Message Cab (MCab): Partition Restoration in MANETs Using Flexible Helping Hosts

Size: px
Start display at page:

Download "Message Cab (MCab): Partition Restoration in MANETs Using Flexible Helping Hosts"

Transcription

1 Wreless Sensor Network, 2010, 2, do: /wsn Publshed Onlne December 2010 ( Message Cab (MCab): Partton Restoraton n MANEs Usng Flexble Helpng Hosts Abstract ng Wang, Chor Png Low School of Electrcal and Electronc Engneerng,Nanyang echnologcal Unversty, Sngapore Emal: wang0235@e.ntu.edu.sg, cplow@e.ntu.edu.sg Receved September 18, 2010; revsed September 26, 2010; accepted October 12, 2010 Helpng hosts are ntensvely used n varous schemes to restore parttoned Moble Ad Hoc Networks (MANEs). Most of the exstng schemes offers only determnstc deployment and fxed routes for the helpng hosts, and are thus not able to deal wth fluctuatng network traffc, whch s a practcal condton n many MANE applcatons. In ths paper, we argue that flexble helpng hosts (referred to as Message Cabs (MCabs)), wth deployment and routes that response to the changes n the traffc demand of the network, may overcome ths drawback and reduce the message delay n the networks. o demonstrate the effectveness of ths observaton, we propose a new helpng host scheme namely the Message Cab (MCab) scheme for partton restoraton n MANEs, and valdate the performance through smulatons. Keywords: Moble Ad Hoc Networks (MANEs), Helpng Host, Message Cab, Adaptve Route Desgn, Dynamc Deployment 1. Introducton A Moble Ad Hoc Network (MANE), as descrbed n [1], s a knd of moble wreless networks whch s comprsed of a collecton of moble hosts connected through wreless channels. he drect connectons between the hosts are referred to as lnks. A host n a MANE exchanges messages wth other nodes as a termnal and forwards the messages as an ntermedate router at the same tme. he routng paths of a message n a MANE are formed by a seres of moble hosts. Messages are forwarded hop-by-hop. Whle the moblty of hosts enables the network to span over a large area, t also causes a hghly dynamc topology, whch s a major challenge to the applcatons of MANEs. When a host moves out of another s communcaton range, the lnk between them breaks, and the entre message routng path may be destroyed by ths broken lnk. hs possblty of lnk breakage may splt hosts nto dfferent parts between whch there s no possble path. hs n turn may result n packets not beng able to reach ther destnatons. We refer to ths phenomenon as network parttonng. Each solated part of a network s referred to as a partton of the network. he parttonng problem makes a crtcal strke on ad hoc routng because most protocols typcally assume that the network s always connected. o enhance the relablty and conserve energy, parttonng should be restored n MANEs. Varous approaches have been proposed for MANEs survvablty and restoraton. In partcular, helpng hosts are deployed to reconnect the network parttons, and we refer to such acton as partton restoraton. he helpng hosts are able to reconnect the network connectvty by movng from one partton to another n a MANE despte the fact that the normal hosts (.e. nonhelpng hosts) may be parttoned. hs dea s also extensvely dscussed n Delay olerant Networks (DN) and Wreless Sensor Networks () n order to collect data from dsconnected parts of the networks. he helpng hosts are addressed by dfferent names n varous schemes. In [2-7], they are referred to as ferres, whle n [8-10], they are called helpng nodes or forwardng nodes. Data MULEs n [11-13] are also known as a knd of helpng hosts, and n the DakNet project [14] buses are used as helpng nodes to connect broadband network to rural vllages. One of the most mportant and commonly dscussed objectves of these schemes s to enhance the connectvty and mze the communcaton delay n the network. On contrary, helpng hosts movement consumes conerable amount of energy and tme, and may cause Copyrght 2010 ScRes.

