] For eample, we are able o oban an nerval such L U as for each varable Then usng he arhmec nervals [,, 4], he conve envelope of a rlnear erm z subec o bo consrans can be calculaed as Each blnear erm can be furher lower bounded b he mamum of wo lnear consrans, and we need o compue he mamum of 3 lnear consrans o oban an underesmae of z In ha case, boh algorhms ma be furher mproved o acheve fas convergence Secondl he generaon of a cung plane a a degenerae ereme pon as n Algorhms, 4 and 6 can be furher nvesgaed snce s a common suaon n compuaonal decson analss [5, 6] Fnall we ma ncorporae he negave eenson of polar cus when ξ Y ( α) n order o generae more effcen cung planes, and focus on he ereme pons belongng o he orgnal feasble se ha are canddaes for he global soluon raher han hose nduced b he generaed cung planes [9] 6 Acnowledgemens Ths wor s parall suppored b Naonal Naural Scence Foundaon of Chna under Gran #7773 and Proec References [] Al-Khaal, FA, Jonl consraned blnear programs and relaed problems: An overvew, Compuers & Mahemacs wh Applcaon, vol9, no, pp53 6, 99 [] Al-Khaal, FA, and JE Fal, Jonl consraned bconve programmng, Mahemacs of Operaons Research, vol8, pp73 86, 983 [3] Alare, S, C Aude, B Jaumard, and G Savard, Concav cus for dson blnear programmng, Mahemacal Programmng, vol9, no, pp , [4] Aude, C, P Hansen, B Jaumard, and G Savard, A smmercal lnear mamn approach o dson blnear programmng, Mahemacal Programmng, vol85, no3, pp573 59, 999 [5] Balas, E, Inersecon cus A new pe of cung planes for neger programmng, Operaons Research, vol9, pp9 39, 97 [6] Carvaal-Moreno, R, Mnmzaon of concave funcons subec o lnear consrans, Operaons Research Cener Repor, Unvers of Berele Calforna, 97 [7] Danelson, M, Generalzed evaluaon n decson analss, European Journal of Operaonal Research, vol6, no, pp44 449, 5 [8] Danelson, M, and L Eenberg, A framewor for analsng decsons under rs, European Journal of Operaonal Research, vol4, no3, pp , 998 [9] Danelson, M, and L Eenberg, Compung upper and lower bounds n nerval decson rees, European Journal of Operaonal Research, vol8, no, pp88 86, 7

10 Journal of Unceran Ssems, Vol4, No, pp4-3, 3 [] Dng,, and FA Al-Khaal, Accelerang convergence of cung plane algorhms for dson blnear programmng, Journal of Global Opmzaon, vol38, no3, pp4 436, 7 [] Gallo, G, and A ülücü, Blnear programmng: An eac algorhm, Mahemacal Programmng, vol, pp73 94, 977 [] Konno, H, A cung plane algorhm for solvng blnear programs, Mahemacal Programmng, vol, pp4 7, 976 [3] Ros Insua, D, and S French, A framewor for sensv analss n dscree mul-obecve decson mang, European Journal of Operaonal Research, vol54, pp76 9, 99 [4] Roo, HS, and NV Sahnds, Analss of bounds for mullnear funcons, Journal of Global Opmzaon, vol9 4, pp43 44, [5] Roo, HS, and NV Sahnds, Global opmzaon of mulplcave programs, Journal of Global Opmzaon, vol6 4, pp387 48, 3 [6] Sage, AP, and CC Whe, ARIADNE: A nowledge-based neracve ssem for plannng and decson suppor, IEEE Transacons on Ssems, Man and Cbernecs, vol4, pp35 47, 984 [7] Salo, A, and RP Hamalanen, Preference programmng hrough appromae rao comparsons, European Journal of Operaonal Research, vol8, no3, pp , 995 [8] Sheral HD, and A Alameddne, A new reformulaon lnearzaon algorhm for blnear programmng problems, Journal of Global Opmzaon, vol, pp379 4, 99 [9] Sheral, HD, and CM She A fnel convergen algorhm for blnear programmng problems usng polar cus and dsuncve face cus, Mahemacal Programmng, vol9, pp4 3, 98 [] Tu H, Concave programmng under lnear consrans, Sove Mah Dol, vol5, pp437 44, 964 [] Vash, H, and CM She A cung plane algorhm for he blnear programmng problem, Naval Research Logscs Quarerl vol4, pp83 94, 977

