European Journal of Operational Research

Size: px
Start display at page:

Download "European Journal of Operational Research"

Transcription

1 European Journal of Operatonal Research 221 (2012) Contents lsts avalable at ScVerse ScenceDrect European Journal of Operatonal Research journal homepage: Producton, Manufacturng and Logstcs Enhanced lateral transshpments n a mult-locaton nventory system Coln Paterson a,, Ruud Teunter b,1, Kevn Glazebrook a,2 a Lancaster Unversty Management School, Lancaster, LA1 4YX, UK b Unversty of Gronngen, PO Box 800, 9700 AV, Gronngen, The Netherlands artcle nfo abstract Artcle hstory: Receved 25 November 2010 Accepted 4 March 2012 Avalable onlne 12 March 2012 Keywords: Inventory control Lateral transshpments Dynamc programmng In managng an nventory network, two approaches to the poolng of stock have been proposed. Reactve transshpments respond to shortages at a locaton by movng nventory from elsewhere wthn the network, whle proactve stock redstrbuton seeks to mnmze the chance of future stockouts. Ths paper s the frst to propose an enhanced reactve approach n whch ndvdual transshpments are vewed as an opportunty for proactve stock redstrbuton. We adopt a quas-myopc approach to the development of a strongly performng enhanced reactve transshpment polcy. In comparson to a purely reactve approach to transshpment, servce levels are mproved whle a reducton n safety stock levels s acheved. The aggregate costs ncurred n managng the system are sgnfcantly reduced, especally so for large networks. Moreover, an optmal polcy s determned for small networks and t s shown that the enhanced reactve polcy substantally closes the gap to optmalty. Ó 2012 Elsever B.V. All rghts reserved. 1. Introducton Lateral transshpments (LTs) are stock movements between locatons n the same echelon of an nventory system. They provde a valuable tool to supply chan managers who are lookng to reduce the penaltes assocated wth a lack of stock at one or more nventory ponts. By strategcally reallocatng excess stock t can be possble to mprove the system wde servce levels and/or lower the cost of operatng the system. These goals have tradtonally been sought wthn spare part networks, where there s a hgh penalty attached to a shortage. However the benefts of LTs have also been realzed n sectors rangng from retal to energy generaton. The challenge that LTs brng s n managng when and where t s benefcal to nstgate a stock movement. An LT may reduce the short term shortage rsk at the recevng locaton but t nevtably ncreases the longer term rsk at the sendng locaton. A transshpment polcy must therefore balance these contrastng rsks and decde when the cost of transshpment s outweghed by the beneft t s expected to delver. The sutablty of a gven LT polcy wll often depend on the attrbutes of the nventory system n whch t s employed. However, a key dstncton wthn the lterature on LTs s that between reactve and proactve polces. Reactve LTs are performed when a shortage or potental shortage occurs, by shppng ether the whole Correspondng author. Tel.: E-mal addresses: c.paterson@lancaster.ac.uk (C. Paterson), r.h.teunter@rug.nl (R. Teunter), k.glazebrook@lancaster.ac.uk (K. Glazebrook). 1 Tel.: Tel.: customer demand or the number of unts short from a dfferent locaton. Proactve transshpments are performed perodcally to rebalance the whole system s stock levels. Ths paper s prncple motvaton s n consderng an enhanced reactve polcy whch falls between these two dstnct sets so as to maxmze the beneft each transshpment can delver. Rather than merely lookng to meet the excess demand, the proposed polcy vews each trggered transshpment as an opportunty to proactvely rebalance the two nteractng locatons nventory. Often when a transshpment occurs the cost assocated wth the stock movement wll prmarly be a fxed cost, ndependent of the sze of the transshpment. The reason for ths s that regardless of whether one tem s transported or several, the costs such as usng a vehcle and the assocated fuel cost of nstgatng the journey wll be hghest for the frst tem. The margnal cost for subsequent tems wll typcally be much lower. When such cost structures exst t s mportant to know how best to carry out transshpments. Economes of scale are consdered throughout nventory management, from orderng n batches to centralzng warehouses. It s therefore natural to want to know how best to operate a transshpment polcy when the opportunty to extract smlar benefts exsts. Wthn the exstng lterature Reactve LTs have been studed under both a perodc and contnuous nventory revew settng. For perodc revew models, Krshnan and Rao (1965) develop optmal transshpments n a sngle perod for a system wth two locatons. Ths s expanded to a mult-locaton, mult-perod settng by Robnson (1990), although here the optmal reactve soluton can only be determned for ether two locatons or multple dentcal locatons and when the transshpment cost structure does not /$ - see front matter Ó 2012 Elsever B.V. All rghts reserved.

2 318 C. Paterson et al. / European Journal of Operatonal Research 221 (2012) nclude a fxed element. Ths hghlghts the complexty of determnng optmal transshpment polces. These papers perform LTs once all demand s known but before t has to be satsfed. In contrast, Archbald (2007) and Archbald et al. (2009, 2010) develop approxmately optmal polces whch can respond to contnuous demand wthn each perod. The former proposes heurstcs to deal wth the transshpment decson process, whle the latter papers look to mprove upon ths and relax some of the restrctons usng dynamc programmng polcy mprovement technques. The results obtaned from these polces show them to be reasonably close to optmal when used n small networks. Ths method of valdaton s one whch ths paper looks to emulate. The above models focus on sngle echelon centralzed models. However, addtonal research n the perodc settng consders the beneft of LTs wthn two echelon models (e.g. Dong and Rud, 2004), decentralzed models (e.g. Rud et al., 2001) and producton based models (e.g. Zhao et al., 2008). The latter of these s closely related to ths paper as t also consders an enhanced transshpment polcy. However, t consders a producton model where stock can be reallocated upon producton whlst also allowng reactve transshpments. It therefore falls between a producton allocaton model and a reactve transshpment model. Much of the lterature on reactve LTs n a contnuous order revew settng s motvated by applcatons n the spare parts ndustry. Here, practcal settngs nclude electronc component manufacturng and electrcty generaton companes. Buldng on the METRIC repars model of Sherbrooke (1968), Lee (1987) proposes a model whch uses complete poolng wthn preset groups of dentcal locatons. Ths shows the beneft of LTs wthn the area and the model s expanded by Axsäter (1990) to allow non-dentcal locatons. Several papers have been wrtten whch further expand these deas by relaxng or tghtenng some constrants such as makng repar capacty fnte (Jung et al., 2003), usng lost sales rather than backorderng (Dada, 1992) or consderng a model where backorders have to be mnmzed rather than costs (Sherbrooke, 1986). In addton to ths, nventory systems that supply more than one type of tem are nvestgated by papers such as Wong et al. (2006b) and Kranenburg and van Houtum (2009). The latter examnes the benefts of partal poolng, where only certan transshpments are performed. All of these papers assume an order-up-to replenshment polcy for each locaton. Wjk et al. (2009) consder a sngle tem system where parts are repared at each locaton and use dynamc programmng to determne an optmal transshpment polcy. Kukreja and Schmdt (2005) consder a more general (s,s) polcy, but have to resort to a smulaton based approach to determne the optmal order polcy. Away from spare parts, Archbald et al. (1997) shows that n a perodc revew model wthout fxed order costs, an order-up-to polcy s optmal. However, postve order costs or mnmum order quanttes often suggest that an (R,Q) polcy s more approprate n practce. Several papers take ths approach. Evers (2001) and Mnner et al. (2003) develop heurstcs that can be used to determne when and how much to transshp for systems wth lost sales. Axsäter (2003) does the same, but for a model wth backorders. He proposes a decson rule whch s constructed to make optmal decsons under an assumpton that no further transshpments wll be made. Ths assumpton enables the exact myopc beneft of transshppng to be calculated and optmzed. A related model by Axsäter et al. (2010) consders an (R, Q) nventory system n practce. They determne approxmately optmal replenshment polcy parameters when transshpments are sourced from a support warehouse. Research on proactve LTs explores ther use to rebalance an entre system s stock on hand. Ths rebalancng s done at a set pont durng a revew perod and before all demand has been realzed. Allen (1958) and Agrawal et al. (2004) consder ths problem ndependently of replenshment decsons. Allen (1958) looks to perform the transshpments at the start of the demand perod, whlst Agrawal et al. (2004) devse a method to calculate the best tme to redstrbute stock durng the perod. Other authors study proactve transshpment and replenshment decsons together. However, due to the perodc nature of proactve redstrbuton all known studes only consder ther use alongsde a perodc revew replenshment polcy. Gross (1963) provdes optmalty results for a two-locaton system, where both orderng and redstrbuton decsons take place at the begnnng of the revew perod. Ths dea s further developed by Das (1975), who allows the redstrbuton pont to occur at an arbtrary tme durng the revew perod. Gross and Das both assume neglgble transshpment tmes. Jönsson and Slver (1987) and Bertrand and Bookbnder (1998) allow postve transshpment tmes. The man dfference between these two studes s that Jönsson and Slver (1987) consder how best to meet servce levels whlst Bertrand and Bookbnder (1998) examne the goal of cost reducton. For a detaled overvew of the lterature we refer to Paterson et al. (2011). However, the hghlghted lterature shows that both reactve and proactve LTs provde cost benefts, but the cost benefts of proactve LTs have only been exploted n a perodc revew settng. In ths study, we analyze the frst hybrd transshpment polcy whch tres to secure the benefts of both under a contnuous revew replenshment polcy by enhancng a tradtonal reactve approach. Our polcy can quckly react to shortages by allowng transshpments at any tme when they occur, as for prevously proposed reactve LT polces. However, the polcy also seeks to proactvely redstrbute stock between the sendng and recevng locatons whenever such an LT s trggered. Ths wll allow maxmum beneft to be extracted from each transshpment nstance and wll be especally benefcal n systems where there s a sgnfcant fxed cost nvolved n carryng out a transshpment. The specfc settng that we consder s as n Axsäter (2003), wth backorderng and an arbtrary number of stockng locatons whch all apply (R, Q) orderng polces. Axsäter (2003) derves an algorthm that determnes near-optmal reactve transshpment decsons. These are shown n a smulaton study on small networks (wth two and three locatons) to provde a sgnfcant cost beneft compared to not transshppng at all and to applyng a smpler transshpment polcy. In ths paper, we generalze ths algorthm wth the goal of determnng an approxmately optmal enhanced reactve transshpment polcy that allows addtonal stock redstrbuton when reactng to a stock out. The results of a comparatve numercal study show that, for small networks, the enhanced polcy sgnfcantly outperforms the orgnal Axsäter reactve proposal, achevng an average 1.6% cost savng over 600 experments. Such a recurrent savng s of major practcal mportance, consderng that nventory costs typcally account for a substantal proporton of a busness s total turnover. To analyze the closeness to optmalty of our enhanced polces, we also develop a dynamc programmng (DP) approach to fndng an -optmal transshpment polcy whch also allows for a proactve element n each transshpment. More sgnfcantly we show through numercal results that the optmalty gap s closed by over 95% on average compared to a polcy of not transshppng and by 88% compared to the orgnal reactve polcy. Ths s strong evdence that our development of an enhanced reactve approach makes an mportant contrbuton to the applcaton of transshpments and enables benefts whch are close to optmal. In a further numercal study, we compare the tradtonal and enhanced algorthms for larger networks wth 5 20 locatons. The exact DP algorthm s too numercally ntensve to be appled n these experments. The results of a comparson of the polces show that the mprovement of the enhanced reactve polcy over the tradtonal reactve polcy s even larger than for small

