A Graphical Approach for Solving Single Machine Scheduling Problems Approximately

Size: px
Start display at page:

Download "A Graphical Approach for Solving Single Machine Scheduling Problems Approximately"

Transcription

1 A Graphica Approach for Soving Singe Machine Scheduing Probems Approximatey Evgeny R Gafarov Aexandre Dogui Aexander A Lazarev Frank Werner Institute of Contro Sciences of the Russian Academy of Sciences, Profsoyuznaya st 65, Moscow, Russia (e-mai: axe73@mairu, azarev@ipuru) Ecoe Nationae Superieure des Mines, FAYOL-EMSE, CNRS:UMR6158, LIMOS, F Saint-Etienne, France (e-mai: dogui@emsefr) Fakutät für Mathematik, Otto-von-Guericke-Universität Magdeburg, PSF 4120, Magdeburg, Germany (e-mai: frankwerner@mathematikuni-magdeburgde) Abstract: For five singe machine tota tardiness probems a fuy poynomia-time approximation scheme (FPTAS) based on a graphica agorithm is presented The FPTAS has the best running time among the known approximation schemes for these probems Keywords: Scheduing agorithms, Dynamic programming, FPTAS 1 INTRODUCTION We consider severa singe machine tota tardiness probems which can be formuated as foows We are given asetn = {1, 2,,n} of n independent jobs that must be processed on a singe machine Job preemption is not aowed The machine can hande ony one job at a time A jobs are assumed to be avaiabe for processing at time 0 For each job j N, a processing time p j > 0, a weight w j > 0 and a due date d j > 0aregiven A feasibe soution is described by a permutation π = (j 1,j 2,,j n ) of the jobs of the set N from which the corresponding schedue can be uniquey determined by starting each job as eary as possibe Let C jk (π) = k =1 p j be the competion time of job j k in the schedue resuting from the sequence π IfC j (π) >d j,thenjobj is tardy If C j (π) d j,thenjobj is on-time Moreover, et T j (π) =max{0,c j (π) d j } be the tardiness of job j in the schedue resuting from sequence π and et GT j (π) = min{max{0,c j (π) d j },p j } In the weighted tota tardiness minimization probem the objective is to find an optima job sequence π that minimizes weighted tota tardiness, ie, F (π) = n w jt j (π) Simiary, for the tota tardiness probem the objective function F (π) = n T j(π) and for the tota ate work probem the objective function F (π) = n GT j(π) have to be minimized The foowing specia cases of the probems are aso considered: Partiay supported by RFBR (Russian Foundation for Basic Research): , The authors are aso gratefu to Prof V Strusevich for his idea to use the GrA as a base for an FPTAS - minimizing weighted tota tardiness when a due dates are equa, ie, d j = d, j = 1, 2,,nThis probem is denoted by 1 d j = d w j T j ; - the specia case B 1 of the tota tardiness probem 1 T j,wherep 1 p 2 p n, d 1 d 2 d n and d n d 1 p n ; - the specia case B 1G of the probem 1 T j with d max d min p min,whered max = max j N {d j }, d min =min j N {d j } and p min =min j N {p j } For the NP-hard probem of maximizing weighted tota tardiness 1(no-ide) max w j T j, the objective is to find an optima job sequence that maximizes weighted tota tardiness, where each feasibe schedue starts at time 0 and there is no ide time between the processing of jobs Such probems with the maximum criterion have practica interpretations and appications, but their investigation is aso a theoreticay significant task A the probems and specia cases mentioned above are NP-hard in the ordinary sense (Gafarov et a (2012)) For the specia case B 1, an FPTAS with a running time of O(n 3 og n+n 3 /ε) is known, see Kouamas (2010) An FPTAS for the tota weighted ate work probem with the same running time was presented in Kovayov et a (1994) An FPTAS for the probem 1 d j = d w j T j with a running time of O((n 6 og w j )/ε 3 ) was presented in Keerer et a (2006) These probems can be considered as particuar cases when a compex function Ψ(t)+F (π, t) has to be minimized The function F (π, t) corresponds to one of the above mentioned functions F (π) if the jobs are processed not from time 0, but from time t Asan aternative, we have to partition the t-axis into intervas with the same optima schedue For exampe, the singe machine probem of minimizing the number of ate jobs,

2 when the starting time of the machine is variabe, was consid ered in Hoogeveen et a (2012) The same situation arises when it is known that some jobs have to be schedued one by one in a batch from an unknown time point t [t 1,t 2 ], eg, a set N contains two subsets N 1, N 2 and an optima job sequence can be represented as a concatenation (π 1,π 2,π 3 ), where {π 1 } {π 3 } = N 1 and {π 2 } = N 2 The jobs from N 2 have to be schedued according to one of the functions mentioned above The graphica and approximation agorithms presented in this paper can be used both for the initia probems and for the probems with variabe starting time Since the main topic of this paper is that of an anaysis of approximation agorithms, we reca some reevant definitions For the scheduing probem of minimizing a function F (π), a poynomia-time agorithm that finds a feasibe soution π such that F (π ) is at most ρ 1 times the optima vaue F (π ) is caed a ρ-approximation agorithm; the vaue of ρ is caed a worst-case ratio bound If a probem admits a ρ-approximation agorithm, it is said to be approximabe within a factor ρ A famiy of ρ- approximation agorithms is caed a fuy poynomia-time approximation scheme, or an FPTAS, if ρ =1+ε for any ε>0 and the running time is poynomia with respect to both the ength of the probem input and 1/ε For a practica reaization of some pseudo-poynomia agorithms with Booean variabes (eg a job can be ontime or tardy) one can use a modification caed a graphica agorithm (GrA) which we present for the probem 1 d j = d w j T j as we as an FPTAS based on this GrA with a running time of O(n 3 /ε) Modifications of these GrA and FPTAS for the cases B 1, B 1G and the probems 1 GT j and 1(no-ide) max w j T j are aso presented 2 DYNAMIC PROGRAMMING FOR THE PROBLEM 1 D J = D W J T J Lemma 1 There exists an optima job sequence π for probem 1 d j = d w j T j that can be represented as a concatenation (G, x, H), where a jobs j H are tardy and S j d for a j H, and a jobs i G are on-time A jobs from set G are processed in non-increasing order of the vaues pj w j and a jobs from set H are processed in non-decreasing order of the vaues pj w j Thejobx starts before time d and is competed no earier than time d The job x is caed stradding Assume that the jobs are numbered as foows: p2 w 2 p3 w 3 pn w n, (*) where the job with number 1 is the stradding job As a coroary from Lemma 1, there is a stradding job x N, to which the number 1 wi be assigned, such that for each {1, 2,,n}, there exists an optima schedue in which a jobs j {1, 2,,} are processed from time t one by one, and there is no job i {+1,+2,,n} which is processed between these jobs Thus, we can present a dynamic programming agorithm (DPA) based on Lemma 1 For each x N, we number the jobs from the set N \{x} according to (*) and perform Agorithm 1 Then we choose a best schedue among the n constructed schedues At each stage, 1 n, of Agorithm 1, we construct a best partia sequence π (t) for the set of jobs {1, 2,,} and for each possibe starting time t of the first job (which represents a possibe state in the DPA) F (t) denotes the weighted tota tardiness vaue for the job sequence π (t) Φ 1 (t) andφ 2 (t) are temporary functions, which are used to compute F (t) Agorithm 1 1 Number the jobs according to order (*); 2 FOR t := 0 TO n j=2 p j DO π 1 (t) :=(1),F 1 (t) :=w 1 max{0,p 1 + t d}; 3 FOR := 2 TO n DO FOR t := 0 TO n j=+1 p j DO π 1 := (, π 1 (t + p )), π 2 := (π 1 (t),); Φ 1 (t) :=w max{0,p + t d} + F 1 (t + p ); Φ 2 (t) :=F 1 (t)+w max{0, p j + t d}; IF Φ 1 (t) < Φ 2 (t) THENF (t) :=Φ 1 (t) and π (t) :=π 1, ELSE F (t) :=Φ 2 (t) andπ (t) :=π 2 ; 4 π n (0) is an optima job sequence for the chosen job x with the objective function vaue F n (0) Theorem 2 By using Agorithm 1 for each x N, an optima job sequence of the type described in Lemma 1 can be found in O(n 2 p j )time It is obvious that for some chosen job x N, in the job sequence π n (0), the job x cannot be stradding (ie, either S x d or C x <d) This means that there exists another job x N for which the vaue F n (0) wi be ess Agorithm 1 can be modified by considering for each =1, 2,,n, ony the interva [0,d p ] instead of the n interva [0, p i ] since for each t > d p, job j is i=+1 tardy in any partia sequence π (t) and the partia sequence π 2 := (π 1 (t),) is optima Thus, the time compexity of the modified Agorithm 1 is equa to O(nd), and an optima schedue can be found in O(n 2 d)time Let UB be an upper bound on the optima function vaue for the probem which is found by the 2-approximation agorithm of Fathi and Nutte, ie, UB 2F (π ) If for some t UB (, + ) wehavef (t UB )=UB,thenfor each t>t UB we have F (t) >UB(since t denotes the starting time of an optima schedue obtained from the job sequence π (t) forjobs1, 2,, and F (t) denotes the corresponding vaue of the monotonic objective function) So, the states t>t UB seem to be unpromising, ie, for any job sequence π constructed using these states, we wi have F (π) >UB, ie, π is not optima Thus, we need to consider different vaues F (t) onyfort [0,t UB ]and assume F (t) =+ for t (t UB, + ) If a parameters p j,w j for a j N and d are integer, then there are at most UB +2differentvauesF (t) In addition, if there is a point t (, + ) such that F (t )=0andF (t +1)> 0, then F (t) = 0 for a t t and F (t) <F (t+1) for a t t +1, since a the functions F (t) are monotonic Thus, we can modify Agorithm 1 as foows If we wi save at each stage instead of a states t [0,t ] ony one state t, then the number of saved states wi be restricted by UB, since for a saved states t we have F (t) <F (t + 1) The running time of