11 8 S Dora: Dfferen Tpes of Convergences for Random Varables Theorem 8 Le (Ω, d) be a merc space and le B be a paron of Ω For ever B B wh posve and fne Hausdorff ouer measure n s dmenson denoe b µ = P (A B) he resrcon o he Borel σ-feld of he upper condonal probabl defned as n Theorem Le L (B) be he class of all Borel measurable random varables on B Then he convergence n µ-dsrbuon of a sequence of random varables of L (B) o a random varable s equvalen o he ponwse convergence of epecaon funconals on all bounded and connuous funcon f ha s lm n fdµn = fdµ Proof: If n and are Borel-measurable random varables and H s a Borelan se hen he ses n (H) and (H) are also Borelan ses; moreover snce ever Hausdorff s-dmensonal ouer measure s counabl addve on he Borel σ-feld hen he (upper) condonal probables µ n and µ nduced respecvel b n and on (R,F ) are probabl measures Then convergence n µ-dsrbuon s equvalen o he ponwse convergence of epecaon funconals on all bounded and connuous funcon f 6 Conclusons Ths paper nvesgaes he relaons among dfferen pes of convergence for random varables when he are based on an upper probabl approach where condonal upper epecaons wh respec o Hausdorff ouer measures are used whenever we have o condon on a se wh probabl zero Upper (lower) condonal prevsons defned wh respec o Hausdorff ouer measures are proven o be he upper (lower) envelopes of all lnear eensons o he class of all random varables of he resrcon o he Borel-measurable random varables of he gven upper condonal prevsons I s proven ha he relaons among dfferen pes of convergences of random varables defned wh respec o upper condonal probabl defned b Hausdorff ouer measures are he same ha hold f convergences are defned wh respec o a probabl measure When he condonng even has fne Hausdorff ouer measure n s dmenson hese resuls are obaned because Hausdorff ouer measures are Borel regular ouer measures and so connuous from below and connuous from above on he Borel σ-feld In general f upper condonal probabl s defned as naural eenson of a coheren merel fnel addve probabl defned on a σ-feld we have ha µ-sochascall convergence does no mpl convergence n µ-dsrbuon snce n hs case he upper condonal probabl s no connuous from above References [] Bllngsle P, Probabl and Measure, Wle 986 [] Couso, I, S Mones, and P Gl, Sochasc convergence, unform negrabl and convergence n mean on fuzz measure spases, Fuzz Ses and Ssems, vol9, pp95 4, [3] de Cooman, G, MCM Troffaes, and E Mranda, n-monoone lower prevsons and lower negrals, Proceedngs of he Fourh Inernaonal Smposum on Imprecse Probabl: Theores and Applcaons, pp45 54, 5 [4] Dora, S, Probablsc ndependence wh respec o upper and lower condonal probables assgned b Hausdorff ouer and nner measures, Inernaonal Journal of Appromae Reasonng, vol46, pp67 635, 7 [5] Denneberg, D, Non-addve Measure and Inegral, Kluwer Academc Publshers, 994 [6] Falconer, KJ, The Geomer of Fracals Ses, Cambrdge Unvers Press, 986 [7] Rogers, CA, Hausdorff Measures, Cambrdge Unvers Press, 998 [8] Sedenfeld, T, M Schervsh, and JB Kadane, Improper regular condonal dsrbuons, The Annals of Probabl vol9, no4, pp6 64, [9] Walle P, Sascal Reasonng wh Imprecse Probables, Chapman and Hall, London, 99

Solution in semi infinite diffusion couples (error function analysis)

Solution in semi infinite diffusion couples (error function analysis) Soluon n sem nfne dffuson couples (error funcon analyss) Le us consder now he sem nfne dffuson couple of wo blocks wh concenraon of and I means ha, n a A- bnary sysem, s bondng beween wo blocks made of

More information

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS R&RATA # Vol.) 8, March FURTHER AALYSIS OF COFIDECE ITERVALS FOR LARGE CLIET/SERVER COMPUTER ETWORKS Vyacheslav Abramov School of Mahemacal Scences, Monash Unversy, Buldng 8, Level 4, Clayon Campus, Wellngon

More information

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany Herarchcal Markov Normal Mxure models wh Applcaons o Fnancal Asse Reurns Appendx: Proofs of Theorems and Condonal Poseror Dsrbuons John Geweke a and Gann Amsano b a Deparmens of Economcs and Sascs, Unversy

More information

Cubic Bezier Homotopy Function for Solving Exponential Equations

Cubic Bezier Homotopy Function for Solving Exponential Equations Penerb Journal of Advanced Research n Compung and Applcaons ISSN (onlne: 46-97 Vol. 4, No.. Pages -8, 6 omoopy Funcon for Solvng Eponenal Equaons S. S. Raml *,,. Mohamad Nor,a, N. S. Saharzan,b and M.

More information

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim Korean J. Mah. 19 (2011), No. 3, pp. 263 272 GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS Youngwoo Ahn and Kae Km Absrac. In he paper [1], an explc correspondence beween ceran

More information

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005 Dynamc Team Decson Theory EECS 558 Proec Shruvandana Sharma and Davd Shuman December 0, 005 Oulne Inroducon o Team Decson Theory Decomposon of he Dynamc Team Decson Problem Equvalence of Sac and Dynamc

More information

Comparison of Differences between Power Means 1

Comparison of Differences between Power Means 1 In. Journal of Mah. Analyss, Vol. 7, 203, no., 5-55 Comparson of Dfferences beween Power Means Chang-An Tan, Guanghua Sh and Fe Zuo College of Mahemacs and Informaon Scence Henan Normal Unversy, 453007,

More information

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS THE PREICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS INTROUCTION The wo dmensonal paral dfferenal equaons of second order can be used for he smulaon of compeve envronmen n busness The arcle presens he

More information

CHAPTER 10: LINEAR DISCRIMINATION

CHAPTER 10: LINEAR DISCRIMINATION CHAPER : LINEAR DISCRIMINAION Dscrmnan-based Classfcaon 3 In classfcaon h K classes (C,C,, C k ) We defned dscrmnan funcon g j (), j=,,,k hen gven an es eample, e chose (predced) s class label as C f g

More information

Approximate Analytic Solution of (2+1) - Dimensional Zakharov-Kuznetsov(Zk) Equations Using Homotopy

Approximate Analytic Solution of (2+1) - Dimensional Zakharov-Kuznetsov(Zk) Equations Using Homotopy Arcle Inernaonal Journal of Modern Mahemacal Scences, 4, (): - Inernaonal Journal of Modern Mahemacal Scences Journal homepage: www.modernscenfcpress.com/journals/jmms.aspx ISSN: 66-86X Florda, USA Approxmae

More information

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Journal of Appled Mahemacs and Compuaonal Mechancs 3, (), 45-5 HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Sansław Kukla, Urszula Sedlecka Insue of Mahemacs,

More information

Notes on the stability of dynamic systems and the use of Eigen Values.