2 892. WANG E AL. long message delay n the network. o effcently utlze them, the deployment and route desgn of helpng hosts are crucal. In ths paper, we propose a new helpng host scheme for partton restoraton n MANEs, namely the Message Cab (MCab) scheme. he helpng hosts are lke cabs, n the sense that they are more flexble and adaptve to traffc than buses or ferres. he MCab scheme conssts of two essental parts, namely the Dynamc Cab Deployment (DCD) algorthm and the Adaptve Cab Route (ACR) algorthm for the purposes of selectng message cabs and mprovng the message cab route, respectvely. Both algorthms are able to deal wth fluctuatng traffc n the network, whch s a practcal scenaro n real lfe, but s not usually dscussed n the exstng works. herefore we say that the message cabs are flexble helpng hosts. We wll demonstrate that the MCab scheme overcomes the drawbacks of exstng schemes by ncurrng lower average message delay and the results are valdated through smulatons. In the followng parts of ths paper, we ntroduce the backgrounds of our work n Secton 2, and the model together wth our objectves n Secton 3. he two parts of the Message Cab (MCab) scheme, namely the Dynamc Cab Deployment (DCD) algorthm and the Adaptve Cab Route (ACR) algorthm are presented n Secton 4 and 5, respectvely. Smulaton and results are dscussed n Secton 6, whle Secton 7 concludes ths paper. 2. Background Before we present our proposed Message Cab (MCab) scheme, we need to explan two mportant concepts, namely the deployment and route desgn of the helpng hosts (or cabs n our scheme), together wth some exstng solutons. In addton, our assumptons and notatons are dscussed n ths secton Deployment Deployment s a procedure for selectng helpng hosts. In most of the exstng works, such as [2,10,13,14], the helpng hosts are a group of specally desgned hosts that have larger storage, more energy and/or hgher movng speed. hey are dfferent from the normal hosts and are assgned as helpng hosts pror to the commencement of network operatons. he deployment s thus statc, or determnstc. Such deployment mechansms allows helpng hosts to have hgher capablty n delverng messages across parttons but s less flexble. Snce the number of helpng hosts s fxed, t may turns out that there s too many, or too few helpng hosts to meet the traffc demand n the network. hus statc deployment may not be adaptve to network traffc demand. o overcome ths drawback, dynamc deployment s used n [8], where normal hosts wth sutable movng drecton and speed are chosen as helpng hosts (a.k.a helpng nodes). However, t s a centralzed scheme and thus not scalable wth network sze. More mportantly, under the assumpton that the network s parttoned, t s nfeasble to make all the hosts movement nformaton avalable to a central server to perform the selecton. Smlar problem can also be observed n [9]. herefore, we propose a localzed algorthm, namely the Dynamc Cab Deployment (DCD) algorthm to deploy helpng hosts (.e. message cabs) n the MCab scheme. he DCD algorthm allows the number of helpng hosts to change dynamcally wth the traffc volume, and thereby reduces the weghted average delay of messages n the network, and enhances the scalablty of the deployment process Route Desgn he route of a helpng host s the path t follows, whch usually connects dfferent parttons or hosts n the network. In [8,12,14], the route s not planned by any algorthm. he hosts random movement s utlzed to delver messages. However, t has been shown n many other works that the message delay can be sgnfcantly reduced f we coner the problem of route desgn as an optmzaton problem. For example, n [2], Zhao et al. have formulated the Massage Ferry Route (MFR) problem to fnd the optmal route for helpng hosts (.e. message ferres). In [4-7], solutons from the well studed ravelng Salesman Problem (SP) and ts varants are adopted as routes for helpnghosts. In these schemes, the route does not change after beng constructed, and s thus a fxed route. It s only sutable when the network traffc among the parttons s constant, whch s an dealzed assumpton. In most of the practcal cases, the traffc between any two parttons wll be a temporal varable, and the route should be able to adapt to the changes. Hence fxed routes are unlkely to be optmal n general. herefore, an adaptve route s more desrable. Ou et al. proposed a centralzed scheme n [10] to let the helpng hosts (a.k.a helpng nodes) move adaptvely to the packets destnatons. However, as we have dscussed n the prevous secton, centralzed schemes may perform poorly when the network s parttoned. In ths paper we propose a dstrbuted algorthm, namely the Adaptve Cab Route (ACR) algorthm, to construct cab routes by adoptng the Shortest Process me Frst (SPF) rule from the Job Sequencng Problem (JSP). Smulaton proves that our proposed scheme s able to effectvely reduce the message delay n the network. Copyrght 2010 ScRes.