3 Fig 1 Function F (t) in the GrA and in Agorithm 1 the modified agorithm is O(n min{d, U B}) If we consider ony the states t [d n p j,d] instead of [0,d]ateach stage, =1, 2,,n, then we obtain an optima soution for the chosen stradding job for each possibe starting time t (,t UB n )ino(nub) time Let π be a job sequence, where a n jobs are processed in non-decreasing order of the vaues pj w j DenotebyF(π,d) the weighted tota tardiness for the job sequence π,where the processing of the jobs starts at time ditisobviousthat for this starting time the schedue π is optima Then, by using the modified Agorithm 1, we can obtain an optima soution for each possibe starting time t (, + ) in O(n 2 F (π,d)) time We note that the inequaity F (t) < F (t + 1) does not necessariy hod for a t>t in the probem 1 GT j, ie, the running time of Agorithm 1 is not restricted by UB for the probem 1 GT j 3 GRA FOR PROBLEM 1 D J = D W J T J The GrA is a modification of Agorithm 1, in which function F (t) is defined for any t (, + ) (not ony for integer t) However, we need to compute these vaues ony at the break points separating intervas in which function F (t) is a inear function of the form F (t) = F k(t) =uk (t tk 1 )+b k F (t) is a continuous piecewise inear function whose parameters can be described in a tabuar form Namey, in each step of the GrA, we store the information on function F (t) for a number of intervas (characterized by the same best partia sequence and by the same tota weight of tardy jobs) in a tabuar form as given in Tabe 1 Tabe 1: Function F (t) k 1 2 m +1 m +2 int (,t 1 ] (t1,t2 ] (tm,t m +1 ] (t m +1, ) b k 0 b 2 b m +1 u k 0 u 2 u m +1 0 π k π 1 π 2 π m +1 (1, 2,, ) In Tabe 1, k denotes the number of the current interva whose vaues range from 1 to m + 2 (where the number of intervas m + 2 is defined for each = 1, 2,,n), (t k 1,t k ]isthekth interva (where t0 =, tm +2 =+ = t UB ), b k, uk are the parameters of the inear and t m +1 function F k (t) defined in the kth interva, and πk is the best sequence of the first jobs if they are processed from time t (t k 1,t k ] These data mean the foowing For each above interva, we store the parameters b k and u k for describing function F (t) and the resuting best partia sequence if the first job starts in this interva For each starting time t (t k 1,t k ] (t 0 = ) of the first job, we have a best partia sequence π k of the jobs 1, 2,, with a tota weight of the tardy jobs u k and the function vaue F (t) =u k (t tk 1 )+b k (see Fig 1(a)) We have F (t) =0fort (t 0,t1 ] Reca that function F (t) is defined not ony for integers t, but aso for rea numbers t For simpicity of the foowing description, we consider the whoe t-axis, ie, t (, + ) This tabe describes a function F (t) which is a continuous, piecewise-inear function in the interva (,t UB ] The points t 1,t2,,tm +1 are caed break points, since there is a change from vaue u k 1 to u k (which means that the sope of the piecewise-inear function changes) Note that some of the break points t k can be non-integer To describe each inear segment, we store its sope u k and its function vaue b k = F (t) atthepointt = t k 1 So,in the tabe b 1 = b 2 <b 3 < < b m +1 <UBhods, since t 1 <t2 <<tm +1 In the GrA, the functions F (t) refect the same functiona equations as in Agorithm 1, ie, for each t Z [0, n j=2 p j], the function F (t) has the same vaue as in Agorithm 1 (see Fig 1), but these functions are now defined for any t (, + ) As a resut, often a arge number of integer states is combined into one interva (describing a new state in the resuting agorithm) with the same best partia sequence In Fig 1 (a), the function F (t) from the GrA is presented and in Fig 1 (b), the function F (t) from Agorithm 1 is dispayed The function Φ 1 (t) is obtained from F 1 (t) by the foowing operations We shift the diagram of function F 1 (t) to the eft by the vaue p and in the tabe for F 1 (t) and add a coumn which resuts from the new break point t = d p Ift s 1 p < t < t s+1 1 p, s m 1, then we have two new intervas of t in the tabe for Φ 1 (t): (t s 1 p,t ] and (t,t s+1 1 p ] (for s = m 1 +1, we have (t UB 1 p,t ]and(t, + )) This means that we first repace each interva (t k 1,tk+1 1 ]by(tk 1 p,t k+1 1 p ] in the tabe for Φ 1 (t), and then repace the coumn with the interva (t s 1 p,t s+1 1 p ] by two new coumns with the intervas (t s 1 p,t ]and(t,t s+1 1 p ] Moreover, we increase the vaues u s+1 1,us+2 1,,um by w, ie, the tota weight of tardy jobs (and thus the sope of the function) increases The corresponding partia sequences π 1 are obtained by adding job as the first job to each previous partia sequence The function Φ 2 (t) is obtained from function F 1 (t) by the foowing operations In the tabe for F 1 (t), we add a coumn which resuts from the new break point t = d i=1 p iift h 1 <t <t h+1 1, h m 1, then there are two new intervas of t in the tabe for Φ 2 (t): (t h 1,t ] and (t,t h+1 1 ](forh = m 1 +1,wehave(t UB 1,t ]and (t, + )) This means that we repace the coumn with the interva (t h 1,th+1 1 ] by two new coumns with the intervas (t h 1,t ]and(t,t h+1 1 ]