Notes on the stability of dynamic systems and the use of Eigen Values. Noes on he sabl of dnamc ssems and he use of Egen Values. Source: Macro II course noes, Dr. Davd Bessler s Tme Seres course noes, zarads (999) Ineremporal Macroeconomcs chaper 4 & Techncal ppend, and Hamlon

More information

GORDON AND NEWELL QUEUEING NETWORKS AND COPULAS

GORDON AND NEWELL QUEUEING NETWORKS AND COPULAS Yugoslav Journal of Operaons Research Vol 9 (009) Number 0- DOI:0.98/YUJOR0900C GORDON AND NEWELL QUEUEING NETWORKS AND COPULAS Danel CIUIU Facul of Cvl Indusral and Agrculural Buldngs Techncal Unvers

More information

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC CH.3. COMPATIBILITY EQUATIONS Connuum Mechancs Course (MMC) - ETSECCPB - UPC Overvew Compably Condons Compably Equaons of a Poenal Vecor Feld Compably Condons for Infnesmal Srans Inegraon of he Infnesmal

More information

Variants of Pegasos. December 11, 2009

Variants of Pegasos. December 11, 2009 Inroducon Varans of Pegasos SooWoong Ryu bshboy@sanford.edu December, 009 Youngsoo Cho yc344@sanford.edu Developng a new SVM algorhm s ongong research opc. Among many exng SVM algorhms, we wll focus on

More information

Testing a new idea to solve the P = NP problem with mathematical induction

Testing a new idea to solve the P = NP problem with mathematical induction Tesng a new dea o solve he P = NP problem wh mahemacal nducon Bacground P and NP are wo classes (ses) of languages n Compuer Scence An open problem s wheher P = NP Ths paper ess a new dea o compare he

More information

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study)

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study) Inernaonal Mahemacal Forum, Vol. 8, 3, no., 7 - HIKARI Ld, www.m-hkar.com hp://dx.do.org/.988/mf.3.3488 New M-Esmaor Objecve Funcon n Smulaneous Equaons Model (A Comparave Sudy) Ahmed H. Youssef Professor

More information

Outline. Probabilistic Model Learning. Probabilistic Model Learning. Probabilistic Model for Time-series Data: Hidden Markov Model

Outline. Probabilistic Model Learning. Probabilistic Model Learning. Probabilistic Model for Time-series Data: Hidden Markov Model Probablsc Model for Tme-seres Daa: Hdden Markov Model Hrosh Mamsuka Bonformacs Cener Kyoo Unversy Oulne Three Problems for probablsc models n machne learnng. Compung lkelhood 2. Learnng 3. Parsng (predcon

More information

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes.

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes. umercal negraon of he dffuson equaon (I) Fne dfference mehod. Spaal screaon. Inernal nodes. R L V For hermal conducon le s dscree he spaal doman no small fne spans, =,,: Balance of parcles for an nernal

More information

On One Analytic Method of. Constructing Program Controls

On One Analytic Method of. Constructing Program Controls Appled Mahemacal Scences, Vol. 9, 05, no. 8, 409-407 HIKARI Ld, www.m-hkar.com hp://dx.do.org/0.988/ams.05.54349 On One Analyc Mehod of Consrucng Program Conrols A. N. Kvko, S. V. Chsyakov and Yu. E. Balyna

More information

Tight results for Next Fit and Worst Fit with resource augmentation

Tight results for Next Fit and Worst Fit with resource augmentation Tgh resuls for Nex F and Wors F wh resource augmenaon Joan Boyar Leah Epsen Asaf Levn Asrac I s well known ha he wo smple algorhms for he classc n packng prolem, NF and WF oh have an approxmaon rao of

More information

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4 CS434a/54a: Paern Recognon Prof. Olga Veksler Lecure 4 Oulne Normal Random Varable Properes Dscrmnan funcons Why Normal Random Varables? Analycally racable Works well when observaon comes form a corruped

More information

On computing differential transform of nonlinear non-autonomous functions and its applications

On computing differential transform of nonlinear non-autonomous functions and its applications On compung dfferenal ransform of nonlnear non-auonomous funcons and s applcaons Essam. R. El-Zahar, and Abdelhalm Ebad Deparmen of Mahemacs, Faculy of Scences and Humanes, Prnce Saam Bn Abdulazz Unversy,

More information

MANY real-world applications (e.g. production

MANY real-world applications (e.g. production Barebones Parcle Swarm for Ineger Programmng Problems Mahamed G. H. Omran, Andres Engelbrech and Ayed Salman Absrac The performance of wo recen varans of Parcle Swarm Opmzaon (PSO) when appled o Ineger

More information

Robustness Experiments with Two Variance Components

Robustness Experiments with Two Variance Components Naonal Insue of Sandards and Technology (NIST) Informaon Technology Laboraory (ITL) Sascal Engneerng Dvson (SED) Robusness Expermens wh Two Varance Componens by Ana Ivelsse Avlés avles@ns.gov Conference

More information

Stochastic Programming handling CVAR in objective and constraint

Stochastic Programming handling CVAR in objective and constraint Sochasc Programmng handlng CVAR n obecve and consran Leondas Sakalaskas VU Inse of Mahemacs and Informacs Lhana ICSP XIII Jly 8-2 23 Bergamo Ialy Olne Inrodcon Lagrangan & KKT condons Mone-Carlo samplng

More information

Department of Economics University of Toronto

Department of Economics University of Toronto Deparmen of Economcs Unversy of Torono ECO408F M.A. Economercs Lecure Noes on Heeroskedascy Heeroskedascy o Ths lecure nvolves lookng a modfcaons we need o make o deal wh he regresson model when some of

More information

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue. Lnear Algebra Lecure # Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons

More information

A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION

A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION S19 A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION by Xaojun YANG a,b, Yugu YANG a*, Carlo CATTANI c, and Mngzheng ZHU b a Sae Key Laboraory for Geomechancs and Deep Underground Engneerng, Chna Unversy

More information

A Principled Approach to MILP Modeling

A Principled Approach to MILP Modeling A Prncpled Approach o MILP Modelng John Hooer Carnege Mellon Unvers Augus 008 Slde Proposal MILP modelng s an ar, bu need no be unprncpled. Slde Proposal MILP modelng s an ar, bu need no be unprncpled.