3 C. Paterson et al. / European Journal of Operatonal Research 221 (2012) networks, wth a cost reducton of over 6.4% on average. It also provdes an average savng of 14.4% over not transshppng at all. A senstvty study provdes further nsghts nto when the cost reducton s most sgnfcant. The study also hghlghts the addtonal benefts that an mproved transshpment polcy can delver. The average servce level wthn the large network study s mproved by 1.5% ponts by the enhanced polcy over the purely reactve polcy and the amount of safety stock requred s reduced by over half when compared to a polcy of no transshpments. The remander of the paper s organzed as follows. In the next secton, we descrbe the model and the algorthm for computng our approxmately optmal enhanced reactve transshpment polcy. The cost beneft of the enhanced reactve polcy s tested numercally n Secton 3 for networks wth two locatons. Secton 4 descrbes the exact DP algorthm, and the optmalty gap for small networks s nvestgated n Secton 5. Larger networks are explored n Secton 6. We end wth conclusons and drectons for further research n Secton An enhanced transshpment polcy An approxmately optmal polcy for determnng purely reactve transshpment decsons n an nventory system, n whch all locatons follow a contnuous revew (R, Q) orderng polcy s derved by Axsäter (2003). The algorthm determnng the polcy s constructed by makng use of an assumpton that the consdered transshpment wll be the last one ever made. Whenever a locaton experences a shortage, the algorthm calculates the most cost effcent amount and locaton to transshp from under ths assumpton. However, the derved polcy only looks to react to a shortage, not to be proactve n future shortage preventon. The polcy proposed n ths paper allows more stock to be transshpped than s needed to meet the mmedate shortage. Ths permts the two locatons whch are partes to a transshpment to redstrbute ther stock and balance future rsk. For the sake of clarty n the remander of the paper we refer to the proposed polcy as the enhanced polcy, wth the exstng polcy referred to as the reactve polcy. Table 1 provdes a lst of notaton needed to defne the nventory system. Table 1 Lst of Notaton. N R Q t s, IL IP L D (L ) A h b c,j (y) c f ;j c u ;j k l f,j f n ;j P n ;j C C (k) X X 0 (d) V (X,t) a (X ) b (IP ) c (X ) Number of demand locatons: locaton 2 {1,...,N} Reorder pont at Batch sze of orders placed at Tme untl sth unt of stock becomes avalable at Inventory level at Inventory poston at Lead tme of orders placed at Stochastc lead tme demand at Order cost at (per order) Holdng cost at (per tem per unt tme) Backorder cost at (per tem per unt tme) Cost of transshppng y unts from to j Fxed cost of transshppng from to j Cost per unt of transshppng from to j Arrval rate of demand nstances at Average sze of each demand at Probablty that a demand at wll be of sze j Probablty that j unts are demanded by n customers at Probablty that nth customer demands the jth unt at Steady state cost of locaton Expected cost rate at tme L gven current IP = k State varable at locaton State varable at locaton less d unts of stock Expected total cost at durng the nterval [0,t] Expected lead tme bas at gven startng state X Expected bas after lead tme at gven current IP Total expected bas at gven startng state X The system has N stockng locatons whch all experence ndependent compound Posson demand wth locaton havng arrval ntensty k. Modelng demand as a compound Posson process enables the polcy to be appled n a wde range of nventory systems that allow customers to demand more than one tem at a tme. The probablty that a demand at locaton s of sze j s denoted by f,j and we do not prescrbe any dstrbutonal form. It should also be noted that P n ;j, the probablty that the nth customer demands at locaton the jth unt of stock, can be calculated usng the recurson P n ;j ¼ Pn ;j 1 f n ;j 1 þ f n 1 ;j 1. To replensh ts stock locaton 2 {1,...,N} places an order of sze Q at the central suppler whenever ts nventory poston drops to or below reorder level R. Ths suppler s assumed to have suffcent nventory capacty so that ths order can always be met and takes a fxed lead tme L to arrve. If a locaton does not have suffcent stock on hand to satsfy a demand, then tems may be transshpped to that locaton from a dfferent locaton wth neglgble lead tme. Ths setup s vald as stockng ponts are often geographcally closer than that of the suppler and do not have the external logstcal delay of placng an order wth a suppler. Any demand that cannot be met mmedately (after transshppng) s backordered. Costs are ncurred for orderng (A ), holdng stock on hand (h per tem and per tme unt), backorderng (b per tem and per tme unt), and for transshppng. The cost for transshppng y unts from locaton to j s gven by c ;j ðyþ ¼c f þ ;j ycu ;j, where cf ;j s the fxed cost per transshpment and c u ;j s the cost per unt transshpped. Under the assumpton that no transshpments take place, Axsäter (2003) derves an expresson for the bas assocated wth each system state. Ths measures the transent effect on costs of startng the system n that state. Performng a transshpment wll nstantly move the system to a new state, so the beneft of a gven transshpment can be dentfed by comparng the bas of the current state wth the aggregate of the bas f the transshpment s enacted and the cost to enact t. By maxmzng the dfference between these quanttes over all possble locatons and transshpment quanttes, the best myopc decson can be dentfed. The lmtaton of the formulaton gven by Axsäter (2003) s that t does not allow the sze of the transshpment to be larger than the shortage. Ths restrcton s mathematcally convenent n that t ensures that the nventory poston of any locaton never exceeds R + Q and thus provdes a closed range on whch to consder the future bas. However, when there s a sgnfcant fxed cost per transshpment t seems ntutve that allowng larger transshpments may well delver greater cost benefts. Not every such enhanced transshpment wll take the nventory poston above R + Q, but to provde a completely general account and to calculate the beneft assocated wth every possble decson t s necessary to extend the methodology to account for such scenaros. Please note that, whle ths extenson s certanly needed n general, we shall encounter n Secton 4 two locaton set-ups where almost all of the (very consderable) benefts from enhanced transshpments can actually be acheved whle keepng the nventory below R + Q. We now descrbe how the above deas can be developed to yeld near-optmal decsons n our enhanced approach. In steady state under an (R,Q ) replenshment polcy wthout transshpments t s known that locaton s nventory poston s unformly dstrbuted over the range [R +1,...,R + Q ]. Wth current (tme zero) nventory poston k and stochastc lead tme demand D (L ) then the mean nventory cost rate at L s gven by C ðkþ ¼h Eðk D ðl ÞÞ þ þ b Eðk D ðl ÞÞ ¼ðh þ b ÞEðk D ðl ÞÞ þ þ b EðD ðl Þ kþ X k 1 ¼ðh þ b Þe k L ðk jþ Xj ðk L Þ n f n ;j n! þ b ðk l L kþ: j¼0 n¼0 ð2:1þ

4 320 C. Paterson et al. / European Journal of Operatonal Research 221 (2012) We nfer that the steady state cost rate for locaton s gven by C ¼ 1 Q R PþQ k¼r þ1 C ðkþ: ð2:2þ The current state X of locaton ncorporates nformaton on the current nventory poston (IP ) together wth the tmes at whch each nventory tem becomes avalable. We wrte V (X,t) for the total expected cost ncurred under an (R,Q ) orderng polcy wth no transshpments durng the tme nterval [0, t] when the system s n state X at tme 0. The bas assocated wth X s defned as c ðx Þ, lm t!1 fv ðx ; tþ tc g: ð2:3þ Suppose that t > 0 and the random state of locaton at tme L s denoted by X L ; we then have V ðx ; L þ tþ ¼V ðx ; L ÞþEV ðx L ; tþ: We now easly see that ð2:4þ c ðx Þ¼lm t!1 fv ðx ; L þ tþ ðl þ tþc g; ð2:5þ has a decomposton gven by c ðx Þ¼a ðx Þþb ðip Þ; where a ðx Þ, V ðx ; L Þ L C ; ð2:6þ ð2:7þ and n o b ðip Þ, lm t!1 EV ðx L ; tþ tc : ð2:8þ The calculaton of a (X ) s gven n the Appendx and s unchanged by the fact that here we may have IP > R + Q. However the earler method for calculatng b s dependent on the nventory poston at tme L beng n the range [R +1,...,R + Q ]. A partcularly large proactve transshpment may nvaldate ths assumpton. Therefore n our formulaton we must take account of the addtonal bas that wll be arse f the nventory poston s taken above R + Q. We now descrbe the calculaton of b. We frst observe that, snce X L s stochastcally ndependent of all nformaton n X save only IP, then b depends upon X only through IP. More specfcally, IL L, the nventory level at tme L, can be calculated through the relaton IL L ¼ IP D ðl Þ: ð2:9þ We also note from Axsäter (2003) the followng notaton to show the relatonshp between IL and IP at any pont n tme. For a gven IL = k the correspondng nventory poston can be gven by hk ¼k þ nq ; ð2:10þ where n s the smallest nteger such that hk P R +1. As locaton evolves under an (R,Q ) orderng polcy wth no transshpments, the assocated nventory poston process regenerates upon every entry nto state R + 1 (or ndeed any other state R + 1 s chosen for convenence). In (2.8) we now deem X L to be the locaton state at tme zero and wrte T for the frst subsequent tme at whch the correspondng nventory poston, gven by rearrangng (2.9), s n the regeneraton state R + 1. Standard theory, for example n Tjms (1986), parttons the nfnte horzon consdered n (2.8) as [0,1) = [0,T) [ [T,1) and demonstrates that the contrbuton to b (IP ) from the latter nterval s zero. Ths reasonng yelds both b ðip Þ¼EV ðx L ; TÞ EðTÞC ; and ð2:11þ b ðr þ 1Þ ¼0: ð2:12þ We now fx IP = k > R + 1 and develop an expresson for b (k) by condtonng upon the sze of the frst demand to occur after tme zero. The expected tme at whch ths frst demand occurs s k 1 and from (2.11) the cost rate contrbuton to b (k) pror to then s C (k) C. Followng demand d the nventory poston at wll be changed to hk d and hence we obtan b ðkþ ¼ fc ðkþ C g k þ P1 d¼1 f ;d b ½< k d >Š: ð2:13þ We now use (2.13) to recover b as the lmt of an teratve scheme whch s akn to DP value teraton. It proceeds as follows: b 0 ðkþ ¼0; k 2½R þ 1;...; R þ Q ;...; SŠ; ð2:14þ b n ðr þ 1Þ ¼0; n P 0; ð2:15þ b nþ1 ðkþ ¼ fc ðkþ C g k þ P1 d¼1 f ;d b n ðhk dþ; k 2½R þ 2;...; R þ Q ;...; SŠ; ð2:16þ where S s a large nventory poston state such that no hgher state s reached. The scheme must converge geometrcally fast, wth b the lmt. We can now use the complete bas functons c for each locaton to dentfy the transshpment that delvers the most cost beneft. We use the notatonal shorthand X 0 ðdþ for the resultng state of a locaton once d unts of nventory have been wthdrawn, ether through customer demand or for transshpment. Suppose that some demand d occurs at locaton when n state X and that ths causes a shortage. The long term cost beneft of transshppng y unts from locaton j (n current state X j ) to locaton n comparson wth performng no transshpment and allowng all demand to be absorbed at s dentfed as the quantty n Dðj; yþ ¼lm V ðx 0 t!1 ðdþ; tþþv o jðx j ; tþ V X 0 ðd yþ; t V j X 0 j ðyþ; t c j; ðyþ yða j =Q j A =Q Þ: From (2.3) we have Dðj; yþ ¼c X 0 ðdþ þ cj ðx j Þ c X 0 ðd yþ cj X 0 j ðyþ c j; ðyþ yða j =Q j A =Q Þ: ð2:17þ ð2:18þ We note that, pror to (2.17), we have taken no account of the order costs n our calculatons. A transshpment of y unts from j to has the effect of adjustng the long-term cost burden from orders by y(a j /Q j A /Q ). We wrte D ¼ max max Dðj; yþ: 16y6IL j j ð2:19þ Our enhanced polcy s as follows: If D 6 0, do not transshp. If D > 0 transshp y unts from j to, where y and j are the maxmzers n (2.19). Ths decson rule s nvoked every tme a demand occurs at a locaton whch would result n a shortage. Ths polcy s quas-myopc n that t s cost mnmzng f no further transshpment s permtted after the current shortage s dealt wth. As the reactve polcy of Axsäter (2003) s a more constraned case of our general enhanced polcy, under the myopc assumpton any enhanced decson wll be at least as good as that of the purely reactve polcy. 3. Two locaton smulaton study: reactve vs. enhanced polcy To analyze the performance of the enhanced transshpment polcy, an ntal smulaton study s conducted. Ths study s restrcted to networks wth two dentcal locatons. In a second study, of