4 Moreover, we increase the vaues u h+1 1,uh+2 1,,um by w, ie, the tota weight of tardy jobs increases The corresponding partia sequences π 2 are obtained by adding job at the end to each previous partia sequence We construct a tabe that corresponds to the function F (t) = min{φ 1 (t), Φ 2 (t)} as foows We consider functions Φ 1 (t) and Φ 2 (t) on a resuting intervas from both tabes and search for intersection points of the diagrams of these functions To be more precise, we construct a ist t 1,t 2,,t e, t 1 < t 2 < < t e, of a break points t from the tabes for Φ 1 (t) andφ 2 (t), which are eft / right boundary points of the intervas given in these tabes Then we consider each interva (t i,t i+1 ],i=1, 2,,e 1, and compare the two functions Φ 1 (t) andφ 2 (t) overthisintervaletthe interva (t i,t i+1 ] be contained in (t z 1,t z ] from the tabe for Φ 1 (t) and in (t y 1,t y ] from the tabe for Φ 2 (t) Then Φ 1 (t) =(t t z 1 ) u z + b z and Φ 2 (t) =(t t y 1 ) u y + b y Choose the coumn from both tabes corresponding to the maximum of the two functions in (t i,t i+1 ]andinsert this coumn into the tabe for F (t) If there exists an intersection point t of Φ 1 (t) andφ 2 (t) inthisinterva, then insert two coumns with the intervas (t i,t ] and (t,t i+1 ] This step requires O(m 1 )operations Let m be the number of coumns in the resuting tabe of F (t) and for k, 1 k < m, the inequaity b k < UB b k+1 hods From the equaity UB = (t UB t k 1 )u k + b k, we compute the vaue tub Inthecoumn with the interva (t k 1,t k ](coumnk), assign tk = t UB and substitute a the coumns k+1,k+2,,m,byonecoumn with the interva (t UB, + ), b k+1 =+, u k+1 =0and π k+1 =(1, 2,,) So, in the resuting tabe, there wi be no more then UB + 2 coumns if a the parameters of the probem are integer In the tabe corresponding to function F n (t) we determine the coumn (t k n,t k+1 n ], where t k n < 0 t k+1 n Thenwehave an optima sequence π = πn k+1 for the chosen job x and the optima function vaue F (π )=b k+1 n +(0 t k n) u k+1 n Theorem 3 By using the GrA for each x N, an optima job sequence of the type described in Lemma 1 can be found in O(n 2 F ) time, where F is the optima objective function vaue 4 ADVANTAGES OF GRA FOR 1 D J = D W J T J In each step = 1, 2,,n of the GrA, we do not consider a points t [0,d], t Z, but ony points from the interva in which the optima partia soution changes or where the resuting functiona equation of the objective function changes So, the main difference is that we operate not with independent vaues F in each of the points t, but with functions which are transformed in each step anayticay (according to their anaytica form), which can have obvious advantages For exampe, et us minimize a function Ψ(t) + F (π, t), where the function F (π, t) corresponds to a function F (π) when the jobs are processed not starting from time 0, but from time t Ifthe function F (t) is presented anayticay (not in a tabuar form (t, F )) and the function Ψ(t) is presented anayticay as we, then the search for the minimum of Ψ(t)+F (π, t) canbemadeinshortertime Moreover, such an approach has the foowing advantages when compared with Agorithm 1 (DPA): 1 GrA can sove instances, where (some of) the parameters p j,w j,, 2,,n or/and d are not in Z 2 The running time of the GrA for two instances with the parameters {p j,w j,d} and {p j 10 const ± 1,w j 10 const ±1,d 10 const ±1},const>1, is the same whie the running time of the DPA wi be 10 const times arger in the second case Thus, one can usuay sove consideraby arger instances with the GrA 3 Properties of an optima soution can be taken into account, and sometimes the GrA has even a poynomia time compexity, or we can at east essentiay reduce the compexity of the standard DPA 4 Unike DPA, it is possibe to construct a FPTAS based on GrA easiy which is presented in Section 5 Let us consider another type of a DPA This agorithm generates iterativey some sets of states In every iteration, = n, n 1,,1, a set of states is generated Each state can be represented by a string of the form (t, F ), where t is the competion time of the ast known job schedued in the beginning of a schedue and F is the vaue of the function, provided that the eary jobs start at time 0 and the ast known ate job competes exacty at time n p This agorithm can be described as foows: Aternative DPA 1 Number the jobs according to order (*); 2 Put the state (0, 0) into the set of states V n+1 3 FOR := n TO 1 DO FOR each state (t, F ) from the set of states V +1 DO Put a state [t+p,f+ w max{0,t+ p d}]; Put a state [t, F + w max{0, n p j ( 1 p j t) d}]; 4 Find F =min{f (t, F ) V 1 } We need to consider ony states, where t d and F UB If in a ist V, there are two states with the same objective function vaue (t 1,F )and(t 2,F )andt 2 >t 1, then the state (t 2,F ) can be removed from consideration So, the running time of the Aternative DPA is O(n min{d, F }) which corresponds to the running time of the GrA (note that the GrA can be easiy modified to consider ony points t [0,d]) However, in the GrA some of the possibe but unpromising states are not considered and, in contrast to the aternative DPA, the GrA finds a optima schedues for a t [,t UB n ]ino(nf ) time So, the aternative DPA is not effective for the probems of minimizing a compex function Ψ(t)+F (π, t) 5 AN FPTAS FOR PROBLEM 1 D J = D W J T J εub 2n The idea of the FPTAS is as foows Let δ = To reduce the time compexity of the GrA, we have to diminish the number of coumns considered, which is the number of different objective function vaues 0 = b 1 = b 2,b3,,bm +1 = UB If we consider not the origina vaues b k but the vaues b k which are rounded up or down to the nearest mutipe of δ vaues b k, there are no more

5 than UB δ = 2n ε different vaues bk Then we wi be abe to convert the tabe F (t) into a simiar tabe with no more than 4 n ε coumns Furthermore, for such a modified tabe (function) F (t), we wi have F (t) F (t) <δ εf (π ) n If we do the rounding and modification after each stage of the GrA, then the cumuative error wi be no more than nδ εf (π ), and the tota running time of n runs of the GrA wi be O( n3 ε ), ie, an FPTAS is obtained By transforming the GrA, we save the approximated functions F A (t) in the same tabuar form but without the ast row describing an optima partia job sequence π k FPTAS (as a modification of the GrA) Assume that we have obtained a tabe for a function F (t) after stage with the objective function vaues b 1,b2,,bm +1 Ifm > 4 n ε, then do the foowing Round a the vaues b k from Tabe 1 to the nearest mutipe of δ Letthevauesb 1, b2,,bm +1 be obtained, where b 1 b 2 b m +1 We modify the tabe F A (t) as foows Assume that, for k 1 <k 2,wehaveb k1 < b k1+1 = = b k2 < b k2+1 We substitute the coumns which correspond to the vaues b k1,,b k2 1 for the two coumns presented in Tabe 2 Tabe 2: Substitution of coumns int k (t k1 1,t k1 ] (t k1,t k2 1 ] (t UB, ) b k b k1 u k u = bk 2 b k 1 t k 1 t k 1 1 b k2 0 0 In fact, in the modification we substitute some inear fragments which foow one by one and are cose (in terms of the absoute error which is δ/2) to the same ine by this ine Let F (t) be the function obtained at stage in the modified agorithm before rounding and et F A (t) describe the rounded function Lemma 4 For a t (t k1 1 (t) <δ/2 F,t k2 1 ], we have F A (t) Before the next stage + 1, we assign F A (t) := F (t) Assume that functions F (t), =1, 2,,n,are exact and constructed by the origina GrA Simiary, F A (t), = 1, 2,,n, are approximated functions constructed by the modified agorithm Lemma 5 For each, = 1, 2,, n, and a t (, + ), we have F A (t) F (t) δ/2 Let π be an optima job sequence for the chosen stradding job x Lemma 6 Inequaity F (π ) F (π ) n δ n 2ε F (π ) 2n hods Theorem 7 By using the modified GrA for each x N, a job sequence π of the type described in Lemma 1 wi be found in O( n3 ε ) time, where F (π ) (1 + ε)f (π ) The running time O( n3 ε ) is obtained as foows The modified GrA is used n times for each job x N The running time of the modified GrA depends on n and the number of coumns in the tabes which describe functions F A (t) The number of coumns does not exceed O( n ε ) 6 ALGORITHMS FOR THREE SPECIAL CASES Lemma 8 There exists an optima job sequence π for the specia case B 1G that can be represented as a concatenation (G, x, H), where a jobs j H are tardy and a jobs i G are on-time A jobs from set G are processed in non-increasing order of the vaues p j (LPT order) and a jobs from set H are processed in nondecreasing order of the vaues p j (SPT order) The job x is caed stradding Lemma 9 (Gafarov et a (2012)) There exists an optima job sequence π for the specia case B 1 that can be represented as a concatenation (G, H) A jobs from the set G are processed in LPT order and a jobs from the set H are processed in SPT order Assume that for the case B 1G, the jobs are numbered as foows: p 1 p 2 p n Lemma 10 For the specia cases B 1 and B 1G andajobsequenceπ SPT =(n, n 1,,1), inequaity F (π SPT ) 3F (π )hods Without oss of generaity, we wi consider ony cases whereinajobsequenceπ SPT at east two jobs are tardy Lemma 11 (Gafarov et a (2012)) There exists an optima job sequence π for the probem 1 GT j that can be represented as a concatenation (G, H), where a jobs j H are tardy and GT j (π) =p j For a jobs i G, we have 0 GT i (π) <p i A jobs from the set G are processed in EDD (eariest due date) order and a jobs from the set H are processed in LDD (ast due date) order Let for the probem 1 GT j thejobsbenumberedas foows: d 1 d 2 d n and et π EDD =(1, 2,,n) Denote by T the maxima tardiness of a job in the sequence π EDD, ie, T =max j N {GT j (π EDD )} Lemma 12 For probem 1 GT j, the foowing inequaity hods: F (π EDD ) nf (π ) Lemma 13 (Gafarov et a (2012)) There exists an optima job sequence π for the probem 1(no-ide) max w j T j that can be represented as a concatenation (G, H), where a jobs j H are tardy and a jobs i G are on-time A jobs from the set G are processed in non-increasing order of the vaues wj p j and a jobs from the set H are processed in non-decreasing order of the vaues wj p j For probem 1(no-ide) max w j T j, the vaue LB maxt T =max j N (w j ( n i=1 p i d j )) is a ower bound Then UB maxt T = nlb maxt T is an upper bound on the optima objective function vaue To sove these probems, Agorithms 1, GrA and FPTAS can be modified as foows For the specia case B 1G: - DPA Use the fact described in Lemma 8 In Agorithm 1, we number the jobs according to the order p 2 p 3 p n Assume that F 1 (t) := max{0,p 1 + t d 1 },Φ 1 (t) :=max{0,p + t d } + F 1 (t+p )andφ 2 (t) :=F 1 (t)+max{0, p j +t