More information

Lecture 18: The Laplace Transform (See Sections and 14.7 in Boas)

Lecture 18: The Laplace Transform (See Sections and 14.7 in Boas) Lecure 8: The Lalace Transform (See Secons 88- and 47 n Boas) Recall ha our bg-cure goal s he analyss of he dfferenal equaon, ax bx cx F, where we emloy varous exansons for he drvng funcon F deendng on

More information

( ) () we define the interaction representation by the unitary transformation () = ()

( ) () we define the interaction representation by the unitary transformation () = () Hgher Order Perurbaon Theory Mchael Fowler 3/7/6 The neracon Represenaon Recall ha n he frs par of hs course sequence, we dscussed he chrödnger and Hesenberg represenaons of quanum mechancs here n he chrödnger

More information

Math 128b Project. Jude Yuen

Math 128b Project. Jude Yuen Mah 8b Proec Jude Yuen . Inroducon Le { Z } be a sequence of observed ndependen vecor varables. If he elemens of Z have a on normal dsrbuon hen { Z } has a mean vecor Z and a varancecovarance marx z. Geomercally

More information

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data Anne Chao Ncholas J Goell C seh lzabeh L ander K Ma Rober K Colwell and Aaron M llson 03 Rarefacon and erapolaon wh ll numbers: a framewor for samplng and esmaon n speces dversy sudes cology Monographs

More information

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading Onlne Supplemen for Dynamc Mul-Technology Producon-Invenory Problem wh Emssons Tradng by We Zhang Zhongsheng Hua Yu Xa and Baofeng Huo Proof of Lemma For any ( qr ) Θ s easy o verfy ha he lnear programmng

More information

Lecture VI Regression

Lecture VI Regression Lecure VI Regresson (Lnear Mehods for Regresson) Conens: Lnear Mehods for Regresson Leas Squares, Gauss Markov heorem Recursve Leas Squares Lecure VI: MLSC - Dr. Sehu Vjayakumar Lnear Regresson Model M

More information

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue. Mah E-b Lecure #0 Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons are

More information

Reactive Methods to Solve the Berth AllocationProblem with Stochastic Arrival and Handling Times

Reactive Methods to Solve the Berth AllocationProblem with Stochastic Arrival and Handling Times Reacve Mehods o Solve he Berh AllocaonProblem wh Sochasc Arrval and Handlng Tmes Nsh Umang* Mchel Berlare* * TRANSP-OR, Ecole Polyechnque Fédérale de Lausanne Frs Workshop on Large Scale Opmzaon November

More information

Advanced time-series analysis (University of Lund, Economic History Department)

Advanced time-series analysis (University of Lund, Economic History Department) Advanced me-seres analss (Unvers of Lund, Economc Hsor Dearmen) 3 Jan-3 Februar and 6-3 March Lecure 4 Economerc echnues for saonar seres : Unvarae sochasc models wh Box- Jenns mehodolog, smle forecasng

More information

FI 3103 Quantum Physics

FI 3103 Quantum Physics /9/4 FI 33 Quanum Physcs Aleander A. Iskandar Physcs of Magnesm and Phooncs Research Grou Insu Teknolog Bandung Basc Conces n Quanum Physcs Probably and Eecaon Value Hesenberg Uncerany Prncle Wave Funcon

More information

An introduction to Support Vector Machine

An introduction to Support Vector Machine An nroducon o Suppor Vecor Machne 報告者 : 黃立德 References: Smon Haykn, "Neural Neworks: a comprehensve foundaon, second edon, 999, Chaper 2,6 Nello Chrsann, John Shawe-Tayer, An Inroducon o Suppor Vecor Machnes,

More information

Lecture 2 L n i e n a e r a M od o e d l e s

Lecture 2 L n i e n a e r a M od o e d l e s Lecure Lnear Models Las lecure You have learned abou ha s machne learnng Supervsed learnng Unsupervsed learnng Renforcemen learnng You have seen an eample learnng problem and he general process ha one

More information

Clustering (Bishop ch 9)

Clustering (Bishop ch 9) Cluserng (Bshop ch 9) Reference: Daa Mnng by Margare Dunham (a slde source) 1 Cluserng Cluserng s unsupervsed learnng, here are no class labels Wan o fnd groups of smlar nsances Ofen use a dsance measure

More information

Lecture 11 SVM cont

Lecture 11 SVM cont Lecure SVM con. 0 008 Wha we have done so far We have esalshed ha we wan o fnd a lnear decson oundary whose margn s he larges We know how o measure he margn of a lnear decson oundary Tha s: he mnmum geomerc

More information

Improvement in Estimating Population Mean using Two Auxiliary Variables in Two-Phase Sampling

Improvement in Estimating Population Mean using Two Auxiliary Variables in Two-Phase Sampling Rajesh ngh Deparmen of ascs, Banaras Hndu Unvers(U.P.), Inda Pankaj Chauhan, Nrmala awan chool of ascs, DAVV, Indore (M.P.), Inda Florenn marandache Deparmen of Mahemacs, Unvers of New Meco, Gallup, UA

More information

Fuzzy Set Theory in Modeling Uncertainty Data. via Interpolation Rational Bezier Surface Function

Fuzzy Set Theory in Modeling Uncertainty Data. via Interpolation Rational Bezier Surface Function Appled Mahemacal Scences, Vol. 7, 013, no. 45, 9 38 HIKARI Ld, www.m-hkar.com Fuzzy Se Theory n Modelng Uncerany Daa va Inerpolaon Raonal Bezer Surface Funcon Rozam Zakara Deparmen of Mahemacs, Faculy