5 C. Paterson et al. / European Journal of Operatonal Research 221 (2012) whch the detals and results wll be presented n later sectons, larger and more vared networks wll be consdered. The two reasons for the more restrcted ntal exploraton are as follows. Frst, there are many model parameters and varyng all of them for larger networks s tme consumng. Ths ntal exploraton allows us to observe how the parameters mpact on costs so that the larger network can focus on key ssues. Secondly, the restrcton wll allow us to determne the optmal transshpment polcy usng DP n Secton 5 for the same set of experments, and therefore to study the relatve reducton of the optmalty gap from the orgnal reactve polcy to the new enhanced transshpment polcy. Snce the two locatons are assumed dentcal, we wll drop the locaton dentfyng subscrpt for cost parameters n the remander of ths secton. We wll do the same for the polcy parameters, and assume that both locatons use the same replenshment and transshpment polcy. Although the transshpment polcy mght have some effect on the optmal orderng quantty, n practce there are often fxed or mnmal order szes. For ths reason and n lne wth prevous studes, pffffffffffffffffffffffffffffffffffffffffffffffff we fx the order cost ($100) and use the EOQ formula ð2aklþ=ðhþ to determne the order sze. The valdty of ths decson s dscussed below. For the szes of successve demands wthn the study we use a geometrc dstrbuton such that f j = p(1 p) j 1, j P 1. The holdng cost rate ($1) s used as the unt cost and all other model parameters are vared n a full factoral study. The full range of parameters examned s shown n Table 2. There are 600 parameter combnatons and hence 600 experments n total. For both the orgnal reactve and enhanced transshpment polces, the transshpment decsons are determned by the algorthms developed n Secton 2. To obtan further nsghts we also consder the no transshpment polcy as a benchmark. For all three (no, reactve and enhanced) transshpment polces, the optmal value of R s found post hoc by conductng smulaton studes on the full range of possble values. Over the 600 experments an average savng of 1.59% s observed for the enhanced polcy when compared to the reactve polcy. Ths s broken down by each parameter n Table 3. For example, the 200 experments whch have an arrval rate of 2.4 customers per unt tme yeld an average savng of 1.77%. The results n Table 3 confrm ntuton n several ways. Once a transshpment s nstgated, a polcy whch looks to transshp more wll see a greater beneft f the margnal cost of addng an extra unt s low. Ths s supported by the c u results. Further, f a polcy s performng each transshpment more effcently then t s natural that the savng wll ncrease when there are more transshpment opportuntes. By ncreasng parameters k or L, or decreasng parameter p we rase the lead tme demand varablty and thus the chance of a shortage. Unsurprsngly, ncreasng the shortage rsk through varyng these parameters n the manner descrbed leads to an observed gan n savngs. Consderng the backorder costs, we see that the savngs are greater as the penalty for not mmedately meetng a demand ncreases. However, as the fxed transshpment cost ncreases the dynamcs are less straghtforward. As the cost per transshpment Table 2 Parameter values. Arrval rate (k) 0.8, 2.4, 4.0 Geometrc dstrbuton parameter for demand sze (p) 0.6, 0.8 Lead tme (L) 2, 3 Backorder cost (b) 10, 20, 30, 40, 50($) Transshpment cost [per tem] (c u ) 1, 2($) Transshpment cost [per transshpment] (c f ) 10, 20, 30, 40, 50($) Table 3 Enhanced rule vs reactve rule. k p L b c f c u % % % % % % % % % % % % % % % % % % % ncreases then naturally so does the reward for carryng out more effectve transshpments but at the same tme the beneft assocated wth carryng out any transshpment decreases. Ths can be observed n Table 3 where the beneft conferred by the enhanced approach ncreases up to a pont before decreasng. Ths s supported by the average number of transshpments per 100 unts of tme, whch falls from 11.2 when c f = 10 to only 3.3 when c f = 50. More detaled results are provded n the Appendx for a sample of parameter values (27 experments). All statstcal comparsons have used pared t-tests at a 95% confdence level, wth common random numbers used for all polces. Overall, the average savng of the new enhanced polcy compared to one whch uses no transshpments s 4.21%, renforcng the vew that transshppng s worthwhle. The standard errors of the costs for the reactve and the new enhanced polces are also gven n Table A.1 (n brackets). There are a few low demand cases where there s no statstcally sgnfcant dfference between the reactve and enhanced polces but no cases were found where the enhanced polcy was outperformed by the exstng polcy. We are not able to prove that the enhanced polcy always outperforms the reactve polcy but we have studed a very large number of problem nstances and have as yet found no counterexamples. A comparson of the safety stock levels R provdes nsght nto how the cost reducton s acheved. The average reorder level s 9.8 unts wthout transshpments, 8.9 unts for the reactve transshpment polcy and 8.2 unts for the new enhanced transshpment polcy. Better transshpment decsons reduce the negatve cost effects of shortages, thereby allowng the system to functon wth lower safety stocks. Further exploraton also showed that n general the enhanced transshpment polcy s more cost effcent, wth transshpments happenng less frequently but wth larger quanttes (4.7 unts for the enhanced polcy compared to 2.1 unts for the reactve polcy on average). To examne the effect of the batch sze on the relatve performance of the polcy, the reorder pont R, whch was optmsed usng a value of Q set by the EOQ, was fxed and used to determne the correspondng optmal batch sze n a smlar fashon. It should be noted that ths par may stll not be the global optmum. The results from ths showed that the dfference between the local optmal value of Q and the EOQ value s relatvely consstent between the polces, wth a dfference of 2 unts for the reactve polcy and 2.2 for the enhanced polcy. In fact the relatve performance of the two polces remaned constant, wth the performance gan remanng at 1.6%. Ths ndcates that the observed gans are robust when the batch sze s vared and that usng the EOQ value gves a sutable overvew of the polces performance. The results n ths secton have clearly shown that the new enhanced transshpment polcy sgnfcantly outperforms the orgnal reactve transshpment polcy. However, t remans of real nterest to dscover how close (n cost terms) t s to an optmal transshpment polcy. To explore ths, we wll develop a dynamc programmng (DP) formulaton for fndng the optmal transshpment polcy n the next secton, and then compare ts cost to that of the new enhanced transshpment polcy n Secton 5, usng the same set of experments that were nvestgated n ths secton.

6 322 C. Paterson et al. / European Journal of Operatonal Research 221 (2012) Table 4 Addtonal notaton. d Tme quantum between each state transton f f;j Truncated probablty of demand j at locaton such that R Q < IL s always true Y Number of quanta untl the outstandng order arrves at locaton Z Locaton of most recent demand where Z = 0 ndcates no demand occurred n the precedng perod I H Indcator functon where H s a logcal statement: f true I H = 1 else I H =0 4. Dynamc programmng formulaton In ths secton, we provde a DP formulaton that can be used to fnd the optmal transshpment polcy usng value teraton. In our system, the resultng polcy determnes how much to transshp, gven the locatons nventory levels and remanng lead tmes of outstandng orders, n order to mnmze the overall cost rate ncurred. We remark that our DP formulaton utlzes a dscrete tme approxmaton of the actual contnuous revew system. However, for a suffcently small tme quantum the approxmaton s very good. Ths wll be verfed n the numercal nvestgaton of Secton 5. For presentatonal ease, we wll refer to the transshpment polcy that mnmzes the cost under the DP formulaton as the optmal transshpment polcy. Our formulaton s for a two locaton system setup. Whle t s possble to generalze the approach to larger systems, value teraton becomes computatonally ntractable very quckly due to the rapd growth n the number of states. In order to lmt the state space, only one order s allowed to be outstandng at each locaton at any tme. Ths assumpton s supported by real world practce and by the smulaton results obtaned n Secton 3. These showed that condtonal upon a locaton havng orders outstandng the probablty of multple orders outstandng was less than 1%. Our smulatons also showed that lmtng the enhanced polcy n a two locaton network so that a locaton could only ever have a maxmum stock of R + Q resulted n no statstcally sgnfcant cost dfference. The reader s referred to the related dscusson n Secton 2. Ths result s due to the fact that t s unlkely that one locaton would have enough excess stock to be able to completely refll the other s safety stock, and thus make IP > R + Q. Therefore t s feasble to lmt the nventory state space to a range of R ± Q.It should be noted that n a mult-locaton network, and n partcular wth non-dentcal locatons, ths assumpton wll not hold. Recall from Secton 3 that we wll apply the DP algorthm for the same set of experments as descrbed there. To develop the DP model we ntroduce addtonal notaton and formulae n Table 4. For suffcently small tme quantum d, we may assume that n a sngle tme slot the system experences an nstance of demand at ether one locaton or nether locaton. The probablty of a demand of sze j at locaton s dk f f;j durng each perod whle the probablty of no demand n the system s 1 d(k 1 + k 2 ). In the lmt d? 0 these probabltes converge to the exact Posson probabltes State defnton A fve dmensonal system state ncorporates the nventory level and the tme untl the outstandng order arrves (f there s one) at both locatons. The ffth dmenson ndcates the locaton where any current demand has occurred. We wrte state s as follows s ¼hIL 1 ; IL 2 ; Y 1 ; Y 2 ; Z ð4:1þ Recall that we do not allow the nventory poston to go above R + Q. Further, the decson to allow a maxmum of one outstandng order bounds the nventory level below. Hence we have that R Q < IL 6 R + Q. Moreover, we clearly have 0 6 Y j < L j d Acton space defnton The acton a ( a) s the amount to transshp from locatons 1(2) 2(1). For a gven state s, the set of actons Act(s) s bounded by zero, and by the mnmum of the amount avalable to transshp and the amount that can be stored at the recevng locaton (.e. that does not take the nventory poston above ts maxmum). We summarze these constrants by. If IL 2 >0,IL 1 < 0 and Z =1: a 2f mnðil 2 ; R 1 þ I ðy1 ¼0Þ Q 1 IL 1 Þ;...; 0g If IL 1 >0,IL 2 < 0 and Z =2: a 2f0;...; mnðil 1 ; R 2 þ I ðy2 ¼0Þ Q 2 IL 2 Þg Else: a 2 {0}. Hence, for example, f IL 1 >0,IL 2 <0,Z = 2 and Y 2 > 0 then a s an nteger n the range from 0 to mn(il 1,R 2 IL 2 ) Cost functon The cost (f s (a)) assocated wth beng n state s and choosng acton a s obtaned by aggregatng the cost of holdng (or backorderng) the current level of nventory after the transshppng acton has taken place wth the cost of the acton tself. There s also the addtonal cost of placng any replenshment order Table 5 Optmalty gap analyss. k p L Parameter value Reactve Enhanced Improvement Parameter value Reactve Enhanced Improvement Parameter value Reactve Enhanced b c f c u Improvement

