Minimization of the Expected Total Net Loss in a Stationary Multistate Flow Network System

Size: px
Start display at page:

Download "Minimization of the Expected Total Net Loss in a Stationary Multistate Flow Network System"

Transcription

1 Appled Mathematcs, 6, 7, Publshed Onlne May 6 n ScRes. Mnmzaton of the Expected Total Net Loss n a Statonary Multstate Flow Networ System Krstna Sutlaberg, Bent Natvg Department of Mathematcs, Unversty of Oslo, Oslo, Norway Receved 5 March 6; accepted May 6; publshed 4 May 6 Copyrght 6 by authors and Scentfc Research Publshng Inc. Ths wor s lcensed under the Creatve Commons Attrbuton Internatonal Lcense (CC BY). Abstract In the present paper, a three-component, statonary, multstate flow networ system s studed. Detaled costs and ncomes are specfed. The am s to mnmze the expected total net loss wth respect to the expected tmes the components spend n each state. Ths represents a novelty n that we connect the expected component tmes spent n each state to the mnmal total net loss of the system, wthout frst fndng the component mportance. Ths s of nterest n the desgn phase where one may tune the components to mnmze the expected total net loss. Due to the complex nature of the problem, we frst study a smplfed verson. There the expected tmes spent n each state are assumed equal for each component. Then a modfed verson of the full model s presented. The optmzaton n ths model s completed n two steps. Frst the optmzaton s carred out for a set of pre-chosen fxed expected lfe cycle lengths. Then the overall mnmum s dentfed by varyng these expectatons. Both the smplfed and the modfed optmzaton problems are nonlnear. The setup used n ths artcle s such that t can easly be modfed to represent other flow networ systems and cost functons. The challenge les n the optmzaton of real lfe systems. Keywords Relablty, Nonlnear Optmzaton, Multstate Flow Networ. Introducton A seres of challenges concernng relablty engneerng s presented n []. Some of these challenges are connected to the representaton and modelng of complex systems, such as multstate systems, and ther How to cte ths paper: Sutlaberg, K. and Natvg, B. (6) Mnmzaton of the Expected Total Net Loss n a Statonary Multstate Flow Networ System. Appled Mathematcs, 7,

2 K. Sutlaberg, B. Natvg operatonal tass, for nstance mantenance optmzaton. Over the past decades varous measures of component mportance have been studed. The use of such measures permts the relablty analyst to prortze the system components n order to allocate resources effcently. In [] a new theory for measures of mportance of system components s presented. Generalzatons of the Brnbaum, Barlow-Proschan and Natvg measures (see [3]-[5] respectvely) from the bnary to the multstate case, both for unreparable and reparable systems are covered. A numercal study of the above mentoned multstate measures of component mportance s also covered n []. Loss of utlty due to the system leavng the dfferent sets of better states are ntroduced n that study. However, no detaled costs or ncomes are specfed. Recently, wor has been done to also nclude costs n the determnaton of component mportance for bnary systems. In [6] and [7] the Brnbaum measure s extended to also nclude both falure nduced and mantenance costs, whle [8] and [9] ntroduce other cost-effectve mportance measures. In mantenance optmzaton studes one s often nterested n choosng a mantenance plan whch mnmzes lfe cycle costs, maxmzes net present value or maxmzes system relablty for a gven system. See for nstance []-[4] for some recent wor on these subjects. In ths artcle we wll loo at one partcular type of mantenance acton, the complete repar. As the components reach the complete falure state, they are repared to what we wll denote the perfect functonng state. The am s to nclude both costs and ncomes n the study of a reparable multstate flow networ system. To acheve ths, we wll defne ncomes and cost functons for the purpose of mnmzng the expected total net loss over a tme perod wth respect to the expected component tmes n the dfferent states. Ths represents a novelty n that we connect the expected component tmes spent n each state to the mnmal total net loss of the system, wthout frst fndng the component mportances. It would of course have been nce to optmze wth respect to probablty dstrbutons nstead of expectatons, but ths s not trval even for a smple three-component system. However, the optmzaton problem consdered n ths artcle s partcularly nterestng n a desgn or re-desgn phase, where one may tune the components n such a way that the expected total net loss s mnmzed. Wth the optmzaton problem consdered n ths artcle we are facng complex dependences. We therefore study both a smplfed verson and a modfed verson of the optmzaton problem. In the smplfed verson we see that the optmal expected tme spent n each state ncreases wth ncreasng operatonal tme for all three cost functon types consdered. However, the extent of the ncrease dffers wth the dfferent basc cost functon types. Due to basc nvestment costs ths s not a trval result. In the modfed verson of the optmzaton problem we only fnd approxmate solutons. We observe that the dfferent types of cost functons nfluence the end results sgnfcantly. For nstance one of the functonng states s redundant for two of the three cost functon types when the cost functon parameter s ncreasng. For both problems we see that the mnmum expected total net loss s ncreasng wth ncreasng component cost per repar. The rest of the artcle s organzed as follows: Secton ntroduces the basc model, the three dfferent types of cost functons and the three-component system of nterest. The smplfed verson of the optmzaton problem wth results s presented n Secton 4. Secton 5 presents the modfed optmzaton problem wth results, and concludng remars are found n Secton 6.. Basc Model Let S be the set of possble system states, and S, =,, n, the set of possble component states. Throughout ths artcle we wll assume that S = S = {,, M}. Snce we are regardng the system as a flow networ, the system state s the amount of flow that can be transported through the networ. In the same way, the component state s the amount of flow that can be transported through each component. Let X ( t) = ( X ( t),, Xn ( t) ) be the vector of component states at tme t. That s X ( t) = f component, =,,, n, s n state S at tme t. A bnary mnmal cut set s a mnmal set of components whch upon falure wll brea the connecton between the endponts of the networ. Let K l, l =,, m, be the bnary mnmal cut sets of the networ. Then, by applyng the max-flow-mn-cut theorem (see [5]), we get that the system state s gven by φ ( X ( t) ) = mn X ( ). t () l m K l 794

3 K. Sutlaberg, B. Natvg Thus, the system state equals the smallest total flow through the mnmal cut sets of the system. Assume now that no components are n seres wth the rest of the system. Then there must be at least two components n every mnmal cut set. If all components are n the perfect functonng state, M, the system state wll be at least M, and therefore we must have that S S. Thus, the assumpton of equalty between the set of system states and the set of component states, S = S, =,, n, mples that at least one component s n seres wth the rest of the system. For ths reason, we wll n Sectons 4 and 5 focus on the three component system gven n Fgure. Assume that the components deterorate by gong through all states n S, from the perfect functonng state M to the complete falure state, before beng repared bac to M. Let be the expected tme component =,, n spends n state =,, M, and let the vector of postve expected component tmes n each state be denoted (,,, M M =,,, n ). Assume for =,, n that the basc nvestment costs of component spendng the expected amount of tme n state,, M g for,, M h for =. = are gven by the cost functons ( ) These basc costs appear once n the tme nterval [ ] = and ( ),T for each combnaton of =,, n and =,, M. For any gven functonng state =,, M, t seems natural that these basc expenses grow when the expected tmes become large. If however, no tme s spent n a functonng state, there wll not be any basc costs of eepng the component n ths state. Smlarly, the shorter the expected tme spent n the complete falure state, the more expensve t should be. In other words, the faster a complete repar s executed, the more g s ncreasng and the cost functon expensve t should be. Therefore, we assume that the cost functon ( ) ( ) g( ) = h( ) =. Throughout ths artcle we assume the cost functons, g ( ) and h ( ) h s assumed to be decreasng; moreover, both functons are assumed to be twce dfferentable wth the followng types: g c Type : ( ) = and h( ) = c. Type : ( ) ln ( g = c + ) and h( ) = ( c + ) Type 3: ( ) e c g c = and ( ) e h =, ln., to be of one of where c >, =,, n, =,, M are constants. These cost functons are constructed by the authors accordng to the above mentoned crtera to represent a varaton n the potental basc cost development. In ths artcle we only consder perfect repars. Let C > denote the cost per repar from the complete falure state to the perfect functonng state of component =,,, 3. The total number of repars of component =,,, 3 n the nterval [,T ] s denoted N ( T ) for =,, n. Let R be the fxed ncome per unt of tme when the system s n state S, and assume that R < R R M. Ths means that the ncome decreases, startng from the perfect functonng state, from one state to the next untl the system reaches the complete falure state, where the ncome s non-postve. Thus, there s a loss per tme unt that the system spends n the complete falure state. Such negatve ncome mght correspond to nterest rate expenses connected to system buldng nvestments. The presence of such costs wll ncrease the ncentve for reparng the faled components. Fgure. A system wth three components. 795