More information

Attribute Reduction Algorithm Based on Discernibility Matrix with Algebraic Method GAO Jing1,a, Ma Hui1, Han Zhidong2,b

Attribute Reduction Algorithm Based on Discernibility Matrix with Algebraic Method GAO Jing1,a, Ma Hui1, Han Zhidong2,b Inernaonal Indusral Informacs and Compuer Engneerng Conference (IIICEC 05) Arbue educon Algorhm Based on Dscernbly Marx wh Algebrac Mehod GAO Jng,a, Ma Hu, Han Zhdong,b Informaon School, Capal Unversy

More information

Lecture 6: Learning for Control (Generalised Linear Regression)

Lecture 6: Learning for Control (Generalised Linear Regression) Lecure 6: Learnng for Conrol (Generalsed Lnear Regresson) Conens: Lnear Mehods for Regresson Leas Squares, Gauss Markov heorem Recursve Leas Squares Lecure 6: RLSC - Prof. Sehu Vjayakumar Lnear Regresson

More information

Performance Analysis for a Network having Standby Redundant Unit with Waiting in Repair

Performance Analysis for a Network having Standby Redundant Unit with Waiting in Repair TECHNI Inernaonal Journal of Compung Scence Communcaon Technologes VOL.5 NO. July 22 (ISSN 974-3375 erformance nalyss for a Nework havng Sby edundan Un wh ang n epar Jendra Sngh 2 abns orwal 2 Deparmen

More information

Existence and Uniqueness Results for Random Impulsive Integro-Differential Equation

Existence and Uniqueness Results for Random Impulsive Integro-Differential Equation Global Journal of Pure and Appled Mahemacs. ISSN 973-768 Volume 4, Number 6 (8), pp. 89-87 Research Inda Publcaons hp://www.rpublcaon.com Exsence and Unqueness Resuls for Random Impulsve Inegro-Dfferenal

More information

Solving the multi-period fixed cost transportation problem using LINGO solver

Solving the multi-period fixed cost transportation problem using LINGO solver Inernaonal Journal of Pure and Appled Mahemacs Volume 119 No. 12 2018, 2151-2157 ISSN: 1314-3395 (on-lne verson) url: hp://www.pam.eu Specal Issue pam.eu Solvng he mul-perod fxed cos ransporaon problem

More information

This document is downloaded from DR-NTU, Nanyang Technological University Library, Singapore.

This document is downloaded from DR-NTU, Nanyang Technological University Library, Singapore. Ths documen s downloaded from DR-NTU, Nanyang Technologcal Unversy Lbrary, Sngapore. Tle A smplfed verb machng algorhm for word paron n vsual speech processng( Acceped verson ) Auhor(s) Foo, Say We; Yong,

More information

Observer Design for Nonlinear Systems using Linear Approximations

Observer Design for Nonlinear Systems using Linear Approximations Observer Desgn for Nonlnear Ssems sng Lnear Appromaons C. Navarro Hernandez, S.P. Banks and M. Aldeen Deparmen of Aomac Conrol and Ssems Engneerng, Unvers of Sheffeld, Mappn Sree, Sheffeld S 3JD. e-mal:

More information

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!") i+1,q - [(!

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!) i+1,q - [(! ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL The frs hng o es n wo-way ANOVA: Is here neracon? "No neracon" means: The man effecs model would f. Ths n urn means: In he neracon plo (wh A on he horzonal

More information

M. Y. Adamu Mathematical Sciences Programme, AbubakarTafawaBalewa University, Bauchi, Nigeria

M. Y. Adamu Mathematical Sciences Programme, AbubakarTafawaBalewa University, Bauchi, Nigeria IOSR Journal of Mahemacs (IOSR-JM e-issn: 78-578, p-issn: 9-765X. Volume 0, Issue 4 Ver. IV (Jul-Aug. 04, PP 40-44 Mulple SolonSoluons for a (+-dmensonalhroa-sasuma shallow waer wave equaon UsngPanlevé-Bӓclund

More information

Advanced Machine Learning & Perception

Advanced Machine Learning & Perception Advanced Machne Learnng & Percepon Insrucor: Tony Jebara SVM Feaure & Kernel Selecon SVM Eensons Feaure Selecon (Flerng and Wrappng) SVM Feaure Selecon SVM Kernel Selecon SVM Eensons Classfcaon Feaure/Kernel

More information

Dual Approximate Dynamic Programming for Large Scale Hydro Valleys

Dual Approximate Dynamic Programming for Large Scale Hydro Valleys Dual Approxmae Dynamc Programmng for Large Scale Hydro Valleys Perre Carpener and Jean-Phlppe Chanceler 1 ENSTA ParsTech and ENPC ParsTech CMM Workshop, January 2016 1 Jon work wh J.-C. Alas, suppored

More information

Improvement in Estimating Population Mean using Two Auxiliary Variables in Two-Phase Sampling

Improvement in Estimating Population Mean using Two Auxiliary Variables in Two-Phase Sampling Improvemen n Esmang Populaon Mean usng Two Auxlar Varables n Two-Phase amplng Rajesh ngh Deparmen of ascs, Banaras Hndu Unvers(U.P.), Inda (rsnghsa@ahoo.com) Pankaj Chauhan and Nrmala awan chool of ascs,

More information

Chapter Lagrangian Interpolation

Chapter Lagrangian Interpolation Chaper 5.4 agrangan Inerpolaon Afer readng hs chaper you should be able o:. dere agrangan mehod of nerpolaon. sole problems usng agrangan mehod of nerpolaon and. use agrangan nerpolans o fnd deraes and

More information

Relative controllability of nonlinear systems with delays in control

Relative controllability of nonlinear systems with delays in control Relave conrollably o nonlnear sysems wh delays n conrol Jerzy Klamka Insue o Conrol Engneerng, Slesan Techncal Unversy, 44- Glwce, Poland. phone/ax : 48 32 37227, {jklamka}@a.polsl.glwce.pl Keywor: Conrollably.