3 . WANG E AL Assumptons and Notatons Whle many exstng schemes (e.g. [2,7,11] ) focus on statonary hosts wth known locatons and predctable network traffc, we try to deal wth a more practcal scenaro, where: the hosts are moble; the traffc n the network s randomly varant and s thus not predctable. As dscussed n [15], movements of hosts n MANEs can be conered as Levy Flghts, n whch the hosts tend to stay n a certan area for a long tme and occasonally make long dstance flghts to places far away. We could thus assume that a host stays n the same cluster for a certan perod of tme and wll also sometmes decdes to move to another cluster. Practcally, ths could be due to the assgent of tasks to the host, such as n a wreless sensor network, a sensor may need to reallocate tself to new postons to collect new data. Let s assume that the MANE of our nterest has n h moble hosts. We assume the network s parttoned to n c components, each of whch forms a cluster wth a chosen cluster head. he hosts n the same cluster are always connected. In order to smplfy the problem, we restrct the movement of a cluster head to be ne a crcular area, whch s referred to as the head zone. he radus of head zone s equal to the communcaton range of the cluster head, denoted as r, as shown n Fgure 1. he center of head zone of cluster s s denoted as pont 1 s n c. We use to denote the tme taken by a host to travel from C s to C d,1 s, d nc. We note that s a postve real number. We say cluster d s a neghbor of cluster s f a host s able to move drectly to cluster d from cluster s n one hop. Usng proper clusterng schemes, such as [16], the cluster head wll always be aware of the changes n ts cluster members, such as the movements, arrvals of new members and departures of exstng members. It also knows the locatons of other clusters, so when a host decdes to move to a new place, the cluster head knows whch s the next cluster the host s gong to jon. In add- Fgure 1. Network topology. ton, the cluster head works as a proxy between other normal hosts n the cluster and the helpng host as well. If a helpng host stops at C s, t s able to communcate wth the head of cluster s and delver the messages to other hosts n cluster s va the cluster head. he route of a helpng host, denoted as R, wll be a sequence of head zone centers ( C s 's ). Each of these head zone centers s lke a waypont (as llustrated n Fgure 1), whch s a temporary destnaton for the helpng host. Upon arrvng a waypont, the helpng host stays for a perod of tme to delver and receve the messages to/from the cluster head, before movng on towards the next waypont. he route segment between two consecutve wayponts s referred to as a leg (shown n Fgure 1). 3. System Model and Objectves o demonstrate our methodology of the MCab scheme, we descrbe an analogy between the MANE and a smple transportaton system whch serves several small towns. It wll also explan the reason why we name our helpng hosts as cabs and how they are dfferent from ferres. he towns represent the clusters, each of whch s dstant away from the others. he cars are lke the hosts they usually move wthn a town, and when t s necessary, they are also able to move from one town to another. he passengers are akn to messages, whch by themselves cannot move from town to town. Buses are provded as an exstng soluton to the nter-town transportaton problem. hey are smlar to the other statcally deployed helpng hosts (such as message ferres) that we have dscussed n the prevous secton. he statc deployment and fxed route desgn of buses perform poorly wth varyng volume of passengers. Hence, t would be preferable f we have a flexble soluton. Bees takng buses, the passengers could coner hrng a cab to travel to another town. he route of a cab wll depend on the passengers demands, and s thus more flexble than the bus routes. Practcally, t s also more convenent n terms of savng tme to hre a cab. he number of cabs s determned by how many passengers there are. When there are more passengers, more cabs can be recruted from the prvate cars (normal hosts) avalable; when the number of passengers drops, some cabs could retre and become prvate cars agan. Moreover, snce t s possble that some cars wll move to another town by ther drvers own decsons, a passenger may also take a rde from a prvate car whch s also movng to the hs/her destnaton town. hs wll utlze the moblty of prvate cars n the transportaton system to serve the passengers needs and save ther travelng tme. A car offerng rde to passengers s lke a Copyrght 2010 ScRes.