7 C. Paterson et al. / European Journal of Operatonal Research 221 (2012) whch s ncluded n the perod drectly after the order has been nstgated. Hence we have f s ðaþ ¼h 1 ðil 1 aþ þ þ h 2 ðil 2 þ aþ þ þ b 1 ðil 1 aþ þ b 2 ðil 2 þ aþ þ A 1 I ðy1 ¼ L 1 d 1Þ þ A 2I ðy2 ¼ L 2 d 1Þ þ cu jajþc f I jaj> State transtons ð4:2þ If the current state s s = hil 1, IL 2, Y 1, Y 2, Z, acton a s undertaken and demand d occurs at locaton 1 then the new state wll be s 0 ¼hIL 0 1 ; IL0 2 ; Y0 1 ; Y 0 2, where IL 0 1 ¼ IL 1 a d þ Q 1 I ðy1 ¼1Þ; IL 0 2 ¼ IL 2 þ a þ Q 2 I ðy2 ¼1Þ; Y 0 1 ¼ðY 1 1ÞI þ L 1 ð16y1 6 L 1 d 1Þ d 1 Y 0 2 ¼ðY 2 1ÞI þ L 2 ð16y2 6 L 2 d 1Þ d 1 I ðil1 a d6r 1 &Y 1 ¼0Þ; I ðil2 þa6r 2 &Y 2 ¼0Þ: ð4:3þ Smlar transtons can be dentfed for the cases when a demand occurs at locaton 2 and when no demand occurs anywhere n the system Value teraton The above are deployed n a value teraton n whch the value functon (V n (s)) s the mnmal cost ncurred over an n-perod horzon from ntal state s. If we wrte w(s,s 0 ) for the probablty of movng from state s to s 0 then the optmalty prncple gves! V n ðsþ ¼ mn a2actðsþ f s ðaþþ X s 0 2Swðs; s 0 ÞV n 1 ðs 0 Þ : ð4:4þ We develop the (V n ) np1 usng backwards nducton and utlze the stoppng crteron recommended by Tjms (1986). The mnmzng actons n the fnal teraton yeld the -optmal polcy. 5. Optmalty gap For each of the 600 experments, the dynamc programmng model was used to determne the optmal transshpment polcy. Ths was then used wthn the smulaton model of Secton 3 along wth the two approxmate polces and the polcy of not transshppng. Ths ensured that lke for lke comparsons could be made usng the same set of randomly generated events. It was necessary to gather results for a range of values of the reorder pont R so that the optmal value could be used for each parameter set Results The beneft of usng the enhanced polcy can now be seen n the large step t takes towards the performance of the -optmal polcy. Table 5 shows that over a range of parameter values there s a consstent level of mprovement n cost performance acheved by ncludng a proactve stock rebalancng element to a reactve polcy. Compared to a polcy of no transshpment the enhanced polcy closes the suboptmalty gap by over 95% over the entre data set. Compared to the reactve polcy t s closed by 89%; t s ths relatve mprovement over the optmalty gap whch s broken down n Table 5. The enhanced polcy appears to perform strongly n comparson to the optmal polcy. In fact n 90 of the 600 experments there s no statstcal dfference between the two polces. It also dsplays very smlar characterstcs to the dynamc programmng based optmal polcy. Both ts average reorder pont (8.2 vs. 8.1) and average transshpment sze (both 4.7) are smlar. The one dfference s the number of transshpment events. For every 100 tme unts the -optmal polcy carres out one extra transshpment on average (6.4 vs. 7.5). Ths mples that the lack of future nformaton results n a slghtly more conservatve approach n the enhanced polcy. A subsecton of the full results can be found n Table A.2. Fg. 1 clearly demonstrates the superorty n performance of the enhanced polcy over the reactve polcy, more so for larger values of the fxed transshpment cost. It also gves an ndcaton of the greater consstency the enhanced polcy acheves n performance Accuracy of Dscrete Tme Assumpton & Computaton Tme The above results arsng from the DP mplementaton all use dscrete tme quantum d ¼ 1. We confrm the acceptablty of ths 8 choce by resolvng wth d ¼ 1 and observng that the resultng 16 changes n cost rates are mnmal and nowhere statstcally sgnfcant. Whle usng the optmal polcy may reduce nventory costs ts development s computatonally expensve when compared to the heurstc polces, whch can be obtaned very rapdly n real tme. The tme taken to compute each experments optmal polcy was recorded. Table 6 gves a breakdown of the tme needed to develop an optmal polcy by arrval rate. It dsplays how halvng the sze of d from d ¼ 1 ncreases the computatonal tme by a factor of 5 or Opt Gap 8 6 Opt Gap No Trans Reactve Enhanced Polcy No Trans Reactve Enhanced Polcy Fg. 1. Senstvty Analyss for c f = 10 (left) and c f = 50 (rght).

8 324 C. Paterson et al. / European Journal of Operatonal Research 221 (2012) Table 6 Computatonal tme (mns). Arrval rate d ¼ 1 8 d ¼ 1 16 Multple Overall more on average. Moreover, the computaton tme ncreases rapdly wth the arrval rate. Table 8 Percentage cost savngs for the enhanced polcy over the reactve polcy: large networks. Average k Identcal Two ter Large network study Table 7 Parameter values. Number of locatons 5, 10, 20 Arrval rate [Identcal Locatons](k) 1.4, 2.2, 3.0, 3.8, 4.6 Low arrval rate [Dfferent Locatons](k l ) 1.0, 1.8, 2.6, 3.4, 4.2 Hgh arrval rate [Dfferent Locatons](k h ) 2.0, 2.8, 3.6, 4.4, 5.2 Dstrbuton of order sze (p) 0.8 Lead tme (L) 3 Backorder cost (b) 30($) Transshpment cost [per tem] (c u ) 1($) Transshpment cost [per transshpment] (c f ) 10($) Fg. 2. Savngs acheved by the enhanced polcy over no transshpment. Havng shown that the enhanced polcy mproves the orgnal reactve polcy and makes a large step towards closng the optmalty gap whle reducng some of the varablty n performance, the next step s to consder ts performance n larger networks. In a smulaton study, desgned n a smlar way to the above small network study, nventory systems wth 5, 10 and 20 locaton are consdered. Rather than a full factoral study the focus s now put on how the sze of the network and the arrval rate (and hence the lead tme demand varablty) nfluences polcy performance. Eght dfferent arrval rates are consdered n networks wth dentcal locatons. Addtonally, networks whch had two dfferent levels of arrvals are consdered. These networks have 40% of the locatons at a hgher demand rate than the other 60%. In the latter case, the overall system arrval rate s set equal to a correspondng dentcal locaton confguraton so that comparsons could be farly drawn (see Table 7). A full set of results s gven n Tables A.3 and A.4 wthn the Appendx. The average percentage savngs, broken down by assgned values of the arrval rate, are gven n Table 8. The overall average results show a savng over the reactve polcy of 6.38% for dentcal locatons and 6.42% for the networks wth a two ter arrval rate structure. As for smaller networks, the greatest savngs occur when the arrval rate s large. The dfference n results between the dentcal locaton setup and the two ter setup s small. Ideally a system wth many dfferent arrval rates could be consdered but ths s challengng to mplement due to the necessty of determnng sutable values of R for each locaton va a post hoc optmzaton. The enhanced polcy offers a consstent level of cost mprovement. Average costs are reduced by between 11% and 17% n comparson wth no transshpment. In the case of the reactve polcy, the cost savng can be as lttle as 6%. The greater consstency n performance of the enhanced polcy aganst ncreasng arrval rate s llustrated by Fg. 2. Whle costs are an mportant part of nventory systems t s not the only performance measure of nterest. Servce levels wthn a system are also a key consderaton and the fll rates (the percentage of demand flled mmedately from stock on hand or va transshpment) were also recorded for the large network study. For the reactve polcy a servce level of 96.8% was acheved, but the enhanced polcy ncreased ths to 98.3%. The large network study also renforces other fndngs from the smaller network results. For dentcal locatons, safety stock s reduced from 12.6 unts on average wth no transshpments to 9.3 unts under the reactve polcy and to 5.6 unts under the new enhanced polcy. Smlar results are obtaned for the two ter networks. The average sze of transshpment agan ncreases, from 1.3 unts to 6.0 unts. These results llustrate the greater effcency possble from antcpatng shortages rather than merely respondng to them. 7. Conclusons We have shown that the benefts of reactve transshpments can be enhanced by the development of a new type of polcy, whch ncorporate a proactve element. System costs can be reduced and the effcency of the transshpment process mproved. Ths has been observed through an extensve study of both small and large (R, Q) replenshment polcy nventory networks, wth the benefts growng wth the number of stock holdng locatons. Moreover the mprovements that ths enhanced transshpment polcy can brng have been shown to sgnfcantly reduce the optmalty gap. The comparson to optmalty has been acheved through a dynamc programmng model that enables the calculaton of an - optmal transshpment polcy and the resultng costs. Whlst ths formulaton s restrcted to small systems t s an mportant step n understandng the transshpment process and n evaluatng the performance of the more easly developed enhanced polcy. More mportantly, the numercal results show that the enhanced polcy performs close to optmal over the examned parameters, wth some scenaros showng no sgnfcant dfference between the enhanced polcy and the -optmal polcy. One possble way to further mprove the enhanced polcy s to relax the myopc assumpton that underpns t. Another avenue would be to develop the redstrbuton element by consderng transshpments at tmes other than at those when shortages occur.

9 C. Paterson et al. / European Journal of Operatonal Research 221 (2012) One clear lmtaton wthn batch orderng systems wth transshpments s the challenge to fnd approprate replenshment polcy parameter values, even more so n systems wth non-dentcal locatons. Our results have shown that the enhanced transshpment polcy can sgnfcantly alter the optmal reorder pont when compared to no-transshpments. Indeed part of the savngs acheved s a consequence of beng able to lower the amount of safety stock requred throughout the system. Future work could develop analytcal approaches to the determnaton of reorder ponts. Ths would enable the full benefts of the mprovements n the transshpment polcy to be realzed n more complex nventory systems wth a larger number of stockng locatons and nondentcal demand rates. Appendx A. Detaled results Tables A.1 A.4. Table A.1 Two locaton results: p = 0.8, L =3,c u =1. k b c f Q R No tran. R Reactve (Err) Savng R Enhanced (Err) Savng Improve (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.03) (0.02) (0.02) (0.03) (0.02) (0.02) (0.03) (0.02) (0.02) (0.03) (0.02) (0.02) (0.03) (0.03) (0.02) (0.03) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.03) (0.02) (0.02) (0.03) (0.02) (0.02) (0.03) (0.02) (0.02) (0.03) (0.03) (0.02) (0.03) (0.02) (0.02) (0.03) (0.03) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.03) (0.02) (0.02) (0.03) (0.03) (0.02) (0.03) (0.03) (0.02) (0.04) (0.03) (0.03) (0.04) (0.03) (0.03) (0.04) (0.03) (0.02) Table A.2 Two locaton results: optmalty comparson p = 0.8, L =3,c u =1. k b c f Q Opt. No tran. Gap Reactve Gap Enhanced Gap

10 326 C. Paterson et al. / European Journal of Operatonal Research 221 (2012) Table A.3 Large network results: dentcal locatons. No. k Q R No tran. R Reactve Savng R Enhanced Savng Improve Table A.4 Large network results: two ter locatons. No. k Q R No tran. R Reactve Savng R Enhanced Savng Improve Appendx B. Calculatng a (X ) If t s s the tme when the sth unt becomes avalable at locaton then at each pont n tme a locaton s state can be descrbed by a varable X, where X ¼ðIP ; t IP ; t IP 1 ;...; t 1 ðil Þ Þ: ðb:1þ For locaton the pdf and cdf of the dstrbuton of the tme when the nth customer arrval nstant occurs can be respectvely gven as: For s 6 0, t s 6 L, x ðs; t s Þ¼b t s : For s 6 0, t s > L, x ðs; t s Þ¼b L : We now defne quantty U j ðtþ as ðb:6þ ðb:7þ g n ðtþ ¼kn tn 1 e k t ; ðn 1Þ! ðb:2þ G n Xn 1 ðk tþ k ðtþ ¼1 e kt : k! ðb:3þ k¼1 Usng these dstrbutons t s possble to obtan the pdf and cdf of the tme when the jth unt s demanded at locaton : r j ðtþ ¼Xj P n ;j gn ðtþ; n¼1 R j ðtþ ¼Xj P n ;j Gn ðtþ; n¼1 ðb:4þ ðb:5þ It s now possble to calculate a (X ). We let x (s,t s ) be the expected holdng and backorder costs, assocated wth the sth tem of stock demanded durng the lead tme L.Ifs 6 0 then the tem has already been demanded, wth s = 0 the most recently demanded tem. Ths gves four specfc cases. The frst two beng: U j ðtþ ¼ Z t 0 r j Xj ðuþudu ¼ P n ;j Gnþ1 ðtþ n : ðb:8þ k n¼1 Ths allows us to express the other two cases as: For s >0,t s 6 L, x ðs; t s Þ¼h For s >0,t s > L, Z L t s r s ðuþðu t sþdu þ 1 R s ðl Þ ðl t s Þ Z ts þ b r s ðuþðt s uþdu; 0 ¼ h U s ðl Þ U s ðt sþ t s ð1 R s ðt sþþ þl ð1 R s ðl ÞÞ þ b R s ðt sþt s U s ðt sþ : ðb:9þ