4 K. Sutlaberg, B. Natvg The contrbuton from the -th component to the total cost connected to the operaton of the system n the tme CN T, n addton to the basc nvestment costs nterval [,T ], s the total repar cost over the nterval, ( ) related to component spendng the expected amount of tme n state,, M. M =, ( ) ( ) g + h = To get the total costs connected to the operaton of the system we sum over all components. Let I{ A } denote the ndcator functon of the event A. Then the ncome at tme t connected to the operaton of the system s gven by M = { φ ( ( t) ) = } I X R. To fnd the total ncome we ntegrate the ncome at tme t over the tme perod [,T ]. Hence, the total net loss connected to the operaton of the system n the tme nterval [,T ] s n M M T ( ) = ( ) + ( ) + ( ) { φ( X ( )) = } M CN T g h I t R d. t = = = Note that a negatve net loss equals a postve net gan. By tang the expectaton we fnd that the correspondng objectve functon s n M M T ( ) = ( ) = ( ) + ( ) + ( ) ( φ( X ( )) = ) O E M C E P d. N T g h t R t () = = = In the remanng parts of ths artcle, we wll focus on statonary multstate systems. Component avalabltes are now gven by a = lm P( X( t) = ) =, t M l l= for =,, n and S. The statonary system avalabltes are gven by Let (,, M M = a a, a,, a n ) ( φ ( ( )) ) a = lmp X t =. (4) φ t a denote the vector of component avalabltes. When the components operate ndependently the statonary system avalabltes, a φ, equals ( ) a φ a for =,, M. The expected number of repars of component s now gven by ( ) = objectve functon, gven n (), therefore becomes ( ) ( ) ( ) M (3) l EN T = T, =,, n. The n M M M l O = CT + g + h aφ RT, (5) = l= = = whch s determned explctly by, =,, n, =,, M. Thus, the optmzaton problem that we wll consder s to mnmze (5) wth respect to, =,, n and =,, M wth dfferent cost functons g ( ) and h ( ). 3. The Three-Component System For smplcty, the system we wll focus on, s the multstate flow networ system consstng of three components where component s n seres wth the parallel structure of components and 3 (see Fgure ). We wll S = S =,,. assume that all components and the system are n one of three states, that s we assume { } The structure functon of the module consstng of components and 3 n parallel s ( x, x ) mn (, x x ) whereas the structure functon of the system s l ψ = +,

5 K. Sutlaberg, B. Natvg ( ) = ( x ψ ( x x3) ) = ( x x + x3) φ x mn,, mn,, (6) snce x. For the system to be n the perfect functonng state, state, both modules, that s both component and the parallel structure of components and 3, must be n the perfect functonng state. For the system to be n state both modules must be functonng, and at least one of the modules must be functonng at level =. For the system to be n the complete falure state, at least one of the modules must be n the complete falure state. The system avalabltes are hence gven by a = a aa + aa + aa + aa + aa + aa = a aa aa aa 4. The Smplfed Problem φ ( ) ( ) ( ) ( ) ( )( 3 ) a = a aa + a aa + aa φ a = a aa = a + aa aaa. φ Because of the complex nature of the problem presented n Secton we frst study a smplfed verson of the problem. Assume the expected tmes spent n each state to be equal for each component. That s, we assume =, for =,, and =,, 3. It s now natural to also assume c = c for =,, and =,, 3. As a consequence, the component avalabltes are equal to a = for all and. Thus, the system 3 avalabltes are equal to a φ = a φ = a φ = As a consequence, the total ncome, gven by the last term n the objectve functon (5), s constant. Let =,, be the vector of expected component tmes. The smplfed objectve functon s gven by ( ) 3 and the correspondng optmzaton problem s 3 φ = = ( ) ( ) ( ) O = CT 3 + g + h a RT, (8) mnmze O ( ) subject to, =,, 3. Ths s a box constraned nonlnear optmzaton problem. Note that the sum ndependent of and wll therefore not affect the mnmum. 4.. Analyss of Convexty = φ Let H O ( ) denote the Hessan matrx related to the objectve functon (8). Ths s a 3 3 matrx. H O ( ) O O O ( ) 3 D O O O = = D ( ) 3 D 3 O O O ( ), (7) (9) a RT, n (8), s 797

6 K. Sutlaberg, B. Natvg where ( ) g( ) ( ) h( ) ( ) CT D = The objectve functon s convex f and only f the Hessan matrx s postve semdefnte (see for nstance [6]). In our case, the Hessan matrx s a dagonal matrx wth D=,,, 3 on the dagonal. Hence, f all the dagonal elements are non-negatve,.e. D, =,, 3, then the objectve functon s convex and a local mnmum wll also be the global mnmum. 4.. The Objectve Functons 4... Type Cost Functons Let the cost functons be gven by g( ) = c and h( ) now gven by () = c respectvely. The objectve functon (8) s 3 c O( ) = CT 3 + c + aφ RT, = () = and the dagonal elements of the Hessan matrx are, for =,, 3, CT c D = + > Thus, n ths case, the objectve functon s convex. Dfferentatng O ( ) gves ( ( )) For =,, 3 the optmal s gven by ( ) ( ) O CT c = + c. 3 C T = +. () c 6 We see from () that the optmal, =,, 3 s dependng on TC, and c. T s the operatonal tme perod, C s the cost per repar of component and c s a constant connected to the basc nvestment costs of component spendng the expected amount of tme,, n state. Furthermore, the optmal s ndependent of j, j. It s ncreasng n T, as seen by the sold lne n Fgure, ncreasng n C, as seen n Table for = and as the basc nvestment cost parameter, c, ncreases, the optmal decreases. Ths s also seen n Table for =. These latter results are reasonable Type Cost Functons Let for,, 3 = the cost functons be logarthmc and gven by g( ) = ln ( c + ) and h( ) ln ( c ) The objectve functon (8) becomes Table. Cost functons of type. C = C3 = c = c3 =., T =, R =., R =., R =. = 5.. Theoretcal s gven by the expresson (). = +., startng values C c C c Theoretcal Computatonal Mn. expected total net loss

7 K. Sutlaberg, B. Natvg Fgure. Optmal as functon of the operatonal tme T for cost functons of type,,3. C = c =. for all. 3 CT O( ) = + ln ( c + ) + ln ( c + ) aφ RT, = 3 (3) = The dagonal elements, (), n the Hessan matrx are n ths case gven by for =,, 3. The numerator, A, s gven by ( c + ) CT c c A D = + = ( c + ) c + c + 3 ( c ) ( ) ( ) + 3 ( ) ( ) ( ) ( )( ) A= CT c + c + 6c c + + 3c c + c +. Ths s a 5-th order polynomal n. As grows large A s domnated by the 5 term whch has negatve sgn. Hence, the numerator, and D, are negatve for large. On the other hand, when approaches, D s postve. The objectve functon (3) s thus nether convex nor concave. We see that O approaches nfnty when, =,, 3, approaches ether or. Therefore the objectve functon (3) ( ) has mnmum values. Dfferentatng (3) wth respect to, =,, 3 gves the followng: The solutons to O ( ) = ( ) O CT c c = c c are the zeroes of the thrd degree polynomal n 799

8 K. Sutlaberg, B. Natvg ( ) ( ( )) c + c c CT c + CT + c CTc (4) , whch can be solved numercally. Every thrd degree polynomal has at least one real root, and snce expected tme component spends n each state, we are only nterested n postve solutons of Type 3 Cost Functons c The cost functons are n ths secton gven by ( ) e g,,, 3 h c ( ) e,,, 3 O ( ) s the =. = = for the two functonng states, and = = for the complete falure state. The objectve functon (8) now becomes 3 CT c c O( ) = + e + e 9 aφ RT = 3 (5) = The dagonal elements of the Hessan matrx are, for all =,, 3, gven by CT c c D = + ( c ) e + + e >. 3 c c 3 3 ( ) ( ) Hence, (5) s convex and therefore t has a global mnmum value Results In ths secton the ncomes per tme unt are chosen to be R =., R =. and R =.. For =,, 3 the startng values for the numercal computatons are chosen to be = 5.. The assumpton, =, =,, 3, =,,, mples that the total ncome term n the objectve functon, (8), s constant, as has already been stated. Thus, the optmal 's only depend on the parameter values, and not on the structural placements of the components. Therefore, only results for component are gven n the followng Effect of T Fgure shows the development of the optmal expected tmes spent n each state (the optmal, =,, 3 ) as functon of the operatonal tme T. Note that n ths case, because of the chosen parameter values C = c =., =,,3, from () vald for cost functon of type, the optmal = = 3 =, and the optmal expected lfe cycle length s 3 for all components. In Fgure we see that, and hence the optmal lfe cycle length, ncreases when the operatonal tme ncreases. Due to the basc nvestment costs ths s not a trval result. However, the extent of the ncrease dffers wth the dfferent cost functon types. For cost functons of type we see some ncrease n the optmal expected. Ths s n complance wth (). We also see that s by far the largest for cost functons of type for all T. For cost functons of type 3, on the other hand, we only observe a slght ncrease n the optmal as T ncreases. From Fgure 3 we see that the mnmum expected net loss as functon of the operatonal tme T behaves dfferently wth dfferent types of cost functons. For cost functons of type and, the mnmum expected net loss s decreasng wth ncreasng operatonal tme T. For type 3 the mnmum expected net loss s ncreasng at frst before t starts to decrease. The mnmum expected net loss s postve for T > when we use cost functons of type Effect of C For cost functons of type, we see from Table that the theoretcal results gven by () are equal to the computatonal results. For constant C c the theoretcal and computatonal results are also constant, whch s n accordance wth (). Even though C c s held constant (see Table for C c =. ), we see an ncrease n the mnmum expected net loss. The mnmum expected net loss s dependent on the values of C and c respectvely. We see that when these values ncrease, the net loss ncreases, as s obvous from (). For cost functons of type we found n Secton.. that the optmal s are the zeroes of the cubc 8