4 894. WANG E AL. one-tme temporary cab that only works for a sngle trp to a partcular town. We refer to such cars as temp-cabs. o dstngush from them, we refer to those cabs that works for multple trps n a longer perod of tme as professonal cabs, or pro-cabs. We mtate the method of hrng pro-cabs and takng rdes from temp-cabs n the MANEs and propose the Message Cab (MCab) scheme. As shown n Fgure 2, the state of a prvate car (normal host), whch s not a cab nor a cluster head, can transform to a pro-cab or a tempcab under the control of the Dynamc Cab Deployment (DCD) algorthm: Pro-Cab Its movement s passvely determned by the Adaptve Cab Route (ACR) algorthm, n whch the route s adaptvely desgned based on the network traffc. It s recruted from a normal host, and contnues movng among the clusters for a certan predetermned perod of tme before retrng to be a normal host. emp-cab Its movement s pro-actvely determned by the host tself. It s a one-tme helpng host that only leaves ts orgnal cluster to ts destnaton cluster. It becomes a normal host agan after t arrves and has delvered the nter-cluster messages to the destnaton cluster. he key dfference between message cabs and the other types of helpng hosts n the exstng works (such as ferres, helpng nodes etc.) s that they are not selected before the network starts. herefore the number of cabs and cab routes can dynamcally change wth the traffc. Although ths knd of deployment does not allow the helpng hosts to have hgher capablty n movng speed or storage space (snce message cabs are just selected normal hosts), we found that t can stll effectvely reduce the message delay n the network. Snce hosts n the same cluster are connected, we can use exstng MANE routng protocols such as AODV or DSR for the communcaton between a cluster head and ts members. We note that the delay ncurred by the message transmsson wthn a cluster (ntra-cluster communcaton) s much smaller than the delvery between dfferent clusters (nter-cluster communcaton), and s less relevant to the cab deployment and route desgn. As a consequence, ntra-cluster communcaton wll be omtted n our followng dscusson. We should note that before a message s collected by a cab, t needs to wat at some cluster head for a certan tme duraton. We refer to ths amount of tme spent on watng as the watng delay of the message. After t s collected by a cab, the cab wll travel from ts source cluster to ts destnaton cluster, and thus ncurs a travelng delay for the message. herefore, the overall delay of a message s. Assume wthn tme duraton (0, t], there are n m messages that have been transmtted by the message cabs. he sze of message 1 s. he watng, travelng and overall delay of message are denoted as and, and.respectvely. Smlar to the objectves n [2-4], we are nterested n reducng the weghted average overall delay of the messages: where the weghted average watng delay ( ) s gven by 1, 1 and the weghted average travelng delay ( ) of all the messages can be taken as 1 1, We can see that f there are more cabs n the network, each cluster could be vsted more frequently, and the watng delay of messages reduces. It shows that the watng delay s closely related to the cab deployment plan. On the other hand, optmzaton of cab routes results n a shorter travelng delay. In the MCab scheme, we bound from above by usng the Dynamc Cab Deployment (DCD) algorthm to control the number of cabs, and s reduced by adoptng optmzaton technques n the Adaptve Cab Route (ACR) algorthm, whch desgns the routes of the cabs. (1) (2) (3) 4. Deployment of Message Cabs Fgure 2. States of a host In the deal scenaro, once a nter-cluster message s receved by the cluster head, there s always a cab ready to Copyrght 2010 ScRes.

5 . WANG E AL. 895 carry t towards ts destnaton, and wll be 0. However, n practce, wth unpredctable and fluctuatng traffc, t s mpossble to have a cab ready for message delvery as soon as a message arrves at the cluster head. It would also be neffcent f a cab carres only one message and does not fully utlze ts storage space. o solve ths problem, we try to make use of the normal hosts moblty as a temp-cab to gve a rde to the messages as well as to recrut pro-cabs from normal hosts n the DCD algorthm. We show that by dong so, we are able to bound the weghted watng delay of messages from above by a predetermned value, denoted as seconds. Snce the change of cluster membershp s handled by the cluster head, a host (say host a ) needs to report to the head of ts current cluster, say cluster s (as source) before t leaves for a new locaton. he cluster head of clusters then knows the new cluster (say cluster d, as destnaton) host a s gong to jon. It becomes possble that the head of cluster s lets host a to carry some messages whch have cluster d as ther destnatons. Host a a wll then delver these messages to the cluster head d when t arrves there and regsters tself as a new cluster member. herefore, when a host s leavng a cluster, ts state changes from normal host to temp-cab, and t s used to delver as many messages as possble across the parttoned clusters. Upon ts arrval, t delvers the messages to the head of the new cluster and ts state changes back to that of a normal host. hs procedure of deployng a temp-cab s depcted n Fgure 3(a). On the other hand, a tmer s used to control the tme when a pro-cab should be selected. Assumng there are n ' m nter-cluster messages stored n cluster head s, of each the sze s. he current watng delay of a message s the length of the tme duraton snce the message s receved by the cluster