More information

Volatility Interpolation

Volatility Interpolation Volaly Inerpolaon Prelmnary Verson March 00 Jesper Andreasen and Bran Huge Danse Mares, Copenhagen wan.daddy@danseban.com brno@danseban.com Elecronc copy avalable a: hp://ssrn.com/absrac=69497 Inro Local

More information

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s Ordnary Dfferenal Equaons n Neuroscence wh Malab eamples. Am - Gan undersandng of how o se up and solve ODE s Am Undersand how o se up an solve a smple eample of he Hebb rule n D Our goal a end of class

More information

Machine Learning Linear Regression

Machine Learning Linear Regression Machne Learnng Lnear Regresson Lesson 3 Lnear Regresson Bascs of Regresson Leas Squares esmaon Polynomal Regresson Bass funcons Regresson model Regularzed Regresson Sascal Regresson Mamum Lkelhood (ML)

More information

SOME NOISELESS CODING THEOREMS OF INACCURACY MEASURE OF ORDER α AND TYPE β

SOME NOISELESS CODING THEOREMS OF INACCURACY MEASURE OF ORDER α AND TYPE β SARAJEVO JOURNAL OF MATHEMATICS Vol.3 (15) (2007), 137 143 SOME NOISELESS CODING THEOREMS OF INACCURACY MEASURE OF ORDER α AND TYPE β M. A. K. BAIG AND RAYEES AHMAD DAR Absrac. In hs paper, we propose

More information

Robust and Accurate Cancer Classification with Gene Expression Profiling

Robust and Accurate Cancer Classification with Gene Expression Profiling Robus and Accurae Cancer Classfcaon wh Gene Expresson Proflng (Compuaonal ysems Bology, 2005) Auhor: Hafeng L, Keshu Zhang, ao Jang Oulne Background LDA (lnear dscrmnan analyss) and small sample sze problem

More information

Comb Filters. Comb Filters

Comb Filters. Comb Filters The smple flers dscussed so far are characered eher by a sngle passband and/or a sngle sopband There are applcaons where flers wh mulple passbands and sopbands are requred Thecomb fler s an example of

More information

PARTICLE SWARM OPTIMIZATION FOR INTERACTIVE FUZZY MULTIOBJECTIVE NONLINEAR PROGRAMMING. T. Matsui, M. Sakawa, K. Kato, T. Uno and K.

PARTICLE SWARM OPTIMIZATION FOR INTERACTIVE FUZZY MULTIOBJECTIVE NONLINEAR PROGRAMMING. T. Matsui, M. Sakawa, K. Kato, T. Uno and K. Scenae Mahemacae Japoncae Onlne, e-2008, 1 13 1 PARTICLE SWARM OPTIMIZATION FOR INTERACTIVE FUZZY MULTIOBJECTIVE NONLINEAR PROGRAMMING T. Masu, M. Sakawa, K. Kao, T. Uno and K. Tamada Receved February

More information

ON THE LINEAR PROGRAMMING AND GROUP RELAXATIONS OF THE UNCAPACITATED FACILITY LOCATION PROBLEM

ON THE LINEAR PROGRAMMING AND GROUP RELAXATIONS OF THE UNCAPACITATED FACILITY LOCATION PROBLEM ON THE LINEAR PROGRAMMING AND GROUP RELAXATIONS O THE UNAPAITATED AILITY LOATION PROLEM MOHAMMAD KHALIL A THESIS PRESENTED TO THE GRADUATE SHOOL O THE UNIVERSITY O LORIDA IN PARTIAL ULILLMENT O THE REQUIREMENTS

More information

Track Properities of Normal Chain

Track Properities of Normal Chain In. J. Conemp. Mah. Scences, Vol. 8, 213, no. 4, 163-171 HIKARI Ld, www.m-har.com rac Propes of Normal Chan L Chen School of Mahemacs and Sascs, Zhengzhou Normal Unversy Zhengzhou Cy, Hennan Provnce, 4544,

More information

CS286.2 Lecture 14: Quantum de Finetti Theorems II

CS286.2 Lecture 14: Quantum de Finetti Theorems II CS286.2 Lecure 14: Quanum de Fne Theorems II Scrbe: Mara Okounkova 1 Saemen of he heorem Recall he las saemen of he quanum de Fne heorem from he prevous lecure. Theorem 1 Quanum de Fne). Le ρ Dens C 2

More information

Computing Relevance, Similarity: The Vector Space Model

Computing Relevance, Similarity: The Vector Space Model Compung Relevance, Smlary: The Vecor Space Model Based on Larson and Hears s sldes a UC-Bereley hp://.sms.bereley.edu/courses/s0/f00/ aabase Managemen Sysems, R. Ramarshnan ocumen Vecors v ocumens are

More information

EEL 6266 Power System Operation and Control. Chapter 5 Unit Commitment

EEL 6266 Power System Operation and Control. Chapter 5 Unit Commitment EEL 6266 Power Sysem Operaon and Conrol Chaper 5 Un Commmen Dynamc programmng chef advanage over enumeraon schemes s he reducon n he dmensonaly of he problem n a src prory order scheme, here are only N

More information

COMPUTER SCIENCE 349A SAMPLE EXAM QUESTIONS WITH SOLUTIONS PARTS 1, 2

COMPUTER SCIENCE 349A SAMPLE EXAM QUESTIONS WITH SOLUTIONS PARTS 1, 2 COMPUTE SCIENCE 49A SAMPLE EXAM QUESTIONS WITH SOLUTIONS PATS, PAT.. a Dene he erm ll-ondoned problem. b Gve an eample o a polynomal ha has ll-ondoned zeros.. Consder evaluaon o anh, where e e anh. e e