9 K. Sutlaberg, B. Natvg Fgure 3. Mnmum expected net loss as functon of the operatonal tme T for cost functons of type,,3. C = c =. for all. polynomal gven n (4). For the parameter values n Table ths polynomal has one postve root, whch equals the results obtaned from the optmzaton routne. We see an ncrease n the optmal as C ncreases. Fgure 4 shows the development of the mnmum expected net loss as the repar costs, C, of component ncreases. We see that the mnmum expected net loss s negatve for cost functons of type. That s, we have a postve maxmum expected net gan for ths cost functon. For the other two types of cost functons we have a postve mnmum expected net loss. For all C, the loss s greater for cost functons of type 3 than t s for cost functons of type. The correspondng optmal, and 3 are shown n Fgure 5. From Fgure 5 we see that for cost functons of type t s optmal to spend longer tme n each state than t s for the other two cost functons. The optmal expected tme spent n each state for component s ncreasng wth ncreasng repar costs C. Ths seems natural. The results for components and 3 are equal because the parameter values concernng these two components are equal. From the rght plot n Fgure 6 we see that the ncrease n the repar costs of component has no nfluence on the optmal expected tme spent n each state for components and 3, thus we see that the optmal, =,3 are constant Effect of c Fgure 6 shows the mnmum expected net loss as a functon of the cost functon parameter c. The mnmum expected net loss s ncreasng wth an ncreasng c. We see that an ncrease n c has much larger effect on the mnmum expected net loss when we use cost functons of type 3 than when we use the other two types of cost functons. Ths s natural snce the type 3 cost functons are exponental. The correspondng optmal, =,, 3 are shown n Fgure 7. We see that s constant., on the other hand, seems to be decreasng wth ncreasng c when we have cost functons of type and 3. Ths behavor seems reasonable. For cost functons of type we see that eventually starts to ncrease when c ncreases, whch seems unnatural. 8

10 K. Sutlaberg, B. Natvg Fgure 4. Mnmum expected net loss as functon of C for cost functons of type,,3. C = C3 = c =. for all and T =. Fgure 5. Optmal as functon of C for cost functons of type,,3. C = C3 = c =. for all and T =. 8

11 K. Sutlaberg, B. Natvg Fgure 6. Mnmum expected net loss as functon of c for cost functons of type,,3. c = c3 = C =. for all and T =. Fgure 7. Optmal as functon of c for cost functons of type,,3. c = c3 = C =. for all and T =. 83

12 K. Sutlaberg, B. Natvg Table. Cost functons of type. C = C3 = c = c = c3 =., T =, R =., R =., R =., startng values = 5.. Theoretcal s the root of the polynomal (4). C Theoretcal Computatonal Mn. expected total net loss Modfcatons of the Full Model Optmzaton Problem The orgnal problem, represented by the objectve functon (5), turned out to be qute complex even though we only consdered a smple three-component system wth three possble system and component states. Thus, the optmzaton of ths problem was not straghtforward. In order to overcome dffcultes wth startng value senstve optmzaton results, we reformulated the orgnal optmzaton problem n order to fnd an approxmate soluton. Let α = + +, =,, 3, be the fxed expected length of the lfe cycle of component. Then, the objectve functon s gven by where (,,, 3 ) 3 φ = = ( ) α ( ) ( ) ( ) O = CT + g + g + h a RT, (6) = 9 s a vector of the expected tmes spent n each state for each component. We are nterested n mnmzng the expected total net loss (6) subject to the fxed expected length of the lfe cycles α, =,, 3. The optmzaton of the modfed problem goes as follows: Step : Choose values for α, =,, 3. For every combnaton of these α, =,, 3, the nonlnear optmzaton problem wth both equalty and nequalty constrants, (7), s solved. mnmze μ O ( ) subject to α =, =,, 3. =, =,,3, =,, Step : Identfy the overall mnmum from the optmzaton results from step. The correspondng wll approxmately mnmze the expected total net loss over the tme perod [,T ]. Optmzaton problems as the ones n step may be solved usng the augmented Lagrange multpler method, for nstance usng the SOLNP algorthm as descrbed n [7]. Ths algorthm s mplemented n the Rsolnp pacage, see [8], n R. Note that mnmzng the expected total net loss s equvalent to maxmzng the expected total net gan, and that a negatve expected total net loss s a postve expected net gan. Snce we are usng mnmzaton algorthms nstead of maxmzaton algorthms, the focus has been on mnmzng the total net loss rather than maxmzng the total net gan. 5.. Results In ths secton lower bounds on, =,, 3 and =,, are chosen to be -5 for cost functons of type and, and. for cost functons of type 3. The upper bound s chosen to be equal to the operatonal tme T. As n the prevous secton, the ncomes per tme unt when the system s n state =,,, are R =., R =. and R =. respectvely. The possble lfe cycle lengths are chosen to be α = ( 4, 6,8,,,4,6 ), =,,3, and the startng values are =. for =,, 3 and =,,. Component and component 3 are n parallel. Ther roles n the system are therefore nterchangeable. Snce (7) 84

13 K. Sutlaberg, B. Natvg we assume that the components cost functons are of the same type and that we are varyng one parameter at a tme, we are n the followng only varyng the parameters connected to component and component. When the parameters of component are vared the results for components and 3 are dentcal. Hence, results for component 3 are then omtted Effect of T Fgure 8 shows the mnmum expected net loss as a functon of T. We see that wth cost functons of type and, the mnmum expected net loss s negatve and decreasng for the chosen values of T. Ths means that for these cost functons we have an ncreasng maxmum expected net gan. The loss s smaller for cost functons of type than t s for the other two types of cost functons. For cost functons of type 3 the mnmum expected net loss s postve for small T. The correspondng optmal s, =, and =,,, are gven n Fgure 9. It seems le the optmal s stablzes as T becomes large Effect of an Increasng Cost Per Repar C, =, For ths, and the followng sectons, the operatonal tme s set to T = 3. Fgure and Fgure shows the mnmum expected net loss as a functon of the repar cost of component and respectvely. We see that the mnmum expected net loss s ncreasng wth ncreasng repar costs C, =,, for all three cost functons. Ths seems natural. The correspondng optmal, =,, =,, as functons of C are shown n Fgure and for C n Fgure 3 for =,, 3, =,,. As the repar costs of component, C, ncreases, the optmal expected lfe cycle length ( α = ) = for component ncreases for cost functon types and 3. For cost functons of type the optmal,,, = are constant. For cost functons of type and type, t s optmal to eep Fgure 8. Mnmum expected net loss as functon of T for cost functons of type,,3. C = C = C3 = c =. for =,,3, =,,. 85

14 K. Sutlaberg, B. Natvg Fgure 9. Optmal =,,3, =,,., =,, =,,, as functon of T for cost functons of type,,3. C = C = C3 = c =. for Fgure. Mnmum expected net loss as functon of C for cost functons of type,,3. C = C3 = c =. for =,,3, =,,. 86

15 K. Sutlaberg, B. Natvg Fgure. Mnmum expected net loss as functon of C for cost functons of type,,3. C = C3 = c =. for =,,3, =,,. Fgure. Optmal, =,, =,,, as functon of C for cost functons of type,,3. C = C3 = c =. for =,,3, =,,. 87

16 K. Sutlaberg, B. Natvg Fgure 3. Optmal, =,,3, =,, as functon of C for cost functons of type,,3. C = C3 = c =. for =,,3, =,,. component n the perfect functonng state for as long as possble. Hence, we see a large for these two cost functon types. For cost functons of type 3, on the other hand, the ncrease n expected lfe cycle length s placed n state =, and we see an ncrease n for ths cost functon. The extra costs connected to an ncrease n the expected tmes spent n ether of the two functonng states are much larger for cost functons of type 3 than for the other two types of cost functons. As the repar costs of component, C, ncreases, we see from Fgure 3 that the results for component are constant. Hence, for component the optmal expected tme spent n each state s ndependent of the repar cost of component. We also see that an ncreasng repar cost results n an ncreasng optmal expected lfe cycle length for component for all cost functons. For cost functons of type and the ncrease s on the expected tme spent n state,, whle the ncrease s on the expected repar tme,, for cost functons of type 3. Snce component s crtcal to the functonng of the system we see from Fgure an ncrease n for cost functons of type 3 as C ncreases. Component s n parallel wth component 3. It s therefore possble to extend the expected repar tme of ths component when C ncreases, whle at the same tme the expected repar tmes of component 3,, are ept low. Ths s seen n Fgure 3 for cost functons of type Effect of an Increasng c, =,, As the cost functon parameter c, =, for component ncreases, the mnmum expected net loss also ncreases for cost functons of type and 3. The mnmum expected net loss remans unchanged when cost functons of type are used. See Fgure 4 and Fgure 5. An ncrease n c,, = has largest mpact on the mnmum expected net loss when the cost functons are of type 3, that s when we have exponental cost 88

17 K. Sutlaberg, B. Natvg Fgure 4. Mnmum expected net loss as functon of c for cost functons of type,,3. C =. for =,,3 and c =. for =,3, =,, and =, =,. Fgure 5. Mnmum expected net loss as functon of c for cost functons of type,,3. C =. for =,,3 and c =. for =,3, =,, and =, =,. 89

18 K. Sutlaberg, B. Natvg functons. In state = there was no ncrease n the mnmum expected net loss wth ncreasng Fgure 6. Fgures 7-9 show the development n the optmal ncreases. The effect of an ncreasng c s evdent n the ncreasng optmal seen n Fgure 7, where there s also a slght ncrease n and c, as seen n, =,, =,, as c,,, = for cost functons of type 3. Ths s for cost functons of type. The behavour of dffers for the three cost functon types as the cost functon parameter c ncreases. In state =, s close to for both cost functons of type and, whle postve for c > for cost functons of type 3. For = the optmal s hgh and (close to) constant for cost functons of type (). For cost functons of type 3 the optmal s lower. An ncreasng c has no effect on the optmal, =,, =,,. Ths s seen n Fgure 8. As c ncreases, we see from Fgure 9 that the optmal expected tme spent n state for component,, s decreasng for both cost functons of type and 3. As the costs of eepng at a fxed level s ncreasng, t becomes less desrng to mantan ths level, and we see a decrease. The decrease s faster n for cost functons of type 3. For ths cost functon we see an ncrease n wth ncreasng c that s n contrast to the results for cost functons of type and whch are ndependent of the ncrease n to. c and equal Fgure 6. Mnmum expected net loss as functon of c for cost functons of type,,3. C =. for =,,3 and c =. for =,3, =,, and =, =,. 8