head, and s denoted as '. he tmer s set to t n' m ', 1 n' m 1 so that f a pro-cab s recruted rght before the tmer expres, the average weghted delay of the messages stored n cluster head s wll be less than, whch s the requred upper bound. We note that the value of t needs to be updated every tme there s a change n the number of nter-cluster messages whch are stored n cluster head s, such as when a new nter-cluster message s receved and stored, or when some messages have been uploaded to a cab (ether temp or pro) by cluster head s. We let the tme duraton between the recrutment and retrement of a cab be denoted as. After havng served for seconds as a pro-cab, t retres. Upon ts re trement, a pro-cab does not stop mmedately. It contnues movng among the clusters to delver the remanng messages t has already collected, but wthout collectng new messages. When all the messages are delvered, t moves back to the cluster where t has been recruted, and regster wth the head as a normal host agan. Fgure 3 depcts the flowcharts for the DCD algorthm. o avod gong nto too much detals, we make some smplfcatons to the algorthm: he pro-cabs are randomly selected from the normal hosts n the cluster; If a cab does not have enough space to store all the messages for nter-cluster delvery, t delvers those messages wth larger weghted watng delay ' frst; When the memory of a cluster head overflows, those messages wth larger weghted watng delay ' wll also be dropped frst (.e. the hot-potato rule 1 ); he value of (.e. the length of tme duraton between the recrutment and retrement of the cabs) s a constant. Wth these smplfcatons, the DCD algorthm s a fully dstrbuted scheme whch does not requre any global nformaton. Lemma 1. he worst case complexty of the DCD algorthm s On m, where s the number of messages that have been generated durng the perod of network operaton. Proof: Based on the algorthm depcted n Fgure 3, we can observe the fact that the complexty of the procedures s ndependent of the number of clusters and the number of hosts. o update the tmer (refer to step 5 n Fgure 3(a) and step 3 n Fgure 3(b)), the cluster head needs to go through the messages that have been stored n ts memory space ( n ' m messages). In the worst case, all the messages are stored, and t thus requres On m tme. On the other hand, to upload the messages to the cab (step 4 n Fgure 3(a) and step 2 n Fgure 3(b)), and to delver the messages to another cluster head (step 7 n n Fgure 3(a), step 5 and step 8 n Fgure 3(b)), t requres O n m tme as well. We note that the plannng of routes (step 6 n n Fgure 3(a), step 4 and step 7 n Fgure 3(b)) wll be handled by the ACR algorthm, and s not part of cab deployment process handled by the DCD algorthm; thus ts complexty wll not be counted here. In addton, step 1 n Fgure 3(a) s handled by the clusterng scheme, and the pro-cab selecton (step 1 n Fgure 3(b)) s random accordng to our assumptons. herefore they do not ncur addtonal com- 1 We deal wth the elements that have heavest mpact to the system by ether processng them or droppng them frst. By droppng those messages wth heaver weghted delay, the average weghted delay of the success-fully delvered messages can be reduced. Copyrght 2010 ScRes.

6 896. WANG E AL. Cab Route (MCR) problem as a generalzaton of the exstng Massage Ferry Route (MFR) problem, and propose the Adaptve Cab Route (ACR) algorthm as a soluton. Snce only the pro-cabs route wll be desgned by our algorthm, we refer to them as cabs for the sake of convenence n ths secton he Message Cab Route (MCR) Problem In [2], Zhao et al. have formulated the Massage Ferry Route (MFR) problem to fnd the optmal route for helpng hosts (.e. message ferres), and ts objectve s to mze the average delay of the traffc: R R 1 s, dnb c = (4) b 1, nc where b s the average traffc per unt tme from s to d and R s the tme spent by the ferry to travel from (a)emp-cab (b) Pro-Cab Fgure 3. Flowcharts of the dynamc cab deployment (DCD) algorthm. plexty to the DCD algorthm. In addton, the f clauses (steps 2, 3 n Fgure 3(a) and step 6, 9 n Fgure 3(b) does not ncur addtonal complexty to the algorthm. As a result, the overall complexty of the DCD algorthm s O n m 5. Message CAB Route Desgn o mprove the route of cabs, we defne the Message s to d on the statc route R. But n our case, b could not be determned beforehand because we have assumed that the traffc s unpredctable. Moreover, when R s not fxed (.e. beng adaptve nstead), the travelng delay from one cluster to another s not a constant. herefore the solutons for the MFR problem may not perform well for the problem that we are addressng. Hence we defne a generalzed verson of the MFR problem whch s able to adapt to both statc and dynamc traffc demand and routes. We refer to ths problem as the Message Cab Route (MCR) problem: Defnton 1. (he Message Cab Route (MCR) problem) For a gven group of clusters, fnd the optmal vstng sequence for a cab, so that the weghted average travelng delay to delver n m messages, whch s expressed by 1 1 where s the travelng delay of message, can be mzed. We could make the followng observatons from the MCR problem: he MCR problem s a generalzaton of the MFR problem. In the MFR problem, there are two addtonal assumptons made as compared to the MCR problem, namely the traffc rate b between two clusters s constant, and the dstance R from cluster s to d on the gven route R s also a fxed value. hat s, for any message sent from cluster s to cluster d, ts travelng delay wll be. Let s say the network R operates (5) Copyrght 2010 ScRes.