6 d } A other steps of the agorithm remain the same Remember that w j =1foraj N for the probem 1 T j The running time of the modified Agorithm 1 is O(nd max ) Since it is necessary to consider n stradding jobs x N, an optima job sequence can be found in O(n 2 d max )timebyusingthemodified Agorithm 1; - GrA The GrA remains the same The parameters u k denote the number of tardy jobs which is equa to the tota weight of the tardy jobs, since w j =1fora j N Inaddition,assumethatUB = F (π SPT ) By using the GrA, an optima schedue can be found in O(n 2 min{d max,f })time; εf (πsp T ) 3n - FPTAS In the FPTAS, assume that δ = Since the GrA is without changes, the time compexity of the FPTAS based on the GrA for the specia case B 1G has a running time of O(n 3 /ε), which is ess than the running time O(n 7 /ε) ofthefptasfor the genera case presented by Lawer For the specia case B 1: - DPA Use the fact described in Lemma 9 Agorithm 1 is modified as for the specia case B 1G The running time of the modified Agorithm 1 is O(nd max ) Since there is no stradding job, an optima job sequence can be found in O(nd max )timebyusing the modified Agorithm 1 ony once; - GrA The GrA remains the same as for the specia case B 1G By the GrA, an optima schedue can be found in O(n min{d max,f })time; - FPTAS The FPTAS remains the same as for the specia case B 1G Since there is no stradding job, the FPTAS for the specia case B 1 has a running time of O(n 2 /ε), which is ess than the running time of O(n 3 og n + n 3 /ε) of the FPTAS mentioned by Kouamas For the probem 1 GT j : - DPA Use the fact described in Lemma 11 In Agorithm 1, we number the jobs according to the order d 1 d 2 d n Assume that F 1 (t) := min{p 1, max{0,p 1 + t d 1 }}, Φ 1 (t) := min{p, max{0,p +t d }}+F 1 (t+p )andφ 2 (t) := F 1 (t)+min{p, max{0, p j +t d }} The running time of the modified Agorithm 1 is O(nd max ) Since there is no stradding job, an optima job sequence can be found in O(nd max )timebyusingthemodified Agorithm 1 ony once; - GrA The GrA remains amost the same In addition to the break points t and t,twonewbreak points τ = d and τ = d 1 p j are considered The sope u k of the function F (t) is changed according to the function min{p, max{0,p + t d }} By the GrA, an optima schedue can be found in O(n min{d max,nf }) time, since UB = F (π EDD ) nf (π )andthereareatmost2ub + 2 coumns in each tabe F (t) considered in the GrA; εf (πedd) - FPTAS Here, we assume that δ = n So,the 2 FPTAS has a running time of O(n 3 /ε) For the probem 1(no-ide) max w j T j : - DPA We use the fact described in Lemma 13 In Agorithm 1, we enumerate the jobs according to the order w1 p 1 w2 p 2 wn p n We assume that F 1 (t) := w 1 max{0,p 1 + t d 1 }, Φ 1 (t) := w max{0,p + t d } { + F 1 (t + p ) and Φ 2 (t) := F 1 (t) + w max 0, p i + t d } Since tota tardiness is i=1 maximized, we have F (t) := max{φ 1 (t), Φ 2 (t)} The running time of the modified Agorithm 1 is O(nd max ) Since there is no stradding job, an optima job sequence can be found in O(nd max )timebyusing the modified Agorithm 1 ony once; - GrA The GrA remains the same as for the probem 1 d j = d w j T j WehaveF (t) =max{φ 1 (t), Φ 2 (t)} It is known (Gafarov et a (2012)) that the functions F (t) represent continuous, piecewise-inear and convex functions By the GrA, an optima schedue can be found in O(n min{d max,nf, w j })time - FPTAS In the FPTAS, assume that δ = εubmaxt n 2 T So, the FPTAS has a running time of O(n 3 /ε) In Gafarov et a (2012), GrA for 8 scheduing probems and FPTAS for 6 scheduing probems are presented 7 CONCLUSION In this paper, an FPTAS was presented, which can be used with some simpe modifications for severa singe machine probems The FPTAS is based on a graphica approach The idea of such a modification of graphica agorithms enabes us to construct an FPTAS easiy The graphica approach can be appied to probems, where a pseudo-poynomia agorithm exists and Booean variabes are used in the sense that yes/no decisions have to be made (eg, a job may be competed on-time or not) For the singe machine probem of maximizing tota tardiness, the graphica agorithm improved the compexity from O(n p j )too(n 2 ) Thus, the graphica approach has not ony a practica but aso a theoretica importance REFERENCES ER Gafarov, A Dogui and F Werner Dynamic Programming Approach to Design FPTAS for Singe Machine Scheduing Probems CNRS Report, 26 pp (2012) C Kouamas The Singe-Machine Tota Tardiness Scheduing Probem: Review and Extensions European Journa of Operationa Research, 202, 1 7 (2010) M Y Kovayov, C N Potts and L N van Wassenhove A Fuy Poynomia Approximation Scheme for Scheduing a Singe Machine to Minimize Tota Weighted Late Work Mathematics of Operations Research, Vo 19, No 1, (1994) H Hoogeveen and V T Kindt Minimizing the Number of Late Jobs When the Starting Time of the Machine is Variabe Proceedings PMS, (2010) H Keerer and VA Strusevich A Fuy Poynomia Approximation Scheme for the Singe Machine Weighted Tota Tardiness Probem with a Common Due Date Theoretica Computer Science, 369, (2006)

Single Machine Scheduling with Generalized Total Tardiness Objective Function

Single Machine Scheduling with Generalized Total Tardiness Objective Function Single Machine Scheduling with Generalized Total Tardiness Objective Function Evgeny R. Gafarov a, Alexander A. Lazarev b Institute of Control Sciences of the Russian Academy of Sciences, Profsoyuznaya

More information

CS229 Lecture notes. Andrew Ng

CS229 Lecture notes. Andrew Ng CS229 Lecture notes Andrew Ng Part IX The EM agorithm In the previous set of notes, we taked about the EM agorithm as appied to fitting a mixture of Gaussians. In this set of notes, we give a broader view

More information

Uniprocessor Feasibility of Sporadic Tasks with Constrained Deadlines is Strongly conp-complete

Uniprocessor Feasibility of Sporadic Tasks with Constrained Deadlines is Strongly conp-complete Uniprocessor Feasibiity of Sporadic Tasks with Constrained Deadines is Strongy conp-compete Pontus Ekberg and Wang Yi Uppsaa University, Sweden Emai: {pontus.ekberg yi}@it.uu.se Abstract Deciding the feasibiity

More information

Minimizing Total Weighted Completion Time on Uniform Machines with Unbounded Batch

Minimizing Total Weighted Completion Time on Uniform Machines with Unbounded Batch The Eighth Internationa Symposium on Operations Research and Its Appications (ISORA 09) Zhangiaie, China, September 20 22, 2009 Copyright 2009 ORSC & APORC, pp. 402 408 Minimizing Tota Weighted Competion

More information

Problem set 6 The Perron Frobenius theorem.

Problem set 6 The Perron Frobenius theorem. Probem set 6 The Perron Frobenius theorem. Math 22a4 Oct 2 204, Due Oct.28 In a future probem set I want to discuss some criteria which aow us to concude that that the ground state of a sef-adjoint operator

More information

Single Machine Scheduling with a Non-renewable Financial Resource

Single Machine Scheduling with a Non-renewable Financial Resource Single Machine Scheduling with a Non-renewable Financial Resource Evgeny R. Gafarov a, Alexander A. Lazarev b Institute of Control Sciences of the Russian Academy of Sciences, Profsoyuznaya st. 65, 117997

More information

XSAT of linear CNF formulas

XSAT of linear CNF formulas XSAT of inear CN formuas Bernd R. Schuh Dr. Bernd Schuh, D-50968 Kön, Germany; bernd.schuh@netcoogne.de eywords: compexity, XSAT, exact inear formua, -reguarity, -uniformity, NPcompeteness Abstract. Open

More information

A Brief Introduction to Markov Chains and Hidden Markov Models

A Brief Introduction to Markov Chains and Hidden Markov Models A Brief Introduction to Markov Chains and Hidden Markov Modes Aen B MacKenzie Notes for December 1, 3, &8, 2015 Discrete-Time Markov Chains You may reca that when we first introduced random processes,

More information

MARKOV CHAINS AND MARKOV DECISION THEORY. Contents

MARKOV CHAINS AND MARKOV DECISION THEORY. Contents MARKOV CHAINS AND MARKOV DECISION THEORY ARINDRIMA DATTA Abstract. In this paper, we begin with a forma introduction to probabiity and expain the concept of random variabes and stochastic processes. After