11 C. Paterson et al. / European Journal of Operatonal Research 221 (2012) x ðs; t s Þ¼b R s ðl ÞL U s ðl Þ : ðb:10þ We calculate a (X ) usng a ðx Þ¼ XIP s¼1 ðilþ x ðs; t s Þþ X 1 s¼ipþ1 x ðs; L Þ L C ; ðb:11þ where L C s the steady state cost ncurred durng the lead tme perod. The frst summaton calculates the expected cost over the perod L assocated wth each unt of nventory ether n stock, on order or already demanded. Smlarly the second summaton calculates the expected cost for any potental demand where the correspondng tem of stock has not yet been ordered from the suppler. The fnal term subtracts the steady state cost for the gven perod of tme so that overall a (X ) gves the bas over perod L. References Agrawal, V., Chao, X., Seshadr, S., Dynamc balancng of nventory n supply chans. European Journal of Operatonal Research 159, Allen, S.G., Redstrbuton of total stock over several user locatons. Naval Research Logstcs Quarterly 5 (4), Archbald, T.W., Sassen, S.A.E., Thomas, L.C., An optmal polcy for a two depot nventory problem wth stock transfer. Management Scence 43 (2), Archbald, T.W., Modellng replenshment and transshpment decsons n perodc revew multlocaton nventory systems. Journal of the Operatonal Research Socety 58 (7), Archbald, T.W., Black, D., Glazebrook, K.D., An ndex heurstc for transshpment decsons n mult-locaton nventory systems based on a parwse decomposton. European Journal of Operatonal Research 192, Archbald, T.W., Black, D., Glazebrook, K.D., The use of smple calbratons of ndvdual locatons n makng transshpment decsons n a mult-locaton nventory network. Journal of th Operatonal Research Socety 61 (2), Axsäter, S., Modellng emergency lateral transshpments n nventory systems. Management Scence 36 (11), Axsäter, S., A new decson rule for lateral transshpments n nventory systems. Management Scence 49 (9), Axsäter, S., Howard, C., Marklund, J., A Dstrbuton Inventory Model wth Transshpments from a Support Warehouse. Workng Paper, Lund Unversty. Bertrand, L.P., Bookbnder, J.H., Stock redstrbuton n two-echelon logstcs system. Journal of the Operatonal Research Socety 49, Dada, M., A two-echelon nventory system wth prorty shpments. Management Scence 38 (8), Das, C., Supply and redstrbuton rules for two-locaton nventory systems: one perod analyss. Management Scence 21 (7), Dong, L., Rud, N., Who benefts from transshpment? Exogenous vs. endogenous wholesale prces. Management Scence 50 (5), Evers, P.T., Heurstcs for assessng emergency transshpments. European Journal of Operatonal Research 129 (2), Gross, D., Centralzed nventory control n multlocaton supply systems. In: Scarf, H.E., Glford, D.M., Shelly, M.W. (Eds.), Multstage Inventory Models and Technques. Stanford Unversty Press, Stanford, Calforna, pp Jung, B., Sun, B., Km, J., Ahn, S., Modelng lateral transshpments n multechelon reparable-tem nventory systems wth fnte repar channels. Computers and Operatons Research 30, Jönsson, H., Slver, E.A., Analyss of a two echelon nventory control system wth complete redstrbuton. Management Scence 33, Kranenburg, A.A., van Houtum, G.J., A new partal poolng structure for spare parts networks. European Journal of Operatonal Research 199, Krshnan, K., Rao, V., Inventory control n N warehouses. Journal of Industral Engneerng XVI (3), Kukreja, A., Schmdt, C.P., A model for lumpy demand parts n a mult-locaton nventory system wth transshpments. Computers and Operatons Research 32 (8), Lee, H.L., A mult-echelon nventory model for reparable tems wth emergency lateral transshpments. Management Scence 33 (10), Mnner, S., Slver, E.A., Robb, D.J., An mproved heurstc for decdng on emergency transshpments. European Journal of Operatonal Research 148, Paterson, C., Kesmüller, G., Teunter, R., Glazebrook, K., Inventory models wth lateral transshpments: a revew. European Journal of Operatonal Research 210 (2), Robnson, L.W., Optmal an approxmate polces n multperod multlocaton nventory models wth transshpments. Operatons Research 38 (2), Rud, N., Kapur, S., Pyke, D.F., A two-locaton nventory model wth transshpment and local decson makng. Management Scence 47 (12), Tjms, H.C., Stochastc Modellng and Analyss: A Computatonal Approach. Wley, Chchester, New York. Sherbrooke, C.C., METRIC: a mult-echelon technque for recoverable tem control. Operatons Research 16, Sherbrooke, C.C., VARI-METRIC: mproved approxmatons for multndenture, mult-echelon avalablty models. Operatons Research 34 (2), Zhao, H., Ryan, J.K., Deshpande, V., Optmal dynamc producton and nventory transshpment polces for a two-locaton make-to-stock system. Operatons Research 56 (2), Wjk, A.C.C., Adan, I.J.B.F., van Houtum, G.J., Optmal Lateral Transshpment Polcy for a Two Locaton Inventory Problem. Workng Paper, EURANDOM. Wong, H., van Houtum, G.J., Cattrysse, D., van Oudheusden, D., 2006b. Mult-tem spare parts systems wth lateral transshpments and watng tme constrants. European Journal of Operatonal Research 171,

Power-of-Two Policies for Single- Warehouse Multi-Retailer Inventory Systems with Order Frequency Discounts

Power-of-Two Policies for Single- Warehouse Multi-Retailer Inventory Systems with Order Frequency Discounts Power-of-wo Polces for Sngle- Warehouse Mult-Retaler Inventory Systems wth Order Frequency Dscounts José A. Ventura Pennsylvana State Unversty (USA) Yale. Herer echnon Israel Insttute of echnology (Israel)

More information

In some supply chains, materials are ordered periodically according to local information. This paper investigates

In some supply chains, materials are ordered periodically according to local information. This paper investigates MANUFACTURING & SRVIC OPRATIONS MANAGMNT Vol. 12, No. 3, Summer 2010, pp. 430 448 ssn 1523-4614 essn 1526-5498 10 1203 0430 nforms do 10.1287/msom.1090.0277 2010 INFORMS Improvng Supply Chan Performance:

More information

An Alternative Way to Measure Private Equity Performance

An Alternative Way to Measure Private Equity Performance An Alternatve Way to Measure Prvate Equty Performance Peter Todd Parlux Investment Technology LLC Summary Internal Rate of Return (IRR) s probably the most common way to measure the performance of prvate

More information

Retailers must constantly strive for excellence in operations; extremely narrow profit margins

Retailers must constantly strive for excellence in operations; extremely narrow profit margins Managng a Retaler s Shelf Space, Inventory, and Transportaton Gerard Cachon 300 SH/DH, The Wharton School, Unversty of Pennsylvana, Phladelpha, Pennsylvana 90 cachon@wharton.upenn.edu http://opm.wharton.upenn.edu/cachon/

More information

benefit is 2, paid if the policyholder dies within the year, and probability of death within the year is ).

benefit is 2, paid if the policyholder dies within the year, and probability of death within the year is ). REVIEW OF RISK MANAGEMENT CONCEPTS LOSS DISTRIBUTIONS AND INSURANCE Loss and nsurance: When someone s subject to the rsk of ncurrng a fnancal loss, the loss s generally modeled usng a random varable or

More information

Project Networks With Mixed-Time Constraints

Project Networks With Mixed-Time Constraints Project Networs Wth Mxed-Tme Constrants L Caccetta and B Wattananon Western Australan Centre of Excellence n Industral Optmsaton (WACEIO) Curtn Unversty of Technology GPO Box U1987 Perth Western Australa

More information

2008/8. An integrated model for warehouse and inventory planning. Géraldine Strack and Yves Pochet

2008/8. An integrated model for warehouse and inventory planning. Géraldine Strack and Yves Pochet 2008/8 An ntegrated model for warehouse and nventory plannng Géraldne Strack and Yves Pochet CORE Voe du Roman Pays 34 B-1348 Louvan-la-Neuve, Belgum. Tel (32 10) 47 43 04 Fax (32 10) 47 43 01 E-mal: corestat-lbrary@uclouvan.be

More information

Outsourcing inventory management decisions in healthcare: Models and application

Outsourcing inventory management decisions in healthcare: Models and application European Journal of Operatonal Research 154 (24) 271 29 O.R. Applcatons Outsourcng nventory management decsons n healthcare: Models and applcaton www.elsever.com/locate/dsw Lawrence Ncholson a, Asoo J.

More information

The OC Curve of Attribute Acceptance Plans

The OC Curve of Attribute Acceptance Plans The OC Curve of Attrbute Acceptance Plans The Operatng Characterstc (OC) curve descrbes the probablty of acceptng a lot as a functon of the lot s qualty. Fgure 1 shows a typcal OC Curve. 10 8 6 4 1 3 4

More information

On the Optimal Control of a Cascade of Hydro-Electric Power Stations

On the Optimal Control of a Cascade of Hydro-Electric Power Stations On the Optmal Control of a Cascade of Hydro-Electrc Power Statons M.C.M. Guedes a, A.F. Rbero a, G.V. Smrnov b and S. Vlela c a Department of Mathematcs, School of Scences, Unversty of Porto, Portugal;

More information

Institute of Informatics, Faculty of Business and Management, Brno University of Technology,Czech Republic

Institute of Informatics, Faculty of Business and Management, Brno University of Technology,Czech Republic Lagrange Multplers as Quanttatve Indcators n Economcs Ivan Mezník Insttute of Informatcs, Faculty of Busness and Management, Brno Unversty of TechnologCzech Republc Abstract The quanttatve role of Lagrange

More information

J. Parallel Distrib. Comput.

J. Parallel Distrib. Comput. J. Parallel Dstrb. Comput. 71 (2011) 62 76 Contents lsts avalable at ScenceDrect J. Parallel Dstrb. Comput. journal homepage: www.elsever.com/locate/jpdc Optmzng server placement n dstrbuted systems n

More information

The Development of Web Log Mining Based on Improve-K-Means Clustering Analysis

The Development of Web Log Mining Based on Improve-K-Means Clustering Analysis The Development of Web Log Mnng Based on Improve-K-Means Clusterng Analyss TngZhong Wang * College of Informaton Technology, Luoyang Normal Unversty, Luoyang, 471022, Chna wangtngzhong2@sna.cn Abstract.

More information

Multiple-Period Attribution: Residuals and Compounding

Multiple-Period Attribution: Residuals and Compounding Multple-Perod Attrbuton: Resduals and Compoundng Our revewer gave these authors full marks for dealng wth an ssue that performance measurers and vendors often regard as propretary nformaton. In 1994, Dens

More information

A DYNAMIC CRASHING METHOD FOR PROJECT MANAGEMENT USING SIMULATION-BASED OPTIMIZATION. Michael E. Kuhl Radhamés A. Tolentino-Peña

A DYNAMIC CRASHING METHOD FOR PROJECT MANAGEMENT USING SIMULATION-BASED OPTIMIZATION. Michael E. Kuhl Radhamés A. Tolentino-Peña Proceedngs of the 2008 Wnter Smulaton Conference S. J. Mason, R. R. Hll, L. Mönch, O. Rose, T. Jefferson, J. W. Fowler eds. A DYNAMIC CRASHING METHOD FOR PROJECT MANAGEMENT USING SIMULATION-BASED OPTIMIZATION

More information

Module 2 LOSSLESS IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur

Module 2 LOSSLESS IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur Module LOSSLESS IMAGE COMPRESSION SYSTEMS Lesson 3 Lossless Compresson: Huffman Codng Instructonal Objectves At the end of ths lesson, the students should be able to:. Defne and measure source entropy..