19 K. Sutlaberg, B. Natvg Fgure 7. Optmal, =,, =,, as functon of c =. for =,3, =,, and =, =,. c for cost functons of type,,3. C =. for =,,3 and Fgure 8. Optmal, =,, =,, as functon of c for cost functons of type,,3. C =. for all and c =. for =,3, =,, and =, =,. 8

20 K. Sutlaberg, B. Natvg Fgure 9. Optmal, =,, =,, as functon of c =. for =,3, =,, and =, =,. c for cost functons of type,,3. C =. for =,,3 and Effect of an Increasng c, =,, An ncreasng c, =,, has lttle effect on the mnmal expected net loss. Ths s seen n Fgures - where the mnmal expected net loss wth cost functons of type seems to be constant and the result for cost functons of type and 3 s ncreasng slghtly for small c,,, = before t seems to be constant. Fgures 3-5 show the optmal expected tmes spent n each state for each component as the cost functon parameter, c,,, =, ncreases. The optmal expected tmes spent n each state for component,,,, = reman unchanged for all types of cost functons. For every cost functon type we see from Fgure 3, for component, an ncrease n and a decrease n as c ncreases. For component 3 we have the opposte behavour. As the costs of eepng the expected repar tmes of component low ncreases, t s optmal to spend more expected tme reparng ths component. At the same tme, t wll be more mportant to eep the expected repar tmes of component 3 low. We see from Fgure 4 that the optmal s slghtly decreasng wth ncreasng c for both cost functons of type and 3. 3 s also slghtly decreasng for these two cost functons, but the optmal s ncreasng for these cost functons. For cost functons of type we see from Fgure 4 that the optmal s hgh (approxmately 5) for c >, whle s below 5 for all values of c for cost functons of type 3. For component 3 we see that the results for cost functons of type and are lower than the correspondng results for cost functons of type 3 n state =, and above n state =. The optmal, =,, 3 and =,, for ncreasng c are shown n Fgure 5. We see that the results for cost functons of type are constant. Furthermore, for cost functons of type and 3, component 8

21 K. Sutlaberg, B. Natvg Fgure. Mnmum expected net loss as functon of c for cost functons of type,,3. C =. for =,,3 and c =. for =,3, =,, and =, =,. Fgure. Mnmum expected net loss as functon of c for cost functons of type,,3. C =. for =,,3 and c =. for =,3, =,, and =, =,. 83

22 K. Sutlaberg, B. Natvg Fgure. Mnmum expected net loss as functon of c =. for =,3, =,, and =, =,. c for cost functons of type,,3. C =. for all =,,3 and Fgure 3. Optmal, =,,3, =,, as functon of c for cost functons of type,,3. C =. for =,,3 and c =. for =,3, =,, and =, =,. 84

23 K. Sutlaberg, B. Natvg Fgure 4. Optmal, =,,3, =,, as functon of c for cost functons of type,,3. C =. for =,,3 and c =. for =, 3, =,, and =, =,. Fgure 5. Optmal, =,, 3, =,, as functon of c for cost functons of type,, 3. C =. for all =,,3 and c =. for =,3, =,, and =, =,. 85

24 K. Sutlaberg, B. Natvg has ncreasng and decreasng wth ncreasng c, whereas, component 3 has decreasng 3 and ncreasng 3. Ths s the same behavour as observed for ncreasng 6. Concludng Remars c for these cost functons. In the present paper we have been mnmzng the expected total net loss over a tme perod [,T ] as a functon of the expected component tmes n each state for a three-component flow networ system. Frst the basc model was presented. The assumpton of equalty between the set of system states and the set of component states mpled that an approprate flow networ system would have at least one component n seres wth the rest of the system. Hence, the three-component system gven n Fgure was chosen as a case. Wth three possble system and component states and three components, the orgnal box-constraned optmzaton problem had 9 varables. Due to the complexty of ths problem, we frst studed the smplfed problem, where = and c = c for =,, 3 and =,,, and then a modfcaton of the orgnal problem where the optmzaton was done n two steps (see Secton 5). Ths method found an approxmate soluton. The ndcaton of lac of constructve conclusons s manly due to that we are facng complex dependences. The varables TC, and c for =, n the smplfed problem, and the varables TC, and c for =, and =,, n the modfed full problem, were vared one at a tme wth three dfferent types of cost functons. For the smplfed problem we were able to fnd expressons for the optmal, =,, 3 for cost functons of type. For cost functons of type and 3, the objectve functon, (8), turned out to be a convex functon. Wth cost functons of type the objectve functon s nether convex nor concave. The type cost functons are logarthmc, and hence concave whle the other two types are convex. For ths cost functon, we saw n Fgure 7 that the optmal has a mnmum as c ncreased, whch seems unnatural. In both the smplfed problem and the modfed full model, the mnmum expected net loss was ncreasng wth ncreasng C, =, for every cost functon type (as seen n Fgure 4, Fgure and Fgure respectvely). As the operatonal tme T ncreased we saw a decrease n the mnmum expected net loss n the modfed full model for all three cost functons (as seen n Fgure 8). Ths s n contrast to the results wth the smplfed model when the exponental cost functons were used. Then, the mnmum expected net loss ncreased at frst, before t started to decrease (as seen n Fgure 3). For every cost functon parameter, c, =,,, we vared, we saw n Fgures 7-9 that the optmal, =, was constant for cost functons of type. The same observaton of constant, =,, 3 for cost functons of type was done n Fgures 3-5 where c,,, = were vared. The values were also close to zero. Hence, t was, for cost functons of type, optmal to spend as lttle tme as possble n state ndependent of the values of the parameters. Wth cost functons of type we observed the same, except from when c was ncreasng where decreased from around 5 to close to as c ncreased from to. For c > stayed constant and close to. Thus, t seems le the functonng component state s n a way redundant for cost functons of type and. Ths was not the case wth cost functons of type 3. The general objectve functon (5) can qute easly be modfed to represent the expected net loss of other networ flow systems, and to nclude other types of cost functons. However, wth larger systems, wth more components and possbly more component states, the optmzaton problem qucly becomes large. Hence, the real challenge les n the optmzaton of real lfe systems. Acnowledgements The authors than Professor Ger Dahl for the dea on how to modfy the full optmzaton problem and Ph.D Olav Sutlaberg for valuable feedbac throughout the process. References [] Zo, E. (9) Relablty Engneerng: Old Problems and New Challenges. Relablty Engneerng and System Safety, 86

25 K. Sutlaberg, B. Natvg 94, [] Natvg, B. () Multstate Systems Relablty Theory wth Applcatons. Wley, New Yor. [3] Brnbaum, Z.W. (969) On the Importance of Dfferent Components n a Multcomponent System. In: Krshnaah, P.R., Ed., Multvarate Analyss II, Academc Press, Waltham, [4] Barlow, R.E. and Proschan. F. (975) Importance of System Components and Fault Tree Events. Stochastc Processes and ther Applcatons, 3, [5] Natvg, B. (979) A Suggeston of a New Measure of Importance of System Components. Stochastc Processes and ther Applcatons, 9, [6] Gao, X., Barabady, J. and Mareset, T. () Crtcalty Analyss of a Producton Faclty Usng Cost Importance Measures. Internatonal Journal of Systems Assurance Engneerng and Management,, [7] Wu, S. and Coolen, F. (3) A Cost-Based Importance Measure for System Components: An Extenson of the Brnbaum Importance. European Journal of Operatonal Research, 5, [8] Xu, M., Zhao, W. and Yang, X. () Cost-Related Importance Measure. Proceedngs IEEE Internatonal Conference on Informaton and Automaton (ICIA), Shenzhen, 6-8 June, [9] Gupta, S., Bhattacharya, J., Barabady, J. and Kumar, U. (3) Cost-Effectve Importance Measure: A New Approach for Resource Prortzaton n a Producton Plant. Internatonal Journal of Qualty and Relablty Management, 3, [] Nourelfath, M., Châtelet, E. and Nahas, N. () Jont Redundancy and Imperfect Preventve Mantenance Optmzaton For Seres Parallel Mult-State Degraded Systems. Relablty Engneerng and System Safety, 3, [] Gomes, W.J.S., Bec, A.T. and Hauaas, T. (3) Optmal Inspecton Plannng for Onshore Ppelnes Subject to External Corroson. Relablty Engneerng and System Safety, 8, [] Maras, K.B. (3) Value Maxmzng Mantenance Polces Under General Repar. Relablty Engneerng and System Safety, 9, [3] Doostparast, M., Kolahan, F. and Doostparast, M. (4) A Relablty-Based Approach to Optmze Preventve Mantenance Schedulng for Coherent Systems. Relablty Engneerng and System Safety, 6, [4] Mendes, A.A., Cot, D.W. and Rbero, J.L.D. (4) Establshment of the Optmal Tme Interval between Perodc Inspectons for Redundant Systems. Establshment of the Optmal Tme Interval between Perodc Inspectons for Redundant Systems, 3, [5] Ford, L.R. and Fulerson, D.R. (956) Maxmal Flow through a Networ. Canadan Journal of Mathematcs, 8, [6] Magnus, J.R. and Neudecer, N. (999) Matrx Dfferental Calculus wth Applcatons n Statstcs and Economcs. Wley, New Yor. [7] Ye, Y. (987) Interor Algorthms for Lnear, Quadratc, and Lnearly Constraned Non-Lnear Programmng. PhD Thess, Department of ESS, Stanford Unversty, Stanford. [8] Ghalanos, A. and Theussl, S. () Rsolnp: General Non-lnear Optmzaton Usng Augmented Lagrange Multpler Method. R Pacage Verson.. 87

Support Vector Machines

Support Vector Machines /9/207 MIST.6060 Busness Intellgence and Data Mnng What are Support Vector Machnes? Support Vector Machnes Support Vector Machnes (SVMs) are supervsed learnng technques that analyze data and recognze patterns.