7 . WANG E AL. 897 from tme 0 to t. he total sze of all the messages that have been generated durng ths perod can be expressed as 1 s, dntb c. On the other hand, usng our notaton, snce the number of these messages s n m, and the sze of each message s de- noted by and thus, the total sze s thus 1, n c 1 t b 1 s, dn c tb n m R 1 n m 1. herefore Equaton (5) becomes equvalent to Equaton 4 under the assumptons of the MFR problem. Hence t s easy to see that the MCR problem s a generalzed verson of the MFR problem. he MCR problem s MAX-SNP-hard, and thus polynomal tme approxmaton schemes are unlkely to exsts to ths problem. We coner a smpler specal case of the MCR problem. Coner a partcular nstance where a cab collects nc 1 messages, each for a dstnct destnaton cluster, from cluster s. Before t fnshes delverng these messages, other clusters do not upload any new message to the cab. he objectve functon, Equaton (5) therefore becomes nc 1 1 nc 1 1 hs s exactly the objectve functon of the Weghted Mum Latency Problem (WMLP) [17], where s referred to as the latency of the tour startng from s to the destnaton cluster of message, and s the weght. In the WMLP, we try to fnd the tour that vsts all the vertces n a graph that mzes the weghted latency, represented by the above equaton. Unfortunately, the WMLP s known to be a MAX-SNP-hard problem [18]. Generally, there s no polynomal tme approxmaton scheme for MAX-SNP-hard problems [19]. Even for a more smplfed scenaro, where s constant and the traffc s unform among the clusters, to mze n c 1 1., we stll need to solve the rav- elng Salesman Problem (SP), whch s also known to be NP-hard. herefore, t s nfeasble to completely generate an optmzed cab route. A statc route cannot be an optmal soluton. hs s because n the network, traffc cannot be constant and unform all the tme. If some cluster receves messages more frequently than others, t wll be a waste of tme to vst all the other clusters before returnng to, whch s what that wll take place wth a statc route. Moreover, due to the nature of some of the MANE tasks, such as area exploraton or survellance, the frequency and szes of messages may vary over tme. A statc route could not adapt tself to such changes. he key to fndng a sutable cab route s to determne the next leg. he cab stops to delver and collect messages to/from a cluster head after each leg. herefore, we do not need to plan the route beyond the next leg, as the future legs should be adaptvely determned accordng to the messages the ferry collects n the future and have not been unveled to the ferry yet. herefore, f a scheme can produce each leg optmally, t offers a good soluton to the MCR problem. Based on the above observatons, we use a weghtng scheme to choose the best 2 cluster head among all the neghbors as the next waypont for a cab to move to and ths n turn wll eventually forms an adaptve route. he Shortest Process me Frst (SPF) rule from the Job Schedulng Problem (JSP) s adopted to compute the weght of the clusters, and we refer to our proposed approach as the Adaptve Cab Route (ACR) algorthm he Shortest Process me Frst (SPF) Rule In the Job Sequencng Problem (JSP) [20] of a sngle machne, we need to fnd the optmal sequence of a seres of jobs k wth weght 3 l k and processng tme p k to mze the average delay of these jobs whch s gven by: klkk jsp klk Smth proved n [21] that the optmal soluton to JSP can be produced by applyng the Shortest Process me Frst (SPF) rule to sequence the jobs: SPF Rule: Sequencng the jobs n order of nondecreasng rato pk l k produces an optmal schedule to mze jsp. Comparng Equaton (6) wth Equaton (3), we can observe that the objectve functons of JSP and MCR are smlar (detals wll be dscussed n the next Secton). Hence SPF would be a useful tool to construct adaptve ferry routes. In ths paper, to ensure that the equatons that we derve n the next Secton are well defned, we 2 As the most sutable waypont n a cab route to mze 3 Larger l k represents hgher mportance and vce versa. Copyrght 2010 ScRes.