More information

F-Tests and Analysis of Variance (ANOVA) in the Simple Linear Regression Model. 1. Introduction

F-Tests and Analysis of Variance (ANOVA) in the Simple Linear Regression Model. 1. Introduction ECOOMICS 35* -- OTE 9 ECO 35* -- OTE 9 F-Tess and Analyss of Varance (AOVA n he Smple Lnear Regresson Model Inroducon The smple lnear regresson model s gven by he followng populaon regresson equaon, or

More information

Stochastic Maxwell Equations in Photonic Crystal Modeling and Simulations

Stochastic Maxwell Equations in Photonic Crystal Modeling and Simulations Sochasc Maxwell Equaons n Phoonc Crsal Modelng and Smulaons Hao-Mn Zhou School of Mah Georga Insue of Technolog Jon work wh: Al Adb ECE Majd Bade ECE Shu-Nee Chow Mah IPAM UCLA Aprl 14-18 2008 Parall suppored

More information

Conflict-Free Routing of AGVs on the Mesh Topology Based on a Discrete-Time Model

Conflict-Free Routing of AGVs on the Mesh Topology Based on a Discrete-Time Model Conflc-Free Roung of AGVs on he Mesh Topolog Based on a Dscree-Tme Model Zeng Janang zengj@pmalnuedusg Wen-Jng Hsu hsu@nuedusg Cener for Advanced Informaon Ssems School of Compuer Engneerng Nanang Technologcal

More information

How about the more general "linear" scalar functions of scalars (i.e., a 1st degree polynomial of the following form with a constant term )?

How about the more general linear scalar functions of scalars (i.e., a 1st degree polynomial of the following form with a constant term )? lmcd Lnear ransformaon of a vecor he deas presened here are que general hey go beyond he radonal mar-vecor ype seen n lnear algebra Furhermore, hey do no deal wh bass and are equally vald for any se of

More information

Part II CONTINUOUS TIME STOCHASTIC PROCESSES

Part II CONTINUOUS TIME STOCHASTIC PROCESSES Par II CONTINUOUS TIME STOCHASTIC PROCESSES 4 Chaper 4 For an advanced analyss of he properes of he Wener process, see: Revus D and Yor M: Connuous marngales and Brownan Moon Karazas I and Shreve S E:

More information

@FMI c Kyung Moon Sa Co.

@FMI c Kyung Moon Sa Co. Annals of Fuzzy Mahemacs and Informacs Volume 8, No. 2, (Augus 2014), pp. 245 257 ISSN: 2093 9310 (prn verson) ISSN: 2287 6235 (elecronc verson) hp://www.afm.or.kr @FMI c Kyung Moon Sa Co. hp://www.kyungmoon.com

More information

HIERARCHICAL DECISIONS FOR LINEAR/NON-LINEAR DISJUNCTIVE PROBLEMS

HIERARCHICAL DECISIONS FOR LINEAR/NON-LINEAR DISJUNCTIVE PROBLEMS 2 nd Mercosur Congress on Chemcal Engneerng 4 h Mercosur Congress on Process Sysems Engneerng HIERARCHICAL DECISIONS FOR LINEAR/NON-LINEAR DISJUNCTIVE PROLEMS Jorge M. Monagna and Aldo R. Vecche * INGAR

More information

Combining Optimization and Constraint Programming

Combining Optimization and Constraint Programming Combnng Opmzaon and Consran Programmng J. N. Hooker Carnege Mellon Unvers 4 h Inernaonal Conference on Operaons Research March Havana Some work s jon wh Ignaco Grossmann Carnege Mellon Unvers Hak-Jn Km

More information

P R = P 0. The system is shown on the next figure:

P R = P 0. The system is shown on the next figure: TPG460 Reservor Smulaon 08 page of INTRODUCTION TO RESERVOIR SIMULATION Analycal and numercal soluons of smple one-dmensonal, one-phase flow equaons As an nroducon o reservor smulaon, we wll revew he smples

More information

ON THE WEAK LIMITS OF SMOOTH MAPS FOR THE DIRICHLET ENERGY BETWEEN MANIFOLDS

ON THE WEAK LIMITS OF SMOOTH MAPS FOR THE DIRICHLET ENERGY BETWEEN MANIFOLDS ON THE WEA LIMITS OF SMOOTH MAPS FOR THE DIRICHLET ENERGY BETWEEN MANIFOLDS FENGBO HANG Absrac. We denfy all he weak sequenal lms of smooh maps n W (M N). In parcular, hs mples a necessary su cen opologcal

More information

Fall 2010 Graduate Course on Dynamic Learning

Fall 2010 Graduate Course on Dynamic Learning Fall 200 Graduae Course on Dynamc Learnng Chaper 4: Parcle Flers Sepember 27, 200 Byoung-Tak Zhang School of Compuer Scence and Engneerng & Cognve Scence and Bran Scence Programs Seoul aonal Unversy hp://b.snu.ac.kr/~bzhang/

More information

Including the ordinary differential of distance with time as velocity makes a system of ordinary differential equations.