More information

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS ARPN Journal of Engneerng and Appled Scences 006-017 Asan Research Publshng Network (ARPN). All rghts reserved. NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS Igor Grgoryev, Svetlana

More information

GSLM Operations Research II Fall 13/14

GSLM Operations Research II Fall 13/14 GSLM 58 Operatons Research II Fall /4 6. Separable Programmng Consder a general NLP mn f(x) s.t. g j (x) b j j =. m. Defnton 6.. The NLP s a separable program f ts objectve functon and all constrants are

More information

Load Balancing for Hex-Cell Interconnection Network

Load Balancing for Hex-Cell Interconnection Network Int. J. Communcatons, Network and System Scences,,, - Publshed Onlne Aprl n ScRes. http://www.scrp.org/journal/jcns http://dx.do.org/./jcns.. Load Balancng for Hex-Cell Interconnecton Network Saher Manaseer,

More information

A mathematical programming approach to the analysis, design and scheduling of offshore oilfields

A mathematical programming approach to the analysis, design and scheduling of offshore oilfields 17 th European Symposum on Computer Aded Process Engneerng ESCAPE17 V. Plesu and P.S. Agach (Edtors) 2007 Elsever B.V. All rghts reserved. 1 A mathematcal programmng approach to the analyss, desgn and

More information

An Optimal Algorithm for Prufer Codes *

An Optimal Algorithm for Prufer Codes * J. Software Engneerng & Applcatons, 2009, 2: 111-115 do:10.4236/jsea.2009.22016 Publshed Onlne July 2009 (www.scrp.org/journal/jsea) An Optmal Algorthm for Prufer Codes * Xaodong Wang 1, 2, Le Wang 3,

More information

Support Vector Machines

Support Vector Machines Support Vector Machnes Decson surface s a hyperplane (lne n 2D) n feature space (smlar to the Perceptron) Arguably, the most mportant recent dscovery n machne learnng In a nutshell: map the data to a predetermned

More information

6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour

6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour 6.854 Advanced Algorthms Petar Maymounkov Problem Set 11 (November 23, 2005) Wth: Benjamn Rossman, Oren Wemann, and Pouya Kheradpour Problem 1. We reduce vertex cover to MAX-SAT wth weghts, such that the

More information

Parallelism for Nested Loops with Non-uniform and Flow Dependences

Parallelism for Nested Loops with Non-uniform and Flow Dependences Parallelsm for Nested Loops wth Non-unform and Flow Dependences Sam-Jn Jeong Dept. of Informaton & Communcaton Engneerng, Cheonan Unversty, 5, Anseo-dong, Cheonan, Chungnam, 330-80, Korea. seong@cheonan.ac.kr

More information

Feature Reduction and Selection

Feature Reduction and Selection Feature Reducton and Selecton Dr. Shuang LIANG School of Software Engneerng TongJ Unversty Fall, 2012 Today s Topcs Introducton Problems of Dmensonalty Feature Reducton Statstc methods Prncpal Components

More information

Solving two-person zero-sum game by Matlab

Solving two-person zero-sum game by Matlab Appled Mechancs and Materals Onlne: 2011-02-02 ISSN: 1662-7482, Vols. 50-51, pp 262-265 do:10.4028/www.scentfc.net/amm.50-51.262 2011 Trans Tech Publcatons, Swtzerland Solvng two-person zero-sum game by

More information

S1 Note. Basis functions.

S1 Note. Basis functions. S1 Note. Bass functons. Contents Types of bass functons...1 The Fourer bass...2 B-splne bass...3 Power and type I error rates wth dfferent numbers of bass functons...4 Table S1. Smulaton results of type

More information

Sum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints

Sum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints Australan Journal of Basc and Appled Scences, 2(4): 1204-1208, 2008 ISSN 1991-8178 Sum of Lnear and Fractonal Multobjectve Programmng Problem under Fuzzy Rules Constrants 1 2 Sanjay Jan and Kalash Lachhwan

More information

Classification / Regression Support Vector Machines

Classification / Regression Support Vector Machines Classfcaton / Regresson Support Vector Machnes Jeff Howbert Introducton to Machne Learnng Wnter 04 Topcs SVM classfers for lnearly separable classes SVM classfers for non-lnearly separable classes SVM

More information

A New Approach For the Ranking of Fuzzy Sets With Different Heights

A New Approach For the Ranking of Fuzzy Sets With Different Heights New pproach For the ankng of Fuzzy Sets Wth Dfferent Heghts Pushpnder Sngh School of Mathematcs Computer pplcatons Thapar Unversty, Patala-7 00 Inda pushpndersnl@gmalcom STCT ankng of fuzzy sets plays

More information

Hermite Splines in Lie Groups as Products of Geodesics

Hermite Splines in Lie Groups as Products of Geodesics Hermte Splnes n Le Groups as Products of Geodescs Ethan Eade Updated May 28, 2017 1 Introducton 1.1 Goal Ths document defnes a curve n the Le group G parametrzed by tme and by structural parameters n the

More information

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique //00 :0 AM Outlne and Readng The Greedy Method The Greedy Method Technque (secton.) Fractonal Knapsack Problem (secton..) Task Schedulng (secton..) Mnmum Spannng Trees (secton.) Change Money Problem Greedy

More information

Quality Improvement Algorithm for Tetrahedral Mesh Based on Optimal Delaunay Triangulation

Quality Improvement Algorithm for Tetrahedral Mesh Based on Optimal Delaunay Triangulation Intellgent Informaton Management, 013, 5, 191-195 Publshed Onlne November 013 (http://www.scrp.org/journal/m) http://dx.do.org/10.36/m.013.5601 Qualty Improvement Algorthm for Tetrahedral Mesh Based on

More information

Learning the Kernel Parameters in Kernel Minimum Distance Classifier

Learning the Kernel Parameters in Kernel Minimum Distance Classifier Learnng the Kernel Parameters n Kernel Mnmum Dstance Classfer Daoqang Zhang 1,, Songcan Chen and Zh-Hua Zhou 1* 1 Natonal Laboratory for Novel Software Technology Nanjng Unversty, Nanjng 193, Chna Department

More information

For instance, ; the five basic number-sets are increasingly more n A B & B A A = B (1)

For instance, ; the five basic number-sets are increasingly more n A B & B A A = B (1) Secton 1.2 Subsets and the Boolean operatons on sets If every element of the set A s an element of the set B, we say that A s a subset of B, or that A s contaned n B, or that B contans A, and we wrte A

More information

Cluster Analysis of Electrical Behavior

Cluster Analysis of Electrical Behavior Journal of Computer and Communcatons, 205, 3, 88-93 Publshed Onlne May 205 n ScRes. http://www.scrp.org/ournal/cc http://dx.do.org/0.4236/cc.205.350 Cluster Analyss of Electrcal Behavor Ln Lu Ln Lu, School

More information

Course Introduction. Algorithm 8/31/2017. COSC 320 Advanced Data Structures and Algorithms. COSC 320 Advanced Data Structures and Algorithms

Course Introduction. Algorithm 8/31/2017. COSC 320 Advanced Data Structures and Algorithms. COSC 320 Advanced Data Structures and Algorithms Course Introducton Course Topcs Exams, abs, Proects A quc loo at a few algorthms 1 Advanced Data Structures and Algorthms Descrpton: We are gong to dscuss algorthm complexty analyss, algorthm desgn technques

More information

Lecture 5: Multilayer Perceptrons

Lecture 5: Multilayer Perceptrons Lecture 5: Multlayer Perceptrons Roger Grosse 1 Introducton So far, we ve only talked about lnear models: lnear regresson and lnear bnary classfers. We noted that there are functons that can t be represented

More information

An Iterative Solution Approach to Process Plant Layout using Mixed Integer Optimisation

An Iterative Solution Approach to Process Plant Layout using Mixed Integer Optimisation 17 th European Symposum on Computer Aded Process Engneerng ESCAPE17 V. Plesu and P.S. Agach (Edtors) 2007 Elsever B.V. All rghts reserved. 1 An Iteratve Soluton Approach to Process Plant Layout usng Mxed

More information

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz Compler Desgn Sprng 2014 Regster Allocaton Sample Exercses and Solutons Prof. Pedro C. Dnz USC / Informaton Scences Insttute 4676 Admralty Way, Sute 1001 Marna del Rey, Calforna 90292 pedro@s.edu Regster

More information

SENSITIVITY ANALYSIS IN LINEAR PROGRAMMING USING A CALCULATOR

SENSITIVITY ANALYSIS IN LINEAR PROGRAMMING USING A CALCULATOR SENSITIVITY ANALYSIS IN LINEAR PROGRAMMING USING A CALCULATOR Judth Aronow Rchard Jarvnen Independent Consultant Dept of Math/Stat 559 Frost Wnona State Unversty Beaumont, TX 7776 Wnona, MN 55987 aronowju@hal.lamar.edu

More information

A Binarization Algorithm specialized on Document Images and Photos

A Binarization Algorithm specialized on Document Images and Photos A Bnarzaton Algorthm specalzed on Document mages and Photos Ergna Kavalleratou Dept. of nformaton and Communcaton Systems Engneerng Unversty of the Aegean kavalleratou@aegean.gr Abstract n ths paper, a

More information

LECTURE : MANIFOLD LEARNING

LECTURE : MANIFOLD LEARNING LECTURE : MANIFOLD LEARNING Rta Osadchy Some sldes are due to L.Saul, V. C. Raykar, N. Verma Topcs PCA MDS IsoMap LLE EgenMaps Done! Dmensonalty Reducton Data representaton Inputs are real-valued vectors