8 898. WANG E AL. defne an nverse SPF (SPF) rule: SPF Rule: o optmally sequence the jobs n JSP, the job wth hghest lk p k value should be chosen frst. It s easy to see that SPF s logcally equvalent to SPF. hey are both able to solve JSP optmally he Adaptve Cab Route (ACR) Algorthm We may nterpret the event that the cab goes to a destnaton cluster d as job d. Assumng the cab s currently at cluster s, we could say that the requred processng tme of job d s the travelng tme from cluster s to cluster d, e.., whch s smlar to p k n JSP. In order to apply the SPF rule, we need to determne the weght l k of these jobs. We note n JSP, the weght of a job ndcates the how much each job contrbutes to the objectve functon jsp. Smlarly, job d contrbutes to n the MCR problem by delverng the messages wth destnatons n cluster d. Usng dest to denote the ndex of the destnaton cluster of message (.e. message s meant to be delvered to C dest ), the total sze of messages delvered by the cab n job d can be expressed as. d dest : We use ths value as the weght of job d (lke l k n JSP). hen we can choose the cluster wth maxmal rato of Md as the next waypont of the adaptve cab route accordng to the SPF rule. We also note that MCR and JSP are stll two dfferent problems. Because n MCR, vares when the cab s locaton (cluster s) changes, whle n JSP p k s constant. herefore SPF may not provde the exact optmal soluton to MCR. However, we beleve t stll provdes a good ndcaton of whch cluster head should be chosen to be the next waypont. Accordng to SPF, we could defne a weghtng functon W d as d (7) Wd ε A cluster wth the maxmal weght wll be chosen as the next waypont n the adaptve route. o break a te, we defne d as the tme of the most recent arrval of the cab at cluster d. Gven any two clusters C 1 and C 2, f 1 < 2, t means that the tme snce the last cab s arrval to cluster C 1 s longer than that to cluster C 2. If the two clustersboth have the maxmum weght W among all the clusters, C 1 wll be chosen as the next waypont. he algorthm shown n able 1 llustrates how the ACR algorthm determnes the cab s next waypont when t arrves cluster s at tme. Lemma 2. he complexty of the ACR algorthm s able 1. ACR algorthm. ACR(cluster s, tme ){ ntalze W 0 d and M 0 d for all d for each un-delvered message do %Loop 1% dest for each cluster d do %Loop 2% d Wd for all the clusters do %Loop 3% fnd d max of whch W s the maxmum dmax f there are multple that d ' s W W then d' dmax for all the d ' s do %Loop 4% fnd d ' max of whch d ' max s the mum the next waypont s C d else the next waypont s C set s } 'max dmax O nc. Proof: In Loop 1 of ACR, the algorthm goes through the cab s storage to calculate the accumulated sze of messages that wll be delvered to cluster dest. Snce the total number of messages s, Loop 1 takes On m tme n the worst case. o calculate the weght W d for each cluster n Loop 2, the algorthm goes through all the n c clusters and therefore On c tme s requred. Smlarly, n order to fnd the cluster wth maxmum weght, another On c tme s requred n Loop 3. o break the te, Loop 4 wll be executed for Onc tmes n the worst case. he overall complexty for algorthm ACR s thus Onc. We note the complexty of the ACR algorthm (Lemma 2) as well as the DCD algorthm (Lemma 1) s much lower than most of the exstng schemes wth helpng hosts. Moreover, beng a fully localzed scheme, MCab s extremely sutable for the MANEs where the resources are lmted. 6. Smulaton Results o valdate the effectveness of the MCab scheme, extensve smulatons are carred out. We randomly generate the network topology and traffc n C++ programs and compare the results wth exstng schemes. We model a MANE constructed n a 500 m by 500 m area. he number of hosts n each cluster s set to 20 at the begnnng of the smulatons. he communcaton range of the hosts s 5 m, and the storage sze s 10Mb. he hosts move at a speed of 10 m/s. he traffc n the MANE can be specfed by the number of messages per unt tme, but as the sze of messages also vares, t s more convenent to descrbe t as the sze of data transmtted per unt tme,.e. klobts Copyrght 2010 ScRes.