More information

The Stochastic Guaranteed Service Model with Recourse for Multi-Echelon Warehouse Management

The Stochastic Guaranteed Service Model with Recourse for Multi-Echelon Warehouse Management The Stochastc Guaranteed Servce Model wth Recourse for Mult-Echelon Warehouse Management Jörg Rambau, Konrad Schade 1 Lehrstuhl für Wrtschaftsmathematk Unverstät Bayreuth Bayreuth, Germany Abstract The

More information

Answer: A). There is a flatter IS curve in the high MPC economy. Original LM LM after increase in M. IS curve for low MPC economy

Answer: A). There is a flatter IS curve in the high MPC economy. Original LM LM after increase in M. IS curve for low MPC economy 4.02 Quz Solutons Fall 2004 Multple-Choce Questons (30/00 ponts) Please, crcle the correct answer for each of the followng 0 multple-choce questons. For each queston, only one of the answers s correct.

More information

Recurrence. 1 Definitions and main statements

Recurrence. 1 Definitions and main statements Recurrence 1 Defntons and man statements Let X n, n = 0, 1, 2,... be a MC wth the state space S = (1, 2,...), transton probabltes p j = P {X n+1 = j X n = }, and the transton matrx P = (p j ),j S def.

More information

Number of Levels Cumulative Annual operating Income per year construction costs costs ($) ($) ($) 1 600,000 35,000 100,000 2 2,200,000 60,000 350,000

Number of Levels Cumulative Annual operating Income per year construction costs costs ($) ($) ($) 1 600,000 35,000 100,000 2 2,200,000 60,000 350,000 Problem Set 5 Solutons 1 MIT s consderng buldng a new car park near Kendall Square. o unversty funds are avalable (overhead rates are under pressure and the new faclty would have to pay for tself from

More information

DEFINING %COMPLETE IN MICROSOFT PROJECT

DEFINING %COMPLETE IN MICROSOFT PROJECT CelersSystems DEFINING %COMPLETE IN MICROSOFT PROJECT PREPARED BY James E Aksel, PMP, PMI-SP, MVP For Addtonal Informaton about Earned Value Management Systems and reportng, please contact: CelersSystems,

More information

THE DISTRIBUTION OF LOAN PORTFOLIO VALUE * Oldrich Alfons Vasicek

THE DISTRIBUTION OF LOAN PORTFOLIO VALUE * Oldrich Alfons Vasicek HE DISRIBUION OF LOAN PORFOLIO VALUE * Oldrch Alfons Vascek he amount of captal necessary to support a portfolo of debt securtes depends on the probablty dstrbuton of the portfolo loss. Consder a portfolo

More information

CHOLESTEROL REFERENCE METHOD LABORATORY NETWORK. Sample Stability Protocol

CHOLESTEROL REFERENCE METHOD LABORATORY NETWORK. Sample Stability Protocol CHOLESTEROL REFERENCE METHOD LABORATORY NETWORK Sample Stablty Protocol Background The Cholesterol Reference Method Laboratory Network (CRMLN) developed certfcaton protocols for total cholesterol, HDL

More information

Robust Design of Public Storage Warehouses. Yeming (Yale) Gong EMLYON Business School

Robust Design of Public Storage Warehouses. Yeming (Yale) Gong EMLYON Business School Robust Desgn of Publc Storage Warehouses Yemng (Yale) Gong EMLYON Busness School Rene de Koster Rotterdam school of management, Erasmus Unversty Abstract We apply robust optmzaton and revenue management

More information

Managing Cycle Inventories. Matching Supply and Demand

Managing Cycle Inventories. Matching Supply and Demand Managng Cycle Inventores Matchng Supply and Demand 1 Outlne Why to hold cycle nventores? Economes of scale to reduce fxed costs per unt. Jont fxed costs for multple products Long term quantty dscounts

More information

1. Fundamentals of probability theory 2. Emergence of communication traffic 3. Stochastic & Markovian Processes (SP & MP)

1. Fundamentals of probability theory 2. Emergence of communication traffic 3. Stochastic & Markovian Processes (SP & MP) 6.3 / -- Communcaton Networks II (Görg) SS20 -- www.comnets.un-bremen.de Communcaton Networks II Contents. Fundamentals of probablty theory 2. Emergence of communcaton traffc 3. Stochastc & Markovan Processes

More information

Can Auto Liability Insurance Purchases Signal Risk Attitude?

Can Auto Liability Insurance Purchases Signal Risk Attitude? Internatonal Journal of Busness and Economcs, 2011, Vol. 10, No. 2, 159-164 Can Auto Lablty Insurance Purchases Sgnal Rsk Atttude? Chu-Shu L Department of Internatonal Busness, Asa Unversty, Tawan Sheng-Chang

More information

Dynamic Pricing for Smart Grid with Reinforcement Learning

Dynamic Pricing for Smart Grid with Reinforcement Learning Dynamc Prcng for Smart Grd wth Renforcement Learnng Byung-Gook Km, Yu Zhang, Mhaela van der Schaar, and Jang-Won Lee Samsung Electroncs, Suwon, Korea Department of Electrcal Engneerng, UCLA, Los Angeles,

More information

Brigid Mullany, Ph.D University of North Carolina, Charlotte

Brigid Mullany, Ph.D University of North Carolina, Charlotte Evaluaton And Comparson Of The Dfferent Standards Used To Defne The Postonal Accuracy And Repeatablty Of Numercally Controlled Machnng Center Axes Brgd Mullany, Ph.D Unversty of North Carolna, Charlotte

More information

Omega 39 (2011) 313 322. Contents lists available at ScienceDirect. Omega. journal homepage: www.elsevier.com/locate/omega

Omega 39 (2011) 313 322. Contents lists available at ScienceDirect. Omega. journal homepage: www.elsevier.com/locate/omega Omega 39 (2011) 313 322 Contents lsts avalable at ScenceDrect Omega journal homepage: www.elsever.com/locate/omega Supply chan confguraton for dffuson of new products: An ntegrated optmzaton approach Mehd

More information

An MILP model for planning of batch plants operating in a campaign-mode

An MILP model for planning of batch plants operating in a campaign-mode An MILP model for plannng of batch plants operatng n a campagn-mode Yanna Fumero Insttuto de Desarrollo y Dseño CONICET UTN yfumero@santafe-concet.gov.ar Gabrela Corsano Insttuto de Desarrollo y Dseño

More information

Analysis of Premium Liabilities for Australian Lines of Business

Analysis of Premium Liabilities for Australian Lines of Business Summary of Analyss of Premum Labltes for Australan Lnes of Busness Emly Tao Honours Research Paper, The Unversty of Melbourne Emly Tao Acknowledgements I am grateful to the Australan Prudental Regulaton

More information

Awell-known result in the Bayesian inventory management literature is: If lost sales are not observed, the

Awell-known result in the Bayesian inventory management literature is: If lost sales are not observed, the MANUFACTURING & SERVICE OPERATIONS MANAGEMENT Vol. 10, No. 2, Sprng 2008, pp. 236 256 ssn 1523-4614 essn 1526-5498 08 1002 0236 nforms do 10.1287/msom.1070.0165 2008 INFORMS Dynamc Inventory Management

More information

Chapter 4 ECONOMIC DISPATCH AND UNIT COMMITMENT

Chapter 4 ECONOMIC DISPATCH AND UNIT COMMITMENT Chapter 4 ECOOMIC DISATCH AD UIT COMMITMET ITRODUCTIO A power system has several power plants. Each power plant has several generatng unts. At any pont of tme, the total load n the system s met by the

More information

8.5 UNITARY AND HERMITIAN MATRICES. The conjugate transpose of a complex matrix A, denoted by A*, is given by

8.5 UNITARY AND HERMITIAN MATRICES. The conjugate transpose of a complex matrix A, denoted by A*, is given by 6 CHAPTER 8 COMPLEX VECTOR SPACES 5. Fnd the kernel of the lnear transformaton gven n Exercse 5. In Exercses 55 and 56, fnd the mage of v, for the ndcated composton, where and are gven by the followng

More information

Staff Paper. Farm Savings Accounts: Examining Income Variability, Eligibility, and Benefits. Brent Gloy, Eddy LaDue, and Charles Cuykendall

Staff Paper. Farm Savings Accounts: Examining Income Variability, Eligibility, and Benefits. Brent Gloy, Eddy LaDue, and Charles Cuykendall SP 2005-02 August 2005 Staff Paper Department of Appled Economcs and Management Cornell Unversty, Ithaca, New York 14853-7801 USA Farm Savngs Accounts: Examnng Income Varablty, Elgblty, and Benefts Brent

More information

Simulation and optimization of supply chains: alternative or complementary approaches?

Simulation and optimization of supply chains: alternative or complementary approaches? Smulaton and optmzaton of supply chans: alternatve or complementary approaches? Chrstan Almeder Margaretha Preusser Rchard F. Hartl Orgnally publshed n: OR Spectrum (2009) 31:95 119 DOI 10.1007/s00291-007-0118-z

More information

Financial Mathemetics

Financial Mathemetics Fnancal Mathemetcs 15 Mathematcs Grade 12 Teacher Gude Fnancal Maths Seres Overvew In ths seres we am to show how Mathematcs can be used to support personal fnancal decsons. In ths seres we jon Tebogo,

More information

What is Candidate Sampling

What is Candidate Sampling What s Canddate Samplng Say we have a multclass or mult label problem where each tranng example ( x, T ) conssts of a context x a small (mult)set of target classes T out of a large unverse L of possble

More information

Enabling P2P One-view Multi-party Video Conferencing

Enabling P2P One-view Multi-party Video Conferencing Enablng P2P One-vew Mult-party Vdeo Conferencng Yongxang Zhao, Yong Lu, Changja Chen, and JanYn Zhang Abstract Mult-Party Vdeo Conferencng (MPVC) facltates realtme group nteracton between users. Whle P2P

More information

Preventive Maintenance and Replacement Scheduling: Models and Algorithms

Preventive Maintenance and Replacement Scheduling: Models and Algorithms Preventve Mantenance and Replacement Schedulng: Models and Algorthms By Kamran S. Moghaddam B.S. Unversty of Tehran 200 M.S. Tehran Polytechnc 2003 A Dssertaton Proposal Submtted to the Faculty of the

More information

A hybrid global optimization algorithm based on parallel chaos optimization and outlook algorithm

A hybrid global optimization algorithm based on parallel chaos optimization and outlook algorithm Avalable onlne www.ocpr.com Journal of Chemcal and Pharmaceutcal Research, 2014, 6(7):1884-1889 Research Artcle ISSN : 0975-7384 CODEN(USA) : JCPRC5 A hybrd global optmzaton algorthm based on parallel

More information

ANALYZING THE RELATIONSHIPS BETWEEN QUALITY, TIME, AND COST IN PROJECT MANAGEMENT DECISION MAKING

ANALYZING THE RELATIONSHIPS BETWEEN QUALITY, TIME, AND COST IN PROJECT MANAGEMENT DECISION MAKING ANALYZING THE RELATIONSHIPS BETWEEN QUALITY, TIME, AND COST IN PROJECT MANAGEMENT DECISION MAKING Matthew J. Lberatore, Department of Management and Operatons, Vllanova Unversty, Vllanova, PA 19085, 610-519-4390,

More information

How Sets of Coherent Probabilities May Serve as Models for Degrees of Incoherence

How Sets of Coherent Probabilities May Serve as Models for Degrees of Incoherence 1 st Internatonal Symposum on Imprecse Probabltes and Ther Applcatons, Ghent, Belgum, 29 June 2 July 1999 How Sets of Coherent Probabltes May Serve as Models for Degrees of Incoherence Mar J. Schervsh