More information

Determining the Optimal Bandwidth Based on Multi-criterion Fusion

Determining the Optimal Bandwidth Based on Multi-criterion Fusion Proceedngs of 01 4th Internatonal Conference on Machne Learnng and Computng IPCSIT vol. 5 (01) (01) IACSIT Press, Sngapore Determnng the Optmal Bandwdth Based on Mult-crteron Fuson Ha-L Lang 1+, Xan-Mn

More information

X- Chart Using ANOM Approach

X- Chart Using ANOM Approach ISSN 1684-8403 Journal of Statstcs Volume 17, 010, pp. 3-3 Abstract X- Chart Usng ANOM Approach Gullapall Chakravarth 1 and Chaluvad Venkateswara Rao Control lmts for ndvdual measurements (X) chart are

More information

CMPS 10 Introduction to Computer Science Lecture Notes

CMPS 10 Introduction to Computer Science Lecture Notes CPS 0 Introducton to Computer Scence Lecture Notes Chapter : Algorthm Desgn How should we present algorthms? Natural languages lke Englsh, Spansh, or French whch are rch n nterpretaton and meanng are not

More information

An Application of the Dulmage-Mendelsohn Decomposition to Sparse Null Space Bases of Full Row Rank Matrices

An Application of the Dulmage-Mendelsohn Decomposition to Sparse Null Space Bases of Full Row Rank Matrices Internatonal Mathematcal Forum, Vol 7, 2012, no 52, 2549-2554 An Applcaton of the Dulmage-Mendelsohn Decomposton to Sparse Null Space Bases of Full Row Rank Matrces Mostafa Khorramzadeh Department of Mathematcal

More information

Lecture 4: Principal components

Lecture 4: Principal components /3/6 Lecture 4: Prncpal components 3..6 Multvarate lnear regresson MLR s optmal for the estmaton data...but poor for handlng collnear data Covarance matrx s not nvertble (large condton number) Robustness

More information

Non-Split Restrained Dominating Set of an Interval Graph Using an Algorithm

Non-Split Restrained Dominating Set of an Interval Graph Using an Algorithm Internatonal Journal of Advancements n Research & Technology, Volume, Issue, July- ISS - on-splt Restraned Domnatng Set of an Interval Graph Usng an Algorthm ABSTRACT Dr.A.Sudhakaraah *, E. Gnana Deepka,

More information

Mathematics 256 a course in differential equations for engineering students

Mathematics 256 a course in differential equations for engineering students Mathematcs 56 a course n dfferental equatons for engneerng students Chapter 5. More effcent methods of numercal soluton Euler s method s qute neffcent. Because the error s essentally proportonal to the

More information

Network Coding as a Dynamical System

Network Coding as a Dynamical System Network Codng as a Dynamcal System Narayan B. Mandayam IEEE Dstngushed Lecture (jont work wth Dan Zhang and a Su) Department of Electrcal and Computer Engneerng Rutgers Unversty Outlne. Introducton 2.

More information

CHAPTER 3 SEQUENTIAL MINIMAL OPTIMIZATION TRAINED SUPPORT VECTOR CLASSIFIER FOR CANCER PREDICTION

CHAPTER 3 SEQUENTIAL MINIMAL OPTIMIZATION TRAINED SUPPORT VECTOR CLASSIFIER FOR CANCER PREDICTION 48 CHAPTER 3 SEQUENTIAL MINIMAL OPTIMIZATION TRAINED SUPPORT VECTOR CLASSIFIER FOR CANCER PREDICTION 3.1 INTRODUCTION The raw mcroarray data s bascally an mage wth dfferent colors ndcatng hybrdzaton (Xue

More information

Positive Semi-definite Programming Localization in Wireless Sensor Networks

Positive Semi-definite Programming Localization in Wireless Sensor Networks Postve Sem-defnte Programmng Localzaton n Wreless Sensor etworks Shengdong Xe 1,, Jn Wang, Aqun Hu 1, Yunl Gu, Jang Xu, 1 School of Informaton Scence and Engneerng, Southeast Unversty, 10096, anjng Computer

More information

The Shortest Path of Touring Lines given in the Plane

The Shortest Path of Touring Lines given in the Plane Send Orders for Reprnts to reprnts@benthamscence.ae 262 The Open Cybernetcs & Systemcs Journal, 2015, 9, 262-267 The Shortest Path of Tourng Lnes gven n the Plane Open Access Ljuan Wang 1,2, Dandan He

More information

Wishing you all a Total Quality New Year!

Wishing you all a Total Quality New Year! Total Qualty Management and Sx Sgma Post Graduate Program 214-15 Sesson 4 Vnay Kumar Kalakband Assstant Professor Operatons & Systems Area 1 Wshng you all a Total Qualty New Year! Hope you acheve Sx sgma

More information

Programming in Fortran 90 : 2017/2018

Programming in Fortran 90 : 2017/2018 Programmng n Fortran 90 : 2017/2018 Programmng n Fortran 90 : 2017/2018 Exercse 1 : Evaluaton of functon dependng on nput Wrte a program who evaluate the functon f (x,y) for any two user specfed values

More information

Fast Computation of Shortest Path for Visiting Segments in the Plane

Fast Computation of Shortest Path for Visiting Segments in the Plane Send Orders for Reprnts to reprnts@benthamscence.ae 4 The Open Cybernetcs & Systemcs Journal, 04, 8, 4-9 Open Access Fast Computaton of Shortest Path for Vstng Segments n the Plane Ljuan Wang,, Bo Jang

More information

Clustering on antimatroids and convex geometries

Clustering on antimatroids and convex geometries Clusterng on antmatrods and convex geometres YULIA KEMPNER 1, ILYA MUCNIK 2 1 Department of Computer cence olon Academc Insttute of Technology 52 Golomb tr., P.O. Box 305, olon 58102 IRAEL 2 Department

More information

U.C. Berkeley CS294: Beyond Worst-Case Analysis Handout 5 Luca Trevisan September 7, 2017

U.C. Berkeley CS294: Beyond Worst-Case Analysis Handout 5 Luca Trevisan September 7, 2017 U.C. Bereley CS294: Beyond Worst-Case Analyss Handout 5 Luca Trevsan September 7, 207 Scrbed by Haars Khan Last modfed 0/3/207 Lecture 5 In whch we study the SDP relaxaton of Max Cut n random graphs. Quc

More information

Analysis of Continuous Beams in General

Analysis of Continuous Beams in General Analyss of Contnuous Beams n General Contnuous beams consdered here are prsmatc, rgdly connected to each beam segment and supported at varous ponts along the beam. onts are selected at ponts of support,

More information

CS 534: Computer Vision Model Fitting

CS 534: Computer Vision Model Fitting CS 534: Computer Vson Model Fttng Sprng 004 Ahmed Elgammal Dept of Computer Scence CS 534 Model Fttng - 1 Outlnes Model fttng s mportant Least-squares fttng Maxmum lkelhood estmaton MAP estmaton Robust

More information

Empirical Distributions of Parameter Estimates. in Binary Logistic Regression Using Bootstrap

Empirical Distributions of Parameter Estimates. in Binary Logistic Regression Using Bootstrap Int. Journal of Math. Analyss, Vol. 8, 4, no. 5, 7-7 HIKARI Ltd, www.m-hkar.com http://dx.do.org/.988/jma.4.494 Emprcal Dstrbutons of Parameter Estmates n Bnary Logstc Regresson Usng Bootstrap Anwar Ftranto*

More information

Intra-Parametric Analysis of a Fuzzy MOLP

Intra-Parametric Analysis of a Fuzzy MOLP Intra-Parametrc Analyss of a Fuzzy MOLP a MIAO-LING WANG a Department of Industral Engneerng and Management a Mnghsn Insttute of Technology and Hsnchu Tawan, ROC b HSIAO-FAN WANG b Insttute of Industral

More information

Backpropagation: In Search of Performance Parameters

Backpropagation: In Search of Performance Parameters Bacpropagaton: In Search of Performance Parameters ANIL KUMAR ENUMULAPALLY, LINGGUO BU, and KHOSROW KAIKHAH, Ph.D. Computer Scence Department Texas State Unversty-San Marcos San Marcos, TX-78666 USA ae049@txstate.edu,

More information

Categories and Subject Descriptors B.7.2 [Integrated Circuits]: Design Aids Verification. General Terms Algorithms

Categories and Subject Descriptors B.7.2 [Integrated Circuits]: Design Aids Verification. General Terms Algorithms 3. Fndng Determnstc Soluton from Underdetermned Equaton: Large-Scale Performance Modelng by Least Angle Regresson Xn L ECE Department, Carnege Mellon Unversty Forbs Avenue, Pttsburgh, PA 3 xnl@ece.cmu.edu

More information

CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION

CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION 24 CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION The present chapter proposes an IPSO approach for multprocessor task schedulng problem wth two classfcatons, namely, statc ndependent tasks and

More information

Reducing Frame Rate for Object Tracking

Reducing Frame Rate for Object Tracking Reducng Frame Rate for Object Trackng Pavel Korshunov 1 and We Tsang Oo 2 1 Natonal Unversty of Sngapore, Sngapore 11977, pavelkor@comp.nus.edu.sg 2 Natonal Unversty of Sngapore, Sngapore 11977, oowt@comp.nus.edu.sg

More information

A Simple and Efficient Goal Programming Model for Computing of Fuzzy Linear Regression Parameters with Considering Outliers

A Simple and Efficient Goal Programming Model for Computing of Fuzzy Linear Regression Parameters with Considering Outliers 62626262621 Journal of Uncertan Systems Vol.5, No.1, pp.62-71, 211 Onlne at: www.us.org.u A Smple and Effcent Goal Programmng Model for Computng of Fuzzy Lnear Regresson Parameters wth Consderng Outlers