9 . WANG E AL. 899 per second (kb/s). We refer t as the volume of the traffc. he number of pro-cabs may vary n the smulaton. We defne the effectve number of pro-cabs as the average number of pro-cabs over tme to descrbe how many pro-cabs are deployed n the MANE Performance of the DCD Algorthm We start wth our study on the performance of the DCD algorthm. he man objectve of the DCD algorthm s to dynamcally recrut cabs n the MANE to delver nter-cluster messages. he algorthm s controlled by two parameters, namely (the upper bound of the weghted watng delay) and (the tme duraton that a host serves as a pro-cab). o deploy temp-cabs, the performance of the DCD algorthm s also closely related to the frequency n whch the hosts move from one cluster to another. A varable s defned as the average length of perod (n seconds) that a host stays n a cluster before t moves to another one. Fgure 4 dsplays how the number of recruted cabs changes wth the volume of nter-cluster traffc. he left vertcal axs corresponds to the traffc volume whle the number of pro-cabs are ndcated on the rght axs. o study the mpact of, and on the performance of the DCD algorthm, we choose two (hgh and low) values for each of these varables. Each of the fgures corresponds to a partcular combnaton of the values of and, and plots the two groups of results whch correspond to the two values of. As shown n able 2, the values of are 500 s and 100 s; the values of are 2000 s and 200s; the values of are 5000 s and 200s. he effectve number of pro-cabs are also shown n able 2. 1) he value of : he value of sgnfcantly changes the effectve number of pro-cabs. Wth a hgh value of ( = 5000 s), the hosts tend to stay n the same cluster for a long tme, and thus there are less chances that messages could be delvered by a temp-cab. herefore, more pro-cabs have to be recruted to keep the watng delay below. If the value of s low ( = 200 s), t mples that the hosts frequently move among the clusters. herefore a lot of nter-cluster messages could be delvered by temp-cabs, and there are fewer messages that are left on the cluster head. herefore, fewer pro-cabs need to be recruted. able 2. Effectve Number of Pro-Cabs. 200s 5000s 2000s 200s 200s 2000s 100s s We can observe ths phenomenon n Fgure 4(a) and Fgure 4(c) by comparng lne 1 ( 200 s, 100 s, 5000 s ) wth lne 5 ( 200 s, 100 s, = 200 s). For the same values of and, more procabs wll be recruted when 2000 s, as depcted by lne 1. When decreases to 200 s, more temp-cabs could be used, and thus fewer pro-cabs are needed to delver the messages, as shown by lne 5. Smlar fact can be observed by comparng lne 2 wth lne 6, lne 3 wth lne 7, and lne 4 wth lne 8 n Fgure 4. In able 2, we can also see the effectve number of pro-cabs s much smaller when s low ( 200 s ). 2) he value of : controls the duraton for whch a host serves as a pro-cab. It controls the frequency for trggerng the pro-cab recrutng procedure, and thus affects how fast the system could respond to the change n traffc volume. When the value of s hgh ( 2000 s ), the procabs have a long servce tme and a slow retrement. Comparng lne 1 wth lne 2 n Fgure 4(a) on the tme nterval [7000, 8000], we can see that the number of cabs stays as hgh as 6 when 2000 s (as depcted by lne 2) even though the traffc volume has already dropped to a lower level, where only about 3 pro-cabs wll be deployed f 200 s (as depcted by lne 1), meanng that several cabs may be unnecessary for the purpose of keepng the weghted watng delay low. Smlar phenomena can also be observed n the tme nterval [4000, 5000] n Fgure 4(a), [9000, 10000] n Fgure 4(b), and [8500, 10000] n Fgure 4(c), resultng n the effectve numbers of cabs beng much larger when = 2000 s (as shown n able 2). A low value of ( 200 s ) causes the pro-cabs to only serve for a short perod of tme, and change back to the state of normal host sooner. However, messages are stll beng generated and stored n the cluster heads, causng new pro-cabs to be recruted. he change n the cab number s thus more rapd and frequent than when 2000 s. Snce cabs are frequently recruted and retred and lnes 1, 3, 5 and 7 appear to be more spky than lnes 2, 4, 6 and 8 respectvely n Fgure 4. We should also note that frequent change n the number of cabs may also cause unwanted nterrupton to the other tasks of the hosts. For example, n order to collect relable data, a sensor may need to stay statonary at the same poston for some mum duraton. Hence low values of may affect the effcency of the network task, although t s able to help to reduce the number of pro-cabs. 3) he value of : Comparng Fgure 4(a) wth Fgure 4(b), or Fgure 4(c) wth Fgure 4(d), we can also observe the fact that the value of nfluences the number of