More information

Many e-tailers providing attended home delivery, especially e-grocers, offer narrow delivery time slots to

Many e-tailers providing attended home delivery, especially e-grocers, offer narrow delivery time slots to Vol. 45, No. 3, August 2011, pp. 435 449 ssn 0041-1655 essn 1526-5447 11 4503 0435 do 10.1287/trsc.1100.0346 2011 INFORMS Tme Slot Management n Attended Home Delvery Nels Agatz Department of Decson and

More information

Trade Adjustment and Productivity in Large Crises. Online Appendix May 2013. Appendix A: Derivation of Equations for Productivity

Trade Adjustment and Productivity in Large Crises. Online Appendix May 2013. Appendix A: Derivation of Equations for Productivity Trade Adjustment Productvty n Large Crses Gta Gopnath Department of Economcs Harvard Unversty NBER Brent Neman Booth School of Busness Unversty of Chcago NBER Onlne Appendx May 2013 Appendx A: Dervaton

More information

Data Broadcast on a Multi-System Heterogeneous Overlayed Wireless Network *

Data Broadcast on a Multi-System Heterogeneous Overlayed Wireless Network * JOURNAL OF INFORMATION SCIENCE AND ENGINEERING 24, 819-840 (2008) Data Broadcast on a Mult-System Heterogeneous Overlayed Wreless Network * Department of Computer Scence Natonal Chao Tung Unversty Hsnchu,

More information

How To Solve An Onlne Control Polcy On A Vrtualzed Data Center

How To Solve An Onlne Control Polcy On A Vrtualzed Data Center Dynamc Resource Allocaton and Power Management n Vrtualzed Data Centers Rahul Urgaonkar, Ulas C. Kozat, Ken Igarash, Mchael J. Neely urgaonka@usc.edu, {kozat, garash}@docomolabs-usa.com, mjneely@usc.edu

More information

Credit Limit Optimization (CLO) for Credit Cards

Credit Limit Optimization (CLO) for Credit Cards Credt Lmt Optmzaton (CLO) for Credt Cards Vay S. Desa CSCC IX, Ednburgh September 8, 2005 Copyrght 2003, SAS Insttute Inc. All rghts reserved. SAS Propretary Agenda Background Tradtonal approaches to credt

More information

Efficient Striping Techniques for Variable Bit Rate Continuous Media File Servers æ

Efficient Striping Techniques for Variable Bit Rate Continuous Media File Servers æ Effcent Strpng Technques for Varable Bt Rate Contnuous Meda Fle Servers æ Prashant J. Shenoy Harrck M. Vn Department of Computer Scence, Department of Computer Scences, Unversty of Massachusetts at Amherst

More information

AN APPOINTMENT ORDER OUTPATIENT SCHEDULING SYSTEM THAT IMPROVES OUTPATIENT EXPERIENCE

AN APPOINTMENT ORDER OUTPATIENT SCHEDULING SYSTEM THAT IMPROVES OUTPATIENT EXPERIENCE AN APPOINTMENT ORDER OUTPATIENT SCHEDULING SYSTEM THAT IMPROVES OUTPATIENT EXPERIENCE Yu-L Huang Industral Engneerng Department New Mexco State Unversty Las Cruces, New Mexco 88003, U.S.A. Abstract Patent

More information

"Research Note" APPLICATION OF CHARGE SIMULATION METHOD TO ELECTRIC FIELD CALCULATION IN THE POWER CABLES *

Research Note APPLICATION OF CHARGE SIMULATION METHOD TO ELECTRIC FIELD CALCULATION IN THE POWER CABLES * Iranan Journal of Scence & Technology, Transacton B, Engneerng, ol. 30, No. B6, 789-794 rnted n The Islamc Republc of Iran, 006 Shraz Unversty "Research Note" ALICATION OF CHARGE SIMULATION METHOD TO ELECTRIC

More information

J. Parallel Distrib. Comput. Environment-conscious scheduling of HPC applications on distributed Cloud-oriented data centers

J. Parallel Distrib. Comput. Environment-conscious scheduling of HPC applications on distributed Cloud-oriented data centers J. Parallel Dstrb. Comput. 71 (2011) 732 749 Contents lsts avalable at ScenceDrect J. Parallel Dstrb. Comput. ournal homepage: www.elsever.com/locate/pdc Envronment-conscous schedulng of HPC applcatons

More information

行 政 院 國 家 科 學 委 員 會 補 助 專 題 研 究 計 畫 成 果 報 告 期 中 進 度 報 告

行 政 院 國 家 科 學 委 員 會 補 助 專 題 研 究 計 畫 成 果 報 告 期 中 進 度 報 告 行 政 院 國 家 科 學 委 員 會 補 助 專 題 研 究 計 畫 成 果 報 告 期 中 進 度 報 告 畫 類 別 : 個 別 型 計 畫 半 導 體 產 業 大 型 廠 房 之 設 施 規 劃 計 畫 編 號 :NSC 96-2628-E-009-026-MY3 執 行 期 間 : 2007 年 8 月 1 日 至 2010 年 7 月 31 日 計 畫 主 持 人 : 巫 木 誠 共 同

More information

The Retail Planning Problem Under Demand Uncertainty

The Retail Planning Problem Under Demand Uncertainty Vol., No. 5, September October 013, pp. 100 113 ISSN 1059-1478 EISSN 1937-5956 13 05 100 DOI 10.1111/j.1937-5956.01.0144.x 013 Producton and Operatons Management Socety The Retal Plannng Problem Under

More information

This paper concerns the evaluation and analysis of order

This paper concerns the evaluation and analysis of order ORDER-FULFILLMENT PERFORMANCE MEASURES IN AN ASSEMBLE- TO-ORDER SYSTEM WITH STOCHASTIC LEADTIMES JING-SHENG SONG Unversty of Calforna, Irvne, Calforna SUSAN H. XU Penn State Unversty, Unversty Park, Pennsylvana

More information

BERNSTEIN POLYNOMIALS

BERNSTEIN POLYNOMIALS On-Lne Geometrc Modelng Notes BERNSTEIN POLYNOMIALS Kenneth I. Joy Vsualzaton and Graphcs Research Group Department of Computer Scence Unversty of Calforna, Davs Overvew Polynomals are ncredbly useful

More information

2. SYSTEM MODEL. the SLA (unlike the only other related mechanism [15] we can compare it is never able to meet the SLA).

2. SYSTEM MODEL. the SLA (unlike the only other related mechanism [15] we can compare it is never able to meet the SLA). Managng Server Energy and Operatonal Costs n Hostng Centers Yyu Chen Dept. of IE Penn State Unversty Unversty Park, PA 16802 yzc107@psu.edu Anand Svasubramanam Dept. of CSE Penn State Unversty Unversty

More information

Traffic-light a stress test for life insurance provisions

Traffic-light a stress test for life insurance provisions MEMORANDUM Date 006-09-7 Authors Bengt von Bahr, Göran Ronge Traffc-lght a stress test for lfe nsurance provsons Fnansnspetonen P.O. Box 6750 SE-113 85 Stocholm [Sveavägen 167] Tel +46 8 787 80 00 Fax

More information

APPLICATION OF COMPUTER PROGRAMMING IN OPTIMIZATION OF TECHNOLOGICAL OBJECTIVES OF COLD ROLLING

APPLICATION OF COMPUTER PROGRAMMING IN OPTIMIZATION OF TECHNOLOGICAL OBJECTIVES OF COLD ROLLING Journal Journal of Chemcal of Chemcal Technology and and Metallurgy, 50, 6, 50, 2015, 6, 2015 638-643 APPLICATION OF COMPUTER PROGRAMMING IN OPTIMIZATION OF TECHNOLOGICAL OBJECTIVES OF COLD ROLLING Abdrakhman

More information

Optimal Customized Pricing in Competitive Settings

Optimal Customized Pricing in Competitive Settings Optmal Customzed Prcng n Compettve Settngs Vshal Agrawal Industral & Systems Engneerng, Georga Insttute of Technology, Atlanta, Georga 30332 vshalagrawal@gatech.edu Mark Ferguson College of Management,

More information

Logical Development Of Vogel s Approximation Method (LD-VAM): An Approach To Find Basic Feasible Solution Of Transportation Problem

Logical Development Of Vogel s Approximation Method (LD-VAM): An Approach To Find Basic Feasible Solution Of Transportation Problem INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME, ISSUE, FEBRUARY ISSN 77-866 Logcal Development Of Vogel s Approxmaton Method (LD- An Approach To Fnd Basc Feasble Soluton Of Transportaton

More information

Extending Probabilistic Dynamic Epistemic Logic

Extending Probabilistic Dynamic Epistemic Logic Extendng Probablstc Dynamc Epstemc Logc Joshua Sack May 29, 2008 Probablty Space Defnton A probablty space s a tuple (S, A, µ), where 1 S s a set called the sample space. 2 A P(S) s a σ-algebra: a set

More information

Calculating the high frequency transmission line parameters of power cables

Calculating the high frequency transmission line parameters of power cables < ' Calculatng the hgh frequency transmsson lne parameters of power cables Authors: Dr. John Dcknson, Laboratory Servces Manager, N 0 RW E B Communcatons Mr. Peter J. Ncholson, Project Assgnment Manager,

More information

Lecture 3: Annuity. Study annuities whose payments form a geometric progression or a arithmetic progression.

Lecture 3: Annuity. Study annuities whose payments form a geometric progression or a arithmetic progression. Lecture 3: Annuty Goals: Learn contnuous annuty and perpetuty. Study annutes whose payments form a geometrc progresson or a arthmetc progresson. Dscuss yeld rates. Introduce Amortzaton Suggested Textbook

More information

Revenue Management for a Multiclass Single-Server Queue via a Fluid Model Analysis

Revenue Management for a Multiclass Single-Server Queue via a Fluid Model Analysis OPERATIONS RESEARCH Vol. 54, No. 5, September October 6, pp. 94 93 ssn 3-364X essn 56-5463 6 545 94 nforms do.87/opre.6.35 6 INFORMS Revenue Management for a Multclass Sngle-Server Queue va a Flud Model

More information

SUPPLIER FINANCING AND STOCK MANAGEMENT. A JOINT VIEW.

SUPPLIER FINANCING AND STOCK MANAGEMENT. A JOINT VIEW. SUPPLIER FINANCING AND STOCK MANAGEMENT. A JOINT VIEW. Lucía Isabel García Cebrán Departamento de Economía y Dreccón de Empresas Unversdad de Zaragoza Gran Vía, 2 50.005 Zaragoza (Span) Phone: 976-76-10-00

More information

Joint Scheduling of Processing and Shuffle Phases in MapReduce Systems

Joint Scheduling of Processing and Shuffle Phases in MapReduce Systems Jont Schedulng of Processng and Shuffle Phases n MapReduce Systems Fangfe Chen, Mural Kodalam, T. V. Lakshman Department of Computer Scence and Engneerng, The Penn State Unversty Bell Laboratores, Alcatel-Lucent

More information

Calculation of Sampling Weights

Calculation of Sampling Weights Perre Foy Statstcs Canada 4 Calculaton of Samplng Weghts 4.1 OVERVIEW The basc sample desgn used n TIMSS Populatons 1 and 2 was a two-stage stratfed cluster desgn. 1 The frst stage conssted of a sample

More information

ASSESSING THE AVAILABILITY AND ALLOCATION OF PRODUCTION CAPACITY IN A FABRICATION FACILITY THROUGH SIMULATION MODELING: A CASE STUDY

ASSESSING THE AVAILABILITY AND ALLOCATION OF PRODUCTION CAPACITY IN A FABRICATION FACILITY THROUGH SIMULATION MODELING: A CASE STUDY Internatonal Journal of Industral Engneerng, 15(2), 166-175, 2008. ASSESSING THE AVAILABILITY AND ALLOCATION OF PRODUCTION CAPACITY IN A FABRICATION FACILITY THROUGH SIMULATION MODELING: A CASE STUDY J.H.