More information

A Five-Point Subdivision Scheme with Two Parameters and a Four-Point Shape-Preserving Scheme

A Five-Point Subdivision Scheme with Two Parameters and a Four-Point Shape-Preserving Scheme Mathematcal and Computatonal Applcatons Artcle A Fve-Pont Subdvson Scheme wth Two Parameters and a Four-Pont Shape-Preservng Scheme Jeqng Tan,2, Bo Wang, * and Jun Sh School of Mathematcs, Hefe Unversty

More information

Parallel matrix-vector multiplication

Parallel matrix-vector multiplication Appendx A Parallel matrx-vector multplcaton The reduced transton matrx of the three-dmensonal cage model for gel electrophoress, descrbed n secton 3.2, becomes excessvely large for polymer lengths more

More information

Virtual Machine Migration based on Trust Measurement of Computer Node

Virtual Machine Migration based on Trust Measurement of Computer Node Appled Mechancs and Materals Onlne: 2014-04-04 ISSN: 1662-7482, Vols. 536-537, pp 678-682 do:10.4028/www.scentfc.net/amm.536-537.678 2014 Trans Tech Publcatons, Swtzerland Vrtual Machne Mgraton based on

More information

Biostatistics 615/815

Biostatistics 615/815 The E-M Algorthm Bostatstcs 615/815 Lecture 17 Last Lecture: The Smplex Method General method for optmzaton Makes few assumptons about functon Crawls towards mnmum Some recommendatons Multple startng ponts

More information

Helsinki University Of Technology, Systems Analysis Laboratory Mat Independent research projects in applied mathematics (3 cr)

Helsinki University Of Technology, Systems Analysis Laboratory Mat Independent research projects in applied mathematics (3 cr) Helsnk Unversty Of Technology, Systems Analyss Laboratory Mat-2.08 Independent research projects n appled mathematcs (3 cr) "! #$&% Antt Laukkanen 506 R ajlaukka@cc.hut.f 2 Introducton...3 2 Multattrbute

More information

Complex Numbers. Now we also saw that if a and b were both positive then ab = a b. For a second let s forget that restriction and do the following.

Complex Numbers. Now we also saw that if a and b were both positive then ab = a b. For a second let s forget that restriction and do the following. Complex Numbers The last topc n ths secton s not really related to most of what we ve done n ths chapter, although t s somewhat related to the radcals secton as we wll see. We also won t need the materal

More information

Load-Balanced Anycast Routing

Load-Balanced Anycast Routing Load-Balanced Anycast Routng Chng-Yu Ln, Jung-Hua Lo, and Sy-Yen Kuo Department of Electrcal Engneerng atonal Tawan Unversty, Tape, Tawan sykuo@cc.ee.ntu.edu.tw Abstract For fault-tolerance and load-balance

More information

Type-2 Fuzzy Non-uniform Rational B-spline Model with Type-2 Fuzzy Data

Type-2 Fuzzy Non-uniform Rational B-spline Model with Type-2 Fuzzy Data Malaysan Journal of Mathematcal Scences 11(S) Aprl : 35 46 (2017) Specal Issue: The 2nd Internatonal Conference and Workshop on Mathematcal Analyss (ICWOMA 2016) MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES

More information

AADL : about scheduling analysis

AADL : about scheduling analysis AADL : about schedulng analyss Schedulng analyss, what s t? Embedded real-tme crtcal systems have temporal constrants to meet (e.g. deadlne). Many systems are bult wth operatng systems provdng multtaskng

More information

APPLICATION OF MULTIVARIATE LOSS FUNCTION FOR ASSESSMENT OF THE QUALITY OF TECHNOLOGICAL PROCESS MANAGEMENT

APPLICATION OF MULTIVARIATE LOSS FUNCTION FOR ASSESSMENT OF THE QUALITY OF TECHNOLOGICAL PROCESS MANAGEMENT 3. - 5. 5., Brno, Czech Republc, EU APPLICATION OF MULTIVARIATE LOSS FUNCTION FOR ASSESSMENT OF THE QUALITY OF TECHNOLOGICAL PROCESS MANAGEMENT Abstract Josef TOŠENOVSKÝ ) Lenka MONSPORTOVÁ ) Flp TOŠENOVSKÝ

More information

Kent State University CS 4/ Design and Analysis of Algorithms. Dept. of Math & Computer Science LECT-16. Dynamic Programming

Kent State University CS 4/ Design and Analysis of Algorithms. Dept. of Math & Computer Science LECT-16. Dynamic Programming CS 4/560 Desgn and Analyss of Algorthms Kent State Unversty Dept. of Math & Computer Scence LECT-6 Dynamc Programmng 2 Dynamc Programmng Dynamc Programmng, lke the dvde-and-conquer method, solves problems

More information

Machine Learning. Support Vector Machines. (contains material adapted from talks by Constantin F. Aliferis & Ioannis Tsamardinos, and Martin Law)

Machine Learning. Support Vector Machines. (contains material adapted from talks by Constantin F. Aliferis & Ioannis Tsamardinos, and Martin Law) Machne Learnng Support Vector Machnes (contans materal adapted from talks by Constantn F. Alfers & Ioanns Tsamardnos, and Martn Law) Bryan Pardo, Machne Learnng: EECS 349 Fall 2014 Support Vector Machnes

More information

Smoothing Spline ANOVA for variable screening

Smoothing Spline ANOVA for variable screening Smoothng Splne ANOVA for varable screenng a useful tool for metamodels tranng and mult-objectve optmzaton L. Rcco, E. Rgon, A. Turco Outlne RSM Introducton Possble couplng Test case MOO MOO wth Game Theory

More information

5 The Primal-Dual Method

5 The Primal-Dual Method 5 The Prmal-Dual Method Orgnally desgned as a method for solvng lnear programs, where t reduces weghted optmzaton problems to smpler combnatoral ones, the prmal-dual method (PDM) has receved much attenton

More information

A Saturation Binary Neural Network for Crossbar Switching Problem

A Saturation Binary Neural Network for Crossbar Switching Problem A Saturaton Bnary Neural Network for Crossbar Swtchng Problem Cu Zhang 1, L-Qng Zhao 2, and Rong-Long Wang 2 1 Department of Autocontrol, Laonng Insttute of Scence and Technology, Benx, Chna bxlkyzhangcu@163.com

More information

A fault tree analysis strategy using binary decision diagrams

A fault tree analysis strategy using binary decision diagrams Loughborough Unversty Insttutonal Repostory A fault tree analyss strategy usng bnary decson dagrams Ths tem was submtted to Loughborough Unversty's Insttutonal Repostory by the/an author. Addtonal Informaton:

More information

Analysis of Non-coherent Fault Trees Using Ternary Decision Diagrams

Analysis of Non-coherent Fault Trees Using Ternary Decision Diagrams Analyss of Non-coherent Fault Trees Usng Ternary Decson Dagrams Rasa Remenyte-Prescott Dep. of Aeronautcal and Automotve Engneerng Loughborough Unversty, Loughborough, LE11 3TU, England R.Remenyte-Prescott@lboro.ac.uk

More information

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization Problem efntons and Evaluaton Crtera for Computatonal Expensve Optmzaton B. Lu 1, Q. Chen and Q. Zhang 3, J. J. Lang 4, P. N. Suganthan, B. Y. Qu 6 1 epartment of Computng, Glyndwr Unversty, UK Faclty

More information

Module Management Tool in Software Development Organizations

Module Management Tool in Software Development Organizations Journal of Computer Scence (5): 8-, 7 ISSN 59-66 7 Scence Publcatons Management Tool n Software Development Organzatons Ahmad A. Al-Rababah and Mohammad A. Al-Rababah Faculty of IT, Al-Ahlyyah Amman Unversty,

More information

ON SOME ENTERTAINING APPLICATIONS OF THE CONCEPT OF SET IN COMPUTER SCIENCE COURSE

ON SOME ENTERTAINING APPLICATIONS OF THE CONCEPT OF SET IN COMPUTER SCIENCE COURSE Yordzhev K., Kostadnova H. Інформаційні технології в освіті ON SOME ENTERTAINING APPLICATIONS OF THE CONCEPT OF SET IN COMPUTER SCIENCE COURSE Yordzhev K., Kostadnova H. Some aspects of programmng educaton

More information

Design of Structure Optimization with APDL

Design of Structure Optimization with APDL Desgn of Structure Optmzaton wth APDL Yanyun School of Cvl Engneerng and Archtecture, East Chna Jaotong Unversty Nanchang 330013 Chna Abstract In ths paper, the desgn process of structure optmzaton wth

More information

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance Tsnghua Unversty at TAC 2009: Summarzng Mult-documents by Informaton Dstance Chong Long, Mnle Huang, Xaoyan Zhu State Key Laboratory of Intellgent Technology and Systems, Tsnghua Natonal Laboratory for

More information

TPL-Aware Displacement-driven Detailed Placement Refinement with Coloring Constraints

TPL-Aware Displacement-driven Detailed Placement Refinement with Coloring Constraints TPL-ware Dsplacement-drven Detaled Placement Refnement wth Colorng Constrants Tao Ln Iowa State Unversty tln@astate.edu Chrs Chu Iowa State Unversty cnchu@astate.edu BSTRCT To mnmze the effect of process

More information

R s s f. m y s. SPH3UW Unit 7.3 Spherical Concave Mirrors Page 1 of 12. Notes

R s s f. m y s. SPH3UW Unit 7.3 Spherical Concave Mirrors Page 1 of 12. Notes SPH3UW Unt 7.3 Sphercal Concave Mrrors Page 1 of 1 Notes Physcs Tool box Concave Mrror If the reflectng surface takes place on the nner surface of the sphercal shape so that the centre of the mrror bulges