More information

Robust Sensitivity Analysis for Linear Programming with Ellipsoidal Perturbation

Robust Sensitivity Analysis for Linear Programming with Ellipsoidal Perturbation Robust Sensitivity Anaysis for Linear Programming with Eipsoida Perturbation Ruotian Gao and Wenxun Xing Department of Mathematica Sciences Tsinghua University, Beijing, China, 100084 September 27, 2017

More information

Approximate Bandwidth Allocation for Fixed-Priority-Scheduled Periodic Resources (WSU-CS Technical Report Version)

Approximate Bandwidth Allocation for Fixed-Priority-Scheduled Periodic Resources (WSU-CS Technical Report Version) Approximate Bandwidth Aocation for Fixed-Priority-Schedued Periodic Resources WSU-CS Technica Report Version) Farhana Dewan Nathan Fisher Abstract Recent research in compositiona rea-time systems has focused

More information

Algorithms to solve massively under-defined systems of multivariate quadratic equations

Algorithms to solve massively under-defined systems of multivariate quadratic equations Agorithms to sove massivey under-defined systems of mutivariate quadratic equations Yasufumi Hashimoto Abstract It is we known that the probem to sove a set of randomy chosen mutivariate quadratic equations

More information

Schedulability Analysis of Deferrable Scheduling Algorithms for Maintaining Real-Time Data Freshness

Schedulability Analysis of Deferrable Scheduling Algorithms for Maintaining Real-Time Data Freshness 1 Scheduabiity Anaysis of Deferrabe Scheduing Agorithms for Maintaining Rea-Time Data Freshness Song Han, Deji Chen, Ming Xiong, Kam-yiu Lam, Aoysius K. Mok, Krithi Ramamritham UT Austin, Emerson Process

More information

A. Distribution of the test statistic

A. Distribution of the test statistic A. Distribution of the test statistic In the sequentia test, we first compute the test statistic from a mini-batch of size m. If a decision cannot be made with this statistic, we keep increasing the mini-batch

More information

Componentwise Determination of the Interval Hull Solution for Linear Interval Parameter Systems

Componentwise Determination of the Interval Hull Solution for Linear Interval Parameter Systems Componentwise Determination of the Interva Hu Soution for Linear Interva Parameter Systems L. V. Koev Dept. of Theoretica Eectrotechnics, Facuty of Automatics, Technica University of Sofia, 1000 Sofia,

More information

MATH 172: MOTIVATION FOR FOURIER SERIES: SEPARATION OF VARIABLES

MATH 172: MOTIVATION FOR FOURIER SERIES: SEPARATION OF VARIABLES MATH 172: MOTIVATION FOR FOURIER SERIES: SEPARATION OF VARIABLES Separation of variabes is a method to sove certain PDEs which have a warped product structure. First, on R n, a inear PDE of order m is

More information

Research Article On the Lower Bound for the Number of Real Roots of a Random Algebraic Equation

Research Article On the Lower Bound for the Number of Real Roots of a Random Algebraic Equation Appied Mathematics and Stochastic Anaysis Voume 007, Artice ID 74191, 8 pages doi:10.1155/007/74191 Research Artice On the Lower Bound for the Number of Rea Roots of a Random Agebraic Equation Takashi

More information

Cryptanalysis of PKP: A New Approach

Cryptanalysis of PKP: A New Approach Cryptanaysis of PKP: A New Approach Éiane Jaumes and Antoine Joux DCSSI 18, rue du Dr. Zamenhoff F-92131 Issy-es-Mx Cedex France eiane.jaumes@wanadoo.fr Antoine.Joux@ens.fr Abstract. Quite recenty, in

More information

Week 6 Lectures, Math 6451, Tanveer

Week 6 Lectures, Math 6451, Tanveer Fourier Series Week 6 Lectures, Math 645, Tanveer In the context of separation of variabe to find soutions of PDEs, we encountered or and in other cases f(x = f(x = a 0 + f(x = a 0 + b n sin nπx { a n

More information

Schedulability Analysis of Deferrable Scheduling Algorithms for Maintaining Real-Time Data Freshness

Schedulability Analysis of Deferrable Scheduling Algorithms for Maintaining Real-Time Data Freshness 1 Scheduabiity Anaysis of Deferrabe Scheduing Agorithms for Maintaining Rea- Data Freshness Song Han, Deji Chen, Ming Xiong, Kam-yiu Lam, Aoysius K. Mok, Krithi Ramamritham UT Austin, Emerson Process Management,

More information

C. Fourier Sine Series Overview

C. Fourier Sine Series Overview 12 PHILIP D. LOEWEN C. Fourier Sine Series Overview Let some constant > be given. The symboic form of the FSS Eigenvaue probem combines an ordinary differentia equation (ODE) on the interva (, ) with a

More information

The Price of Bounded Preemption

The Price of Bounded Preemption The Price of Bounded Preemption ABSTRACT Noga Aon * Te-Aviv University Te-Aviv, Israe Princeton University Princeton, New Jersey, USA nogaa@post.tau.ac.i In this paper we provide a tight bound for the

More information

Absolute Value Preconditioning for Symmetric Indefinite Linear Systems

Absolute Value Preconditioning for Symmetric Indefinite Linear Systems MITSUBISHI ELECTRIC RESEARCH LABORATORIES http://www.mer.com Absoute Vaue Preconditioning for Symmetric Indefinite Linear Systems Vecharynski, E.; Knyazev, A.V. TR2013-016 March 2013 Abstract We introduce

More information

4 1-D Boundary Value Problems Heat Equation

4 1-D Boundary Value Problems Heat Equation 4 -D Boundary Vaue Probems Heat Equation The main purpose of this chapter is to study boundary vaue probems for the heat equation on a finite rod a x b. u t (x, t = ku xx (x, t, a < x < b, t > u(x, = ϕ(x

More information

Sequential Decoding of Polar Codes with Arbitrary Binary Kernel

Sequential Decoding of Polar Codes with Arbitrary Binary Kernel Sequentia Decoding of Poar Codes with Arbitrary Binary Kerne Vera Miosavskaya, Peter Trifonov Saint-Petersburg State Poytechnic University Emai: veram,petert}@dcn.icc.spbstu.ru Abstract The probem of efficient

More information

A proposed nonparametric mixture density estimation using B-spline functions

A proposed nonparametric mixture density estimation using B-spline functions A proposed nonparametric mixture density estimation using B-spine functions Atizez Hadrich a,b, Mourad Zribi a, Afif Masmoudi b a Laboratoire d Informatique Signa et Image de a Côte d Opae (LISIC-EA 4491),

More information

Approximated MLC shape matrix decomposition with interleaf collision constraint

Approximated MLC shape matrix decomposition with interleaf collision constraint Approximated MLC shape matrix decomposition with intereaf coision constraint Thomas Kainowski Antje Kiese Abstract Shape matrix decomposition is a subprobem in radiation therapy panning. A given fuence

More information

Asynchronous Control for Coupled Markov Decision Systems

Asynchronous Control for Coupled Markov Decision Systems INFORMATION THEORY WORKSHOP (ITW) 22 Asynchronous Contro for Couped Marov Decision Systems Michae J. Neey University of Southern Caifornia Abstract This paper considers optima contro for a coection of

More information

Approximated MLC shape matrix decomposition with interleaf collision constraint

Approximated MLC shape matrix decomposition with interleaf collision constraint Agorithmic Operations Research Vo.4 (29) 49 57 Approximated MLC shape matrix decomposition with intereaf coision constraint Antje Kiese and Thomas Kainowski Institut für Mathematik, Universität Rostock,

More information

Convergence Property of the Iri-Imai Algorithm for Some Smooth Convex Programming Problems

Convergence Property of the Iri-Imai Algorithm for Some Smooth Convex Programming Problems Convergence Property of the Iri-Imai Agorithm for Some Smooth Convex Programming Probems S. Zhang Communicated by Z.Q. Luo Assistant Professor, Department of Econometrics, University of Groningen, Groningen,

More information

Pattern Frequency Sequences and Internal Zeros

Pattern Frequency Sequences and Internal Zeros Advances in Appied Mathematics 28, 395 420 (2002 doi:10.1006/aama.2001.0789, avaiabe onine at http://www.ideaibrary.com on Pattern Frequency Sequences and Interna Zeros Mikós Bóna Department of Mathematics,

More information

Approximation and Fast Calculation of Non-local Boundary Conditions for the Time-dependent Schrödinger Equation

Approximation and Fast Calculation of Non-local Boundary Conditions for the Time-dependent Schrödinger Equation Approximation and Fast Cacuation of Non-oca Boundary Conditions for the Time-dependent Schrödinger Equation Anton Arnod, Matthias Ehrhardt 2, and Ivan Sofronov 3 Universität Münster, Institut für Numerische