More information

A Lyapunov Optimization Approach to Repeated Stochastic Games

A Lyapunov Optimization Approach to Repeated Stochastic Games PROC. ALLERTON CONFERENCE ON COMMUNICATION, CONTROL, AND COMPUTING, OCT. 2013 1 A Lyapunov Optmzaton Approach to Repeated Stochastc Games Mchael J. Neely Unversty of Southern Calforna http://www-bcf.usc.edu/

More information

The Current Employment Statistics (CES) survey,

The Current Employment Statistics (CES) survey, Busness Brths and Deaths Impact of busness brths and deaths n the payroll survey The CES probablty-based sample redesgn accounts for most busness brth employment through the mputaton of busness deaths,

More information

Optimal resource capacity management for stochastic networks

Optimal resource capacity management for stochastic networks Submtted for publcaton. Optmal resource capacty management for stochastc networks A.B. Deker H. Mlton Stewart School of ISyE, Georga Insttute of Technology, Atlanta, GA 30332, ton.deker@sye.gatech.edu

More information

The Greedy Method. Introduction. 0/1 Knapsack Problem

The Greedy Method. Introduction. 0/1 Knapsack Problem The Greedy Method Introducton We have completed data structures. We now are gong to look at algorthm desgn methods. Often we are lookng at optmzaton problems whose performance s exponental. For an optmzaton

More information

Internet companies extensively use the practice of drop-shipping, where the wholesaler stocks and owns the

Internet companies extensively use the practice of drop-shipping, where the wholesaler stocks and owns the MANAGEMENT SIENE Vol. 52, No. 6, June 26, pp. 844 864 ssn 25-199 essn 1526-551 6 526 844 nforms do 1.1287/mnsc.16.512 26 INFORMS Supply han hoce on the Internet Sergue Netessne The Wharton School, Unversty

More information

Allocating Collaborative Profit in Less-than-Truckload Carrier Alliance

Allocating Collaborative Profit in Less-than-Truckload Carrier Alliance J. Servce Scence & Management, 2010, 3: 143-149 do:10.4236/jssm.2010.31018 Publshed Onlne March 2010 (http://www.scrp.org/journal/jssm) 143 Allocatng Collaboratve Proft n Less-than-Truckload Carrer Allance

More information

Performance Analysis of Energy Consumption of Smartphone Running Mobile Hotspot Application

Performance Analysis of Energy Consumption of Smartphone Running Mobile Hotspot Application Internatonal Journal of mart Grd and lean Energy Performance Analyss of Energy onsumpton of martphone Runnng Moble Hotspot Applcaton Yun on hung a chool of Electronc Engneerng, oongsl Unversty, 511 angdo-dong,

More information

Information Sciences

Information Sciences Informaton Scences 0 (013) 45 441 Contents lsts avalable at ScVerse ScenceDrect Informaton Scences journal homepage: www.elsever.com/locate/ns A hybrd method of fuzzy smulaton and genetc algorthm to optmze

More information

Formulating & Solving Integer Problems Chapter 11 289

Formulating & Solving Integer Problems Chapter 11 289 Formulatng & Solvng Integer Problems Chapter 11 289 The Optonal Stop TSP If we drop the requrement that every stop must be vsted, we then get the optonal stop TSP. Ths mght correspond to a ob sequencng

More information

Joe Pimbley, unpublished, 2005. Yield Curve Calculations

Joe Pimbley, unpublished, 2005. Yield Curve Calculations Joe Pmbley, unpublshed, 005. Yeld Curve Calculatons Background: Everythng s dscount factors Yeld curve calculatons nclude valuaton of forward rate agreements (FRAs), swaps, nterest rate optons, and forward

More information

An Interest-Oriented Network Evolution Mechanism for Online Communities

An Interest-Oriented Network Evolution Mechanism for Online Communities An Interest-Orented Network Evoluton Mechansm for Onlne Communtes Cahong Sun and Xaopng Yang School of Informaton, Renmn Unversty of Chna, Bejng 100872, P.R. Chna {chsun,yang}@ruc.edu.cn Abstract. Onlne

More information

IMPACT ANALYSIS OF A CELLULAR PHONE

IMPACT ANALYSIS OF A CELLULAR PHONE 4 th ASA & μeta Internatonal Conference IMPACT AALYSIS OF A CELLULAR PHOE We Lu, 2 Hongy L Bejng FEAonlne Engneerng Co.,Ltd. Bejng, Chna ABSTRACT Drop test smulaton plays an mportant role n nvestgatng

More information

Lecture 3: Force of Interest, Real Interest Rate, Annuity

Lecture 3: Force of Interest, Real Interest Rate, Annuity Lecture 3: Force of Interest, Real Interest Rate, Annuty Goals: Study contnuous compoundng and force of nterest Dscuss real nterest rate Learn annuty-mmedate, and ts present value Study annuty-due, and

More information

Feasibility of Using Discriminate Pricing Schemes for Energy Trading in Smart Grid

Feasibility of Using Discriminate Pricing Schemes for Energy Trading in Smart Grid Feasblty of Usng Dscrmnate Prcng Schemes for Energy Tradng n Smart Grd Wayes Tushar, Chau Yuen, Bo Cha, Davd B. Smth, and H. Vncent Poor Sngapore Unversty of Technology and Desgn, Sngapore 138682. Emal:

More information

How To Calculate The Accountng Perod Of Nequalty

How To Calculate The Accountng Perod Of Nequalty Inequalty and The Accountng Perod Quentn Wodon and Shlomo Ytzha World Ban and Hebrew Unversty September Abstract Income nequalty typcally declnes wth the length of tme taen nto account for measurement.

More information

IDENTIFICATION AND CORRECTION OF A COMMON ERROR IN GENERAL ANNUITY CALCULATIONS

IDENTIFICATION AND CORRECTION OF A COMMON ERROR IN GENERAL ANNUITY CALCULATIONS IDENTIFICATION AND CORRECTION OF A COMMON ERROR IN GENERAL ANNUITY CALCULATIONS Chrs Deeley* Last revsed: September 22, 200 * Chrs Deeley s a Senor Lecturer n the School of Accountng, Charles Sturt Unversty,

More information

1 Example 1: Axis-aligned rectangles

1 Example 1: Axis-aligned rectangles COS 511: Theoretcal Machne Learnng Lecturer: Rob Schapre Lecture # 6 Scrbe: Aaron Schld February 21, 2013 Last class, we dscussed an analogue for Occam s Razor for nfnte hypothess spaces that, n conjuncton

More information

Lowering Outbound Shipping Costs in an Online Retail Environment by Making Better Fulfillment and Replenishment Decisions

Lowering Outbound Shipping Costs in an Online Retail Environment by Making Better Fulfillment and Replenishment Decisions Lowerng Outbound Shppng Costs n an Onlne Retal Envronment by Makng Better Fulfllment and Replenshment Decsons by Jason Andrew Acmovc B.S. Physcs, Yale Unversty (1999) Submtted to the Sloan School of Management

More information

Time Value of Money Module

Time Value of Money Module Tme Value of Money Module O BJECTIVES After readng ths Module, you wll be able to: Understand smple nterest and compound nterest. 2 Compute and use the future value of a sngle sum. 3 Compute and use the

More information

To manage leave, meeting institutional requirements and treating individual staff members fairly and consistently.

To manage leave, meeting institutional requirements and treating individual staff members fairly and consistently. Corporate Polces & Procedures Human Resources - Document CPP216 Leave Management Frst Produced: Current Verson: Past Revsons: Revew Cycle: Apples From: 09/09/09 26/10/12 09/09/09 3 years Immedately Authorsaton:

More information

Forecasting the Direction and Strength of Stock Market Movement

Forecasting the Direction and Strength of Stock Market Movement Forecastng the Drecton and Strength of Stock Market Movement Jngwe Chen Mng Chen Nan Ye cjngwe@stanford.edu mchen5@stanford.edu nanye@stanford.edu Abstract - Stock market s one of the most complcated systems

More information

DEVELOPMENT & IMPLEMENTATION OF SUPPLY CHAIN MANAGEMENT FOCUSING ON PROCUREMENT PROCESSES & SUPPLIERS 1.

DEVELOPMENT & IMPLEMENTATION OF SUPPLY CHAIN MANAGEMENT FOCUSING ON PROCUREMENT PROCESSES & SUPPLIERS 1. 1. Claudu V. KIFOR, 2. Amela BUCUR, 3. Muhammad Arsalan FAROOQ DEVELOPMENT & IMPLEMENTATION OF SUPPLY CHAIN MANAGEMENT FOCUSING ON PROCUREMENT PROCESSES & SUPPLIERS 1. LUCIAN BLAGA UNIVERSITY SIBIU, RESEARCH

More information

Dynamic Fleet Management for Cybercars

Dynamic Fleet Management for Cybercars Proceedngs of the IEEE ITSC 2006 2006 IEEE Intellgent Transportaton Systems Conference Toronto, Canada, September 17-20, 2006 TC7.5 Dynamc Fleet Management for Cybercars Fenghu. Wang, Mng. Yang, Ruqng.

More information

A Novel Methodology of Working Capital Management for Large. Public Constructions by Using Fuzzy S-curve Regression

A Novel Methodology of Working Capital Management for Large. Public Constructions by Using Fuzzy S-curve Regression Novel Methodology of Workng Captal Management for Large Publc Constructons by Usng Fuzzy S-curve Regresson Cheng-Wu Chen, Morrs H. L. Wang and Tng-Ya Hseh Department of Cvl Engneerng, Natonal Central Unversty,

More information

Marginal Revenue-Based Capacity Management Models and Benchmark 1

Marginal Revenue-Based Capacity Management Models and Benchmark 1 Margnal Revenue-Based Capacty Management Models and Benchmark 1 Qwen Wang 2 Guanghua School of Management, Pekng Unversty Sherry Xaoyun Sun 3 Ctgroup ABSTRACT To effcently meet customer requrements, a

More information

Efficient Project Portfolio as a tool for Enterprise Risk Management

Efficient Project Portfolio as a tool for Enterprise Risk Management Effcent Proect Portfolo as a tool for Enterprse Rsk Management Valentn O. Nkonov Ural State Techncal Unversty Growth Traectory Consultng Company January 5, 27 Effcent Proect Portfolo as a tool for Enterprse

More information

APPLICATION OF PROBE DATA COLLECTED VIA INFRARED BEACONS TO TRAFFIC MANEGEMENT

APPLICATION OF PROBE DATA COLLECTED VIA INFRARED BEACONS TO TRAFFIC MANEGEMENT APPLICATION OF PROBE DATA COLLECTED VIA INFRARED BEACONS TO TRAFFIC MANEGEMENT Toshhko Oda (1), Kochro Iwaoka (2) (1), (2) Infrastructure Systems Busness Unt, Panasonc System Networks Co., Ltd. Saedo-cho

More information

Research Article A Time Scheduling Model of Logistics Service Supply Chain with Mass Customized Logistics Service

Research Article A Time Scheduling Model of Logistics Service Supply Chain with Mass Customized Logistics Service Hndaw Publshng Corporaton Dscrete Dynamcs n Nature and Socety Volume 01, Artcle ID 48978, 18 pages do:10.1155/01/48978 Research Artcle A Tme Schedulng Model of Logstcs Servce Supply Chan wth Mass Customzed

More information

8 Algorithm for Binary Searching in Trees

8 Algorithm for Binary Searching in Trees 8 Algorthm for Bnary Searchng n Trees In ths secton we present our algorthm for bnary searchng n trees. A crucal observaton employed by the algorthm s that ths problem can be effcently solved when the

More information

Fault tolerance in cloud technologies presented as a service

Fault tolerance in cloud technologies presented as a service Internatonal Scentfc Conference Computer Scence 2015 Pavel Dzhunev, PhD student Fault tolerance n cloud technologes presented as a servce INTRODUCTION Improvements n technques for vrtualzaton and performance

More information