Including the ordinary differential of distance with time as velocity makes a system of ordinary differential equations. Soluons o Ordnary Derenal Equaons An ordnary derenal equaon has only one ndependen varable. A sysem o ordnary derenal equaons consss o several derenal equaons each wh he same ndependen varable. An eample

More information

arxiv: v1 [math.oc] 11 Dec 2014

arxiv: v1 [math.oc] 11 Dec 2014 Nework Newon Aryan Mokhar, Qng Lng and Alejandro Rbero Dep. of Elecrcal and Sysems Engneerng, Unversy of Pennsylvana Dep. of Auomaon, Unversy of Scence and Technology of Chna arxv:1412.374v1 [mah.oc] 11

More information

Sampling Procedure of the Sum of two Binary Markov Process Realizations

Sampling Procedure of the Sum of two Binary Markov Process Realizations Samplng Procedure of he Sum of wo Bnary Markov Process Realzaons YURY GORITSKIY Dep. of Mahemacal Modelng of Moscow Power Insue (Techncal Unversy), Moscow, RUSSIA, E-mal: gorsky@yandex.ru VLADIMIR KAZAKOV

More information

Introduction to Boosting

Introduction to Boosting Inroducon o Boosng Cynha Rudn PACM, Prnceon Unversy Advsors Ingrd Daubeches and Rober Schapre Say you have a daabase of news arcles, +, +, -, -, +, +, -, -, +, +, -, -, +, +, -, + where arcles are labeled

More information

Outline. Energy-Efficient Target Coverage in Wireless Sensor Networks. Sensor Node. Introduction. Characteristics of WSN

Outline. Energy-Efficient Target Coverage in Wireless Sensor Networks. Sensor Node. Introduction. Characteristics of WSN Ener-Effcen Tare Coverae n Wreless Sensor Newors Presened b M Trà Tá -4-4 Inroducon Bacround Relaed Wor Our Proosal Oulne Maxmum Se Covers (MSC) Problem MSC Problem s NP-Comlee MSC Heursc Concluson Sensor

More information

Anomaly Detection. Lecture Notes for Chapter 9. Introduction to Data Mining, 2 nd Edition by Tan, Steinbach, Karpatne, Kumar

Anomaly Detection. Lecture Notes for Chapter 9. Introduction to Data Mining, 2 nd Edition by Tan, Steinbach, Karpatne, Kumar Anomaly eecon Lecure Noes for Chaper 9 Inroducon o aa Mnng, 2 nd Edon by Tan, Senbach, Karpane, Kumar 2/14/18 Inroducon o aa Mnng, 2nd Edon 1 Anomaly/Ouler eecon Wha are anomales/oulers? The se of daa

More information

Linear Response Theory: The connection between QFT and experiments

Linear Response Theory: The connection between QFT and experiments Phys540.nb 39 3 Lnear Response Theory: The connecon beween QFT and expermens 3.1. Basc conceps and deas Q: ow do we measure he conducvy of a meal? A: we frs nroduce a weak elecrc feld E, and hen measure

More information

A Tour of Modeling Techniques

A Tour of Modeling Techniques A Tour of Modelng Technques John Hooker Carnege Mellon Unversy EWO Semnar February 8 Slde Oulne Med neger lnear (MILP) modelng Dsuncve modelng Knapsack modelng Consran programmng models Inegraed Models

More information

TSS = SST + SSE An orthogonal partition of the total SS

TSS = SST + SSE An orthogonal partition of the total SS ANOVA: Topc 4. Orhogonal conrass [ST&D p. 183] H 0 : µ 1 = µ =... = µ H 1 : The mean of a leas one reamen group s dfferen To es hs hypohess, a basc ANOVA allocaes he varaon among reamen means (SST) equally

More information

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5 TPG460 Reservor Smulaon 08 page of 5 DISCRETIZATIO OF THE FOW EQUATIOS As we already have seen, fne dfference appromaons of he paral dervaves appearng n he flow equaons may be obaned from Taylor seres

More information

FTCS Solution to the Heat Equation

FTCS Solution to the Heat Equation FTCS Soluon o he Hea Equaon ME 448/548 Noes Gerald Reckenwald Porland Sae Unversy Deparmen of Mechancal Engneerng gerry@pdxedu ME 448/548: FTCS Soluon o he Hea Equaon Overvew Use he forward fne d erence

More information

Single-loop System Reliability-Based Design & Topology Optimization (SRBDO/SRBTO): A Matrix-based System Reliability (MSR) Method

Single-loop System Reliability-Based Design & Topology Optimization (SRBDO/SRBTO): A Matrix-based System Reliability (MSR) Method 10 h US Naonal Congress on Compuaonal Mechancs Columbus, Oho 16-19, 2009 Sngle-loop Sysem Relably-Based Desgn & Topology Opmzaon (SRBDO/SRBTO): A Marx-based Sysem Relably (MSR) Mehod Tam Nguyen, Junho

More information

Chapter 6: AC Circuits

Chapter 6: AC Circuits Chaper 6: AC Crcus Chaper 6: Oulne Phasors and he AC Seady Sae AC Crcus A sable, lnear crcu operang n he seady sae wh snusodal excaon (.e., snusodal seady sae. Complee response forced response naural response.

More information

Epistemic Game Theory: Online Appendix

Epistemic Game Theory: Online Appendix Epsemc Game Theory: Onlne Appendx Edde Dekel Lucano Pomao Marcano Snscalch July 18, 2014 Prelmnares Fx a fne ype srucure T I, S, T, β I and a probably µ S T. Le T µ I, S, T µ, βµ I be a ype srucure ha

More information

Planar truss bridge optimization by dynamic programming and linear programming

Planar truss bridge optimization by dynamic programming and linear programming IABSE-JSCE Jon Conference on Advances n Brdge Engneerng-III, Augus 1-, 015, Dhaka, Bangladesh. ISBN: 978-984-33-9313-5 Amn, Oku, Bhuyan, Ueda (eds.) www.abse-bd.org Planar russ brdge opmzaon by dynamc

More information

, t 1. Transitions - this one was easy, but in general the hardest part is choosing the which variables are state and control variables

, t 1. Transitions - this one was easy, but in general the hardest part is choosing the which variables are state and control variables Opmal Conrol Why Use I - verss calcls of varaons, opmal conrol More generaly More convenen wh consrans (e.g., can p consrans on he dervaves More nsghs no problem (a leas more apparen han hrogh calcls of

More information