More information

8 Digifl'.11 Cth:uits and devices

8 Digifl'.11 Cth:uits and devices 8 Digif'. Cth:uits and devices 8. Introduction In anaog eectronics, votage is a continuous variabe. This is usefu because most physica quantities we encounter are continuous: sound eves, ight intensity,

More information

Two-sample inference for normal mean vectors based on monotone missing data

Two-sample inference for normal mean vectors based on monotone missing data Journa of Mutivariate Anaysis 97 (006 6 76 wwweseviercom/ocate/jmva Two-sampe inference for norma mean vectors based on monotone missing data Jianqi Yu a, K Krishnamoorthy a,, Maruthy K Pannaa b a Department

More information

An explicit Jordan Decomposition of Companion matrices

An explicit Jordan Decomposition of Companion matrices An expicit Jordan Decomposition of Companion matrices Fermín S V Bazán Departamento de Matemática CFM UFSC 88040-900 Forianópois SC E-mai: fermin@mtmufscbr S Gratton CERFACS 42 Av Gaspard Coriois 31057

More information

arxiv: v1 [math.fa] 23 Aug 2018

arxiv: v1 [math.fa] 23 Aug 2018 An Exact Upper Bound on the L p Lebesgue Constant and The -Rényi Entropy Power Inequaity for Integer Vaued Random Variabes arxiv:808.0773v [math.fa] 3 Aug 08 Peng Xu, Mokshay Madiman, James Mebourne Abstract

More information

Lecture Notes 4: Fourier Series and PDE s

Lecture Notes 4: Fourier Series and PDE s Lecture Notes 4: Fourier Series and PDE s 1. Periodic Functions A function fx defined on R is caed a periodic function if there exists a number T > such that fx + T = fx, x R. 1.1 The smaest number T for

More information

Tight Approximation Algorithms for Maximum Separable Assignment Problems

Tight Approximation Algorithms for Maximum Separable Assignment Problems MATHEMATICS OF OPERATIONS RESEARCH Vo. 36, No. 3, August 011, pp. 416 431 issn 0364-765X eissn 156-5471 11 3603 0416 10.187/moor.1110.0499 011 INFORMS Tight Approximation Agorithms for Maximum Separabe

More information

MA 201: Partial Differential Equations Lecture - 10

MA 201: Partial Differential Equations Lecture - 10 MA 201: Partia Differentia Equations Lecture - 10 Separation of Variabes, One dimensiona Wave Equation Initia Boundary Vaue Probem (IBVP) Reca: A physica probem governed by a PDE may contain both boundary

More information

Stochastic Variational Inference with Gradient Linearization

Stochastic Variational Inference with Gradient Linearization Stochastic Variationa Inference with Gradient Linearization Suppementa Materia Tobias Pötz * Anne S Wannenwetsch Stefan Roth Department of Computer Science, TU Darmstadt Preface In this suppementa materia,

More information

CONSISTENT LABELING OF ROTATING MAPS

CONSISTENT LABELING OF ROTATING MAPS CONSISTENT LABELING OF ROTATING MAPS Andreas Gemsa, Martin Nöenburg, Ignaz Rutter Abstract. Dynamic maps that aow continuous map rotations, for exampe, on mobie devices, encounter new geometric abeing

More information

Algorithms for Special Single Machine Total Tardiness Problems

Algorithms for Special Single Machine Total Tardiness Problems Algorithms for Special Single Machine Total Tardiness Problems Alexander A. Lazarev Institute of Control Sciences of the Russian Academy of Sciences, Profsoyuznaya st. 65, 117997 Moscow, Russia, email:

More information

On the Goal Value of a Boolean Function

On the Goal Value of a Boolean Function On the Goa Vaue of a Booean Function Eric Bach Dept. of CS University of Wisconsin 1210 W. Dayton St. Madison, WI 53706 Lisa Heerstein Dept of CSE NYU Schoo of Engineering 2 Metrotech Center, 10th Foor

More information

The EM Algorithm applied to determining new limit points of Mahler measures

The EM Algorithm applied to determining new limit points of Mahler measures Contro and Cybernetics vo. 39 (2010) No. 4 The EM Agorithm appied to determining new imit points of Maher measures by Souad E Otmani, Georges Rhin and Jean-Marc Sac-Épée Université Pau Veraine-Metz, LMAM,

More information

Throughput Optimal Scheduling for Wireless Downlinks with Reconfiguration Delay

Throughput Optimal Scheduling for Wireless Downlinks with Reconfiguration Delay Throughput Optima Scheduing for Wireess Downinks with Reconfiguration Deay Vineeth Baa Sukumaran vineethbs@gmai.com Department of Avionics Indian Institute of Space Science and Technoogy. Abstract We consider

More information

Throughput Rate Optimization in High Multiplicity Sequencing Problems

Throughput Rate Optimization in High Multiplicity Sequencing Problems Throughput Rate Optimization in High Mutipicity Sequencing Probems Aexander Grigoriev A.Grigoriev@KE.unimaas.n Maastricht University, 6200 MD, Maastricht, The Netherands Joris van de Kundert J.vandeKundert@MATH.unimaas.n

More information

CONJUGATE GRADIENT WITH SUBSPACE OPTIMIZATION

CONJUGATE GRADIENT WITH SUBSPACE OPTIMIZATION CONJUGATE GRADIENT WITH SUBSPACE OPTIMIZATION SAHAR KARIMI AND STEPHEN VAVASIS Abstract. In this paper we present a variant of the conjugate gradient (CG) agorithm in which we invoke a subspace minimization

More information

T.C. Banwell, S. Galli. {bct, Telcordia Technologies, Inc., 445 South Street, Morristown, NJ 07960, USA

T.C. Banwell, S. Galli. {bct, Telcordia Technologies, Inc., 445 South Street, Morristown, NJ 07960, USA ON THE SYMMETRY OF THE POWER INE CHANNE T.C. Banwe, S. Gai {bct, sgai}@research.tecordia.com Tecordia Technoogies, Inc., 445 South Street, Morristown, NJ 07960, USA Abstract The indoor power ine network

More information

Iterative Decoding Performance Bounds for LDPC Codes on Noisy Channels

Iterative Decoding Performance Bounds for LDPC Codes on Noisy Channels Iterative Decoding Performance Bounds for LDPC Codes on Noisy Channes arxiv:cs/060700v1 [cs.it] 6 Ju 006 Chun-Hao Hsu and Achieas Anastasopouos Eectrica Engineering and Computer Science Department University

More information

Efficiently Generating Random Bits from Finite State Markov Chains

Efficiently Generating Random Bits from Finite State Markov Chains 1 Efficienty Generating Random Bits from Finite State Markov Chains Hongchao Zhou and Jehoshua Bruck, Feow, IEEE Abstract The probem of random number generation from an uncorreated random source (of unknown

More information

Statistical Learning Theory: A Primer

Statistical Learning Theory: A Primer Internationa Journa of Computer Vision 38(), 9 3, 2000 c 2000 uwer Academic Pubishers. Manufactured in The Netherands. Statistica Learning Theory: A Primer THEODOROS EVGENIOU, MASSIMILIANO PONTIL AND TOMASO

More information

Lecture Note 3: Stationary Iterative Methods

Lecture Note 3: Stationary Iterative Methods MATH 5330: Computationa Methods of Linear Agebra Lecture Note 3: Stationary Iterative Methods Xianyi Zeng Department of Mathematica Sciences, UTEP Stationary Iterative Methods The Gaussian eimination (or

More information

Explicit overall risk minimization transductive bound

Explicit overall risk minimization transductive bound 1 Expicit overa risk minimization transductive bound Sergio Decherchi, Paoo Gastado, Sandro Ridea, Rodofo Zunino Dept. of Biophysica and Eectronic Engineering (DIBE), Genoa University Via Opera Pia 11a,

More information

Another Look at Linear Programming for Feature Selection via Methods of Regularization 1

Another Look at Linear Programming for Feature Selection via Methods of Regularization 1 Another Look at Linear Programming for Feature Seection via Methods of Reguarization Yonggang Yao, The Ohio State University Yoonkyung Lee, The Ohio State University Technica Report No. 800 November, 2007

More information

DISTRIBUTION OF TEMPERATURE IN A SPATIALLY ONE- DIMENSIONAL OBJECT AS A RESULT OF THE ACTIVE POINT SOURCE

DISTRIBUTION OF TEMPERATURE IN A SPATIALLY ONE- DIMENSIONAL OBJECT AS A RESULT OF THE ACTIVE POINT SOURCE DISTRIBUTION OF TEMPERATURE IN A SPATIALLY ONE- DIMENSIONAL OBJECT AS A RESULT OF THE ACTIVE POINT SOURCE Yury Iyushin and Anton Mokeev Saint-Petersburg Mining University, Vasiievsky Isand, 1 st ine, Saint-Petersburg,

More information

Nearly Optimal Constructions of PIR and Batch Codes

Nearly Optimal Constructions of PIR and Batch Codes arxiv:700706v [csit] 5 Jun 07 Neary Optima Constructions of PIR and Batch Codes Hia Asi Technion - Israe Institute of Technoogy Haifa 3000, Israe shea@cstechnionaci Abstract In this work we study two famiies

More information

6 Wave Equation on an Interval: Separation of Variables

6 Wave Equation on an Interval: Separation of Variables 6 Wave Equation on an Interva: Separation of Variabes 6.1 Dirichet Boundary Conditions Ref: Strauss, Chapter 4 We now use the separation of variabes technique to study the wave equation on a finite interva.