More information

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision SLAM Summer School 2006 Practcal 2: SLAM usng Monocular Vson Javer Cvera, Unversty of Zaragoza Andrew J. Davson, Imperal College London J.M.M Montel, Unversty of Zaragoza. josemar@unzar.es, jcvera@unzar.es,

More information

On Some Entertaining Applications of the Concept of Set in Computer Science Course

On Some Entertaining Applications of the Concept of Set in Computer Science Course On Some Entertanng Applcatons of the Concept of Set n Computer Scence Course Krasmr Yordzhev *, Hrstna Kostadnova ** * Assocate Professor Krasmr Yordzhev, Ph.D., Faculty of Mathematcs and Natural Scences,

More information

A HEURISTIC METHOD FOR RELIABILITY REDUNDANCY OPTIMIZATION OF FLOW NETWORKS

A HEURISTIC METHOD FOR RELIABILITY REDUNDANCY OPTIMIZATION OF FLOW NETWORKS Kumar Pardeep and Chaturved D.K., Pahua G.L. - A HEURISTIC METHOD FOR RELIABILITY REDUNDANCY OPTIMIZATION OF FLOW NETWORKS A HEURISTIC METHOD FOR RELIABILITY REDUNDANCY OPTIMIZATION OF FLOW NETWORKS Kumar

More information

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Proceedngs of the Wnter Smulaton Conference M E Kuhl, N M Steger, F B Armstrong, and J A Jones, eds A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Mark W Brantley Chun-Hung

More information

Life Tables (Times) Summary. Sample StatFolio: lifetable times.sgp

Life Tables (Times) Summary. Sample StatFolio: lifetable times.sgp Lfe Tables (Tmes) Summary... 1 Data Input... 2 Analyss Summary... 3 Survval Functon... 5 Log Survval Functon... 6 Cumulatve Hazard Functon... 7 Percentles... 7 Group Comparsons... 8 Summary The Lfe Tables

More information

Machine Learning: Algorithms and Applications

Machine Learning: Algorithms and Applications 14/05/1 Machne Learnng: Algorthms and Applcatons Florano Zn Free Unversty of Bozen-Bolzano Faculty of Computer Scence Academc Year 011-01 Lecture 10: 14 May 01 Unsupervsed Learnng cont Sldes courtesy of

More information

Solitary and Traveling Wave Solutions to a Model. of Long Range Diffusion Involving Flux with. Stability Analysis

Solitary and Traveling Wave Solutions to a Model. of Long Range Diffusion Involving Flux with. Stability Analysis Internatonal Mathematcal Forum, Vol. 6,, no. 7, 8 Soltary and Travelng Wave Solutons to a Model of Long Range ffuson Involvng Flux wth Stablty Analyss Manar A. Al-Qudah Math epartment, Rabgh Faculty of

More information

2x x l. Module 3: Element Properties Lecture 4: Lagrange and Serendipity Elements

2x x l. Module 3: Element Properties Lecture 4: Lagrange and Serendipity Elements Module 3: Element Propertes Lecture : Lagrange and Serendpty Elements 5 In last lecture note, the nterpolaton functons are derved on the bass of assumed polynomal from Pascal s trangle for the fled varable.

More information

NAG Fortran Library Chapter Introduction. G10 Smoothing in Statistics

NAG Fortran Library Chapter Introduction. G10 Smoothing in Statistics Introducton G10 NAG Fortran Lbrary Chapter Introducton G10 Smoothng n Statstcs Contents 1 Scope of the Chapter... 2 2 Background to the Problems... 2 2.1 Smoothng Methods... 2 2.2 Smoothng Splnes and Regresson

More information

An Application of Network Simplex Method for Minimum Cost Flow Problems

An Application of Network Simplex Method for Minimum Cost Flow Problems BALKANJM 0 (0) -0 Contents lsts avalable at BALKANJM BALKAN JOURNAL OF MATHEMATICS journal homepage: www.balkanjm.com An Applcaton of Network Smplex Method for Mnmum Cost Flow Problems Ergun EROGLU *a

More information

Overview. Basic Setup [9] Motivation and Tasks. Modularization 2008/2/20 IMPROVED COVERAGE CONTROL USING ONLY LOCAL INFORMATION

Overview. Basic Setup [9] Motivation and Tasks. Modularization 2008/2/20 IMPROVED COVERAGE CONTROL USING ONLY LOCAL INFORMATION Overvew 2 IMPROVED COVERAGE CONTROL USING ONLY LOCAL INFORMATION Introducton Mult- Smulator MASIM Theoretcal Work and Smulaton Results Concluson Jay Wagenpfel, Adran Trachte Motvaton and Tasks Basc Setup

More information

A Topology-aware Random Walk

A Topology-aware Random Walk A Topology-aware Random Walk Inkwan Yu, Rchard Newman Dept. of CISE, Unversty of Florda, Ganesvlle, Florda, USA Abstract When a graph can be decomposed nto clusters of well connected subgraphs, t s possble

More information

Proper Choice of Data Used for the Estimation of Datum Transformation Parameters

Proper Choice of Data Used for the Estimation of Datum Transformation Parameters Proper Choce of Data Used for the Estmaton of Datum Transformaton Parameters Hakan S. KUTOGLU, Turkey Key words: Coordnate systems; transformaton; estmaton, relablty. SUMMARY Advances n technologes and

More information

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points;

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points; Subspace clusterng Clusterng Fundamental to all clusterng technques s the choce of dstance measure between data ponts; D q ( ) ( ) 2 x x = x x, j k = 1 k jk Squared Eucldean dstance Assumpton: All features

More information

Some Advanced SPC Tools 1. Cumulative Sum Control (Cusum) Chart For the data shown in Table 9-1, the x chart can be generated.

Some Advanced SPC Tools 1. Cumulative Sum Control (Cusum) Chart For the data shown in Table 9-1, the x chart can be generated. Some Advanced SP Tools 1. umulatve Sum ontrol (usum) hart For the data shown n Table 9-1, the x chart can be generated. However, the shft taken place at sample #21 s not apparent. 92 For ths set samples,

More information

EVALUATION OF THE PERFORMANCES OF ARTIFICIAL BEE COLONY AND INVASIVE WEED OPTIMIZATION ALGORITHMS ON THE MODIFIED BENCHMARK FUNCTIONS

EVALUATION OF THE PERFORMANCES OF ARTIFICIAL BEE COLONY AND INVASIVE WEED OPTIMIZATION ALGORITHMS ON THE MODIFIED BENCHMARK FUNCTIONS Academc Research Internatonal ISS-L: 3-9553, ISS: 3-9944 Vol., o. 3, May 0 EVALUATIO OF THE PERFORMACES OF ARTIFICIAL BEE COLOY AD IVASIVE WEED OPTIMIZATIO ALGORITHMS O THE MODIFIED BECHMARK FUCTIOS Dlay

More information

A NOTE ON FUZZY CLOSURE OF A FUZZY SET

A NOTE ON FUZZY CLOSURE OF A FUZZY SET (JPMNT) Journal of Process Management New Technologes, Internatonal A NOTE ON FUZZY CLOSURE OF A FUZZY SET Bhmraj Basumatary Department of Mathematcal Scences, Bodoland Unversty, Kokrajhar, Assam, Inda,

More information

Routing in Degree-constrained FSO Mesh Networks

Routing in Degree-constrained FSO Mesh Networks Internatonal Journal of Hybrd Informaton Technology Vol., No., Aprl, 009 Routng n Degree-constraned FSO Mesh Networks Zpng Hu, Pramode Verma, and James Sluss Jr. School of Electrcal & Computer Engneerng

More information

A Newton-Type Method for Constrained Least-Squares Data-Fitting with Easy-to-Control Rational Curves

A Newton-Type Method for Constrained Least-Squares Data-Fitting with Easy-to-Control Rational Curves A Newton-Type Method for Constraned Least-Squares Data-Fttng wth Easy-to-Control Ratonal Curves G. Cascola a, L. Roman b, a Department of Mathematcs, Unversty of Bologna, P.zza d Porta San Donato 5, 4017

More information

Report on On-line Graph Coloring

Report on On-line Graph Coloring 2003 Fall Semester Comp 670K Onlne Algorthm Report on LO Yuet Me (00086365) cndylo@ust.hk Abstract Onlne algorthm deals wth data that has no future nformaton. Lots of examples demonstrate that onlne algorthm

More information

Line Clipping by Convex and Nonconvex Polyhedra in E 3

Line Clipping by Convex and Nonconvex Polyhedra in E 3 Lne Clppng by Convex and Nonconvex Polyhedra n E 3 Václav Skala 1 Department of Informatcs and Computer Scence Unversty of West Bohema Unverztní 22, Box 314, 306 14 Plzeò Czech Republc e-mal: skala@kv.zcu.cz

More information

CHAPTER 2 DECOMPOSITION OF GRAPHS

CHAPTER 2 DECOMPOSITION OF GRAPHS CHAPTER DECOMPOSITION OF GRAPHS. INTRODUCTION A graph H s called a Supersubdvson of a graph G f H s obtaned from G by replacng every edge uv of G by a bpartte graph,m (m may vary for each edge by dentfyng

More information

Exercises (Part 4) Introduction to R UCLA/CCPR. John Fox, February 2005

Exercises (Part 4) Introduction to R UCLA/CCPR. John Fox, February 2005 Exercses (Part 4) Introducton to R UCLA/CCPR John Fox, February 2005 1. A challengng problem: Iterated weghted least squares (IWLS) s a standard method of fttng generalzed lnear models to data. As descrbed

More information