More information

Partial permutation decoding for MacDonald codes

Partial permutation decoding for MacDonald codes Partia permutation decoding for MacDonad codes J.D. Key Department of Mathematics and Appied Mathematics University of the Western Cape 7535 Bevie, South Africa P. Seneviratne Department of Mathematics

More information

A Fundamental Storage-Communication Tradeoff in Distributed Computing with Straggling Nodes

A Fundamental Storage-Communication Tradeoff in Distributed Computing with Straggling Nodes A Fundamenta Storage-Communication Tradeoff in Distributed Computing with Stragging odes ifa Yan, Michèe Wigger LTCI, Téécom ParisTech 75013 Paris, France Emai: {qifa.yan, michee.wigger} @teecom-paristech.fr

More information

New Efficiency Results for Makespan Cost Sharing

New Efficiency Results for Makespan Cost Sharing New Efficiency Resuts for Makespan Cost Sharing Yvonne Beischwitz a, Forian Schoppmann a, a University of Paderborn, Department of Computer Science Fürstenaee, 3302 Paderborn, Germany Abstract In the context

More information

Primal and dual active-set methods for convex quadratic programming

Primal and dual active-set methods for convex quadratic programming Math. Program., Ser. A 216) 159:469 58 DOI 1.17/s117-15-966-2 FULL LENGTH PAPER Prima and dua active-set methods for convex quadratic programming Anders Forsgren 1 Phiip E. Gi 2 Eizabeth Wong 2 Received:

More information

THE THREE POINT STEINER PROBLEM ON THE FLAT TORUS: THE MINIMAL LUNE CASE

THE THREE POINT STEINER PROBLEM ON THE FLAT TORUS: THE MINIMAL LUNE CASE THE THREE POINT STEINER PROBLEM ON THE FLAT TORUS: THE MINIMAL LUNE CASE KATIE L. MAY AND MELISSA A. MITCHELL Abstract. We show how to identify the minima path network connecting three fixed points on

More information

Distributed average consensus: Beyond the realm of linearity

Distributed average consensus: Beyond the realm of linearity Distributed average consensus: Beyond the ream of inearity Usman A. Khan, Soummya Kar, and José M. F. Moura Department of Eectrica and Computer Engineering Carnegie Meon University 5 Forbes Ave, Pittsburgh,

More information

On colorings of the Boolean lattice avoiding a rainbow copy of a poset arxiv: v1 [math.co] 21 Dec 2018

On colorings of the Boolean lattice avoiding a rainbow copy of a poset arxiv: v1 [math.co] 21 Dec 2018 On coorings of the Booean attice avoiding a rainbow copy of a poset arxiv:1812.09058v1 [math.co] 21 Dec 2018 Baázs Patkós Afréd Rényi Institute of Mathematics, Hungarian Academy of Scinces H-1053, Budapest,

More information

Homogeneity properties of subadditive functions

Homogeneity properties of subadditive functions Annaes Mathematicae et Informaticae 32 2005 pp. 89 20. Homogeneity properties of subadditive functions Pá Burai and Árpád Száz Institute of Mathematics, University of Debrecen e-mai: buraip@math.kte.hu

More information

Bayesian Learning. You hear a which which could equally be Thanks or Tanks, which would you go with?

Bayesian Learning. You hear a which which could equally be Thanks or Tanks, which would you go with? Bayesian Learning A powerfu and growing approach in machine earning We use it in our own decision making a the time You hear a which which coud equay be Thanks or Tanks, which woud you go with? Combine

More information

On the evaluation of saving-consumption plans

On the evaluation of saving-consumption plans On the evauation of saving-consumption pans Steven Vanduffe Jan Dhaene Marc Goovaerts Juy 13, 2004 Abstract Knowedge of the distribution function of the stochasticay compounded vaue of a series of future

More information

Maximum likelihood decoding of trellis codes in fading channels with no receiver CSI is a polynomial-complexity problem

Maximum likelihood decoding of trellis codes in fading channels with no receiver CSI is a polynomial-complexity problem 1 Maximum ikeihood decoding of treis codes in fading channes with no receiver CSI is a poynomia-compexity probem Chun-Hao Hsu and Achieas Anastasopouos Eectrica Engineering and Computer Science Department

More information

Paper presented at the Workshop on Space Charge Physics in High Intensity Hadron Rings, sponsored by Brookhaven National Laboratory, May 4-7,1998

Paper presented at the Workshop on Space Charge Physics in High Intensity Hadron Rings, sponsored by Brookhaven National Laboratory, May 4-7,1998 Paper presented at the Workshop on Space Charge Physics in High ntensity Hadron Rings, sponsored by Brookhaven Nationa Laboratory, May 4-7,998 Noninear Sef Consistent High Resoution Beam Hao Agorithm in

More information

arxiv: v1 [math.co] 17 Dec 2018

arxiv: v1 [math.co] 17 Dec 2018 On the Extrema Maximum Agreement Subtree Probem arxiv:1812.06951v1 [math.o] 17 Dec 2018 Aexey Markin Department of omputer Science, Iowa State University, USA amarkin@iastate.edu Abstract Given two phyogenetic

More information

(This is a sample cover image for this issue. The actual cover is not yet available at this time.)

(This is a sample cover image for this issue. The actual cover is not yet available at this time.) (This is a sampe cover image for this issue The actua cover is not yet avaiabe at this time) This artice appeared in a journa pubished by Esevier The attached copy is furnished to the author for interna

More information

Rate-Distortion Theory of Finite Point Processes

Rate-Distortion Theory of Finite Point Processes Rate-Distortion Theory of Finite Point Processes Günther Koiander, Dominic Schuhmacher, and Franz Hawatsch, Feow, IEEE Abstract We study the compression of data in the case where the usefu information

More information

Symbolic models for nonlinear control systems using approximate bisimulation

Symbolic models for nonlinear control systems using approximate bisimulation Symboic modes for noninear contro systems using approximate bisimuation Giordano Poa, Antoine Girard and Pauo Tabuada Abstract Contro systems are usuay modeed by differentia equations describing how physica

More information

Expectation-Maximization for Estimating Parameters for a Mixture of Poissons

Expectation-Maximization for Estimating Parameters for a Mixture of Poissons Expectation-Maximization for Estimating Parameters for a Mixture of Poissons Brandon Maone Department of Computer Science University of Hesini February 18, 2014 Abstract This document derives, in excrutiating

More information

Asymptotic Properties of a Generalized Cross Entropy Optimization Algorithm

Asymptotic Properties of a Generalized Cross Entropy Optimization Algorithm 1 Asymptotic Properties of a Generaized Cross Entropy Optimization Agorithm Zijun Wu, Michae Koonko, Institute for Appied Stochastics and Operations Research, Caustha Technica University Abstract The discrete

More information

The Binary Space Partitioning-Tree Process Supplementary Material

The Binary Space Partitioning-Tree Process Supplementary Material The inary Space Partitioning-Tree Process Suppementary Materia Xuhui Fan in Li Scott. Sisson Schoo of omputer Science Fudan University ibin@fudan.edu.cn Schoo of Mathematics and Statistics University of

More information

Combining reaction kinetics to the multi-phase Gibbs energy calculation

Combining reaction kinetics to the multi-phase Gibbs energy calculation 7 th European Symposium on Computer Aided Process Engineering ESCAPE7 V. Pesu and P.S. Agachi (Editors) 2007 Esevier B.V. A rights reserved. Combining reaction inetics to the muti-phase Gibbs energy cacuation

More information

Indirect Optimal Control of Dynamical Systems

Indirect Optimal Control of Dynamical Systems Computationa Mathematics and Mathematica Physics, Vo. 44, No. 3, 24, pp. 48 439. Transated from Zhurna Vychisite noi Matematiki i Matematicheskoi Fiziki, Vo. 44, No. 3, 24, pp. 444 466. Origina Russian

More information

Intuitionistic Fuzzy Optimization Technique for Nash Equilibrium Solution of Multi-objective Bi-Matrix Games

Intuitionistic Fuzzy Optimization Technique for Nash Equilibrium Solution of Multi-objective Bi-Matrix Games Journa of Uncertain Systems Vo.5, No.4, pp.27-285, 20 Onine at: www.jus.org.u Intuitionistic Fuzzy Optimization Technique for Nash Equiibrium Soution of Muti-objective Bi-Matri Games Prasun Kumar Naya,,

More information

Lower Bounds for the Relative Greedy Algorithm for Approximating Steiner Trees

Lower Bounds for the Relative Greedy Algorithm for Approximating Steiner Trees This paper appeared in: Networks 47:2 (2006), -5 Lower Bounds for the Reative Greed Agorithm for Approimating Steiner Trees Stefan Hougard Stefan Kirchner Humbodt-Universität zu Berin Institut für Informatik

More information

IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 15, NO. 2, FEBRUARY

IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 15, NO. 2, FEBRUARY IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 5, NO. 2, FEBRUARY 206 857 Optima Energy and Data Routing in Networks With Energy Cooperation Berk Gurakan, Student Member, IEEE, OmurOze,Member, IEEE,

More information

Numerical solution of one dimensional contaminant transport equation with variable coefficient (temporal) by using Haar wavelet

Numerical solution of one dimensional contaminant transport equation with variable coefficient (temporal) by using Haar wavelet Goba Journa of Pure and Appied Mathematics. ISSN 973-1768 Voume 1, Number (16), pp. 183-19 Research India Pubications http://www.ripubication.com Numerica soution of one dimensiona contaminant transport

More information

<C 2 2. λ 2 l. λ 1 l 1 < C 1

<C 2 2. λ 2 l. λ 1 l 1 < C 1 Teecommunication Network Contro and Management (EE E694) Prof. A. A. Lazar Notes for the ecture of 7/Feb/95 by Huayan Wang (this document was ast LaT E X-ed on May 9,995) Queueing Primer for Muticass Optima

More information

Do Schools Matter for High Math Achievement? Evidence from the American Mathematics Competitions Glenn Ellison and Ashley Swanson Online Appendix

Do Schools Matter for High Math Achievement? Evidence from the American Mathematics Competitions Glenn Ellison and Ashley Swanson Online Appendix VOL. NO. DO SCHOOLS MATTER FOR HIGH MATH ACHIEVEMENT? 43 Do Schoos Matter for High Math Achievement? Evidence from the American Mathematics Competitions Genn Eison and Ashey Swanson Onine Appendix Appendix

More information

Statistical Inference, Econometric Analysis and Matrix Algebra

Statistical Inference, Econometric Analysis and Matrix Algebra Statistica Inference, Econometric Anaysis and Matrix Agebra Bernhard Schipp Water Krämer Editors Statistica Inference, Econometric Anaysis and Matrix Agebra Festschrift in Honour of Götz Trenker Physica-Verag

More information

Converting Z-number to Fuzzy Number using. Fuzzy Expected Value

Converting Z-number to Fuzzy Number using. Fuzzy Expected Value ISSN 1746-7659, Engand, UK Journa of Information and Computing Science Vo. 1, No. 4, 017, pp.91-303 Converting Z-number to Fuzzy Number using Fuzzy Expected Vaue Mahdieh Akhbari * Department of Industria

More information

Assignment 7 Due Tuessday, March 29, 2016

Assignment 7 Due Tuessday, March 29, 2016 Math 45 / AMCS 55 Dr. DeTurck Assignment 7 Due Tuessday, March 9, 6 Topics for this week Convergence of Fourier series; Lapace s equation and harmonic functions: basic properties, compuations on rectanges

More information

Power Control and Transmission Scheduling for Network Utility Maximization in Wireless Networks

Power Control and Transmission Scheduling for Network Utility Maximization in Wireless Networks ower Contro and Transmission Scheduing for Network Utiity Maximization in Wireess Networks Min Cao, Vivek Raghunathan, Stephen Hany, Vinod Sharma and. R. Kumar Abstract We consider a joint power contro

More information

Mat 1501 lecture notes, penultimate installment

Mat 1501 lecture notes, penultimate installment Mat 1501 ecture notes, penutimate instament 1. bounded variation: functions of a singe variabe optiona) I beieve that we wi not actuay use the materia in this section the point is mainy to motivate the

More information

A NOTE ON QUASI-STATIONARY DISTRIBUTIONS OF BIRTH-DEATH PROCESSES AND THE SIS LOGISTIC EPIDEMIC

A NOTE ON QUASI-STATIONARY DISTRIBUTIONS OF BIRTH-DEATH PROCESSES AND THE SIS LOGISTIC EPIDEMIC (January 8, 2003) A NOTE ON QUASI-STATIONARY DISTRIBUTIONS OF BIRTH-DEATH PROCESSES AND THE SIS LOGISTIC EPIDEMIC DAMIAN CLANCY, University of Liverpoo PHILIP K. POLLETT, University of Queensand Abstract

More information

2M2. Fourier Series Prof Bill Lionheart

2M2. Fourier Series Prof Bill Lionheart M. Fourier Series Prof Bi Lionheart 1. The Fourier series of the periodic function f(x) with period has the form f(x) = a 0 + ( a n cos πnx + b n sin πnx ). Here the rea numbers a n, b n are caed the Fourier

More information

SydU STAT3014 (2015) Second semester Dr. J. Chan 18

SydU STAT3014 (2015) Second semester Dr. J. Chan 18 STAT3014/3914 Appied Stat.-Samping C-Stratified rand. sampe Stratified Random Samping.1 Introduction Description The popuation of size N is divided into mutuay excusive and exhaustive subpopuations caed

More information

Distributed Optimization With Local Domains: Applications in MPC and Network Flows

Distributed Optimization With Local Domains: Applications in MPC and Network Flows Distributed Optimization With Loca Domains: Appications in MPC and Network Fows João F. C. Mota, João M. F. Xavier, Pedro M. Q. Aguiar, and Markus Püsche Abstract In this paper we consider a network with

More information

An Algorithm for Pruning Redundant Modules in Min-Max Modular Network

An Algorithm for Pruning Redundant Modules in Min-Max Modular Network An Agorithm for Pruning Redundant Modues in Min-Max Moduar Network Hui-Cheng Lian and Bao-Liang Lu Department of Computer Science and Engineering, Shanghai Jiao Tong University 1954 Hua Shan Rd., Shanghai

More information

NOISE-INDUCED STABILIZATION OF STOCHASTIC DIFFERENTIAL EQUATIONS

NOISE-INDUCED STABILIZATION OF STOCHASTIC DIFFERENTIAL EQUATIONS NOISE-INDUCED STABILIZATION OF STOCHASTIC DIFFERENTIAL EQUATIONS TONY ALLEN, EMILY GEBHARDT, AND ADAM KLUBALL 3 ADVISOR: DR. TIFFANY KOLBA 4 Abstract. The phenomenon of noise-induced stabiization occurs

More information

Coupling of LWR and phase transition models at boundary

Coupling of LWR and phase transition models at boundary Couping of LW and phase transition modes at boundary Mauro Garaveo Dipartimento di Matematica e Appicazioni, Università di Miano Bicocca, via. Cozzi 53, 20125 Miano Itay. Benedetto Piccoi Department of

More information

King Fahd University of Petroleum & Minerals

King Fahd University of Petroleum & Minerals King Fahd University of Petroeum & Mineras DEPARTMENT OF MATHEMATICAL SCIENCES Technica Report Series TR 369 December 6 Genera decay of soutions of a viscoeastic equation Saim A. Messaoudi DHAHRAN 3161

More information

Wavelet shrinkage estimators of Hilbert transform

Wavelet shrinkage estimators of Hilbert transform Journa of Approximation Theory 163 (2011) 652 662 www.esevier.com/ocate/jat Fu ength artice Waveet shrinkage estimators of Hibert transform Di-Rong Chen, Yao Zhao Department of Mathematics, LMIB, Beijing

More information

Homework 5 Solutions

Homework 5 Solutions Stat 310B/Math 230B Theory of Probabiity Homework 5 Soutions Andrea Montanari Due on 2/19/2014 Exercise [5.3.20] 1. We caim that n 2 [ E[h F n ] = 2 n i=1 A i,n h(u)du ] I Ai,n (t). (1) Indeed, integrabiity

More information

Codes between MBR and MSR Points with Exact Repair Property

Codes between MBR and MSR Points with Exact Repair Property Codes between MBR and MSR Points with Exact Repair Property 1 Toni Ernva arxiv:1312.5106v1 [cs.it] 18 Dec 2013 Abstract In this paper distributed storage systems with exact repair are studied. A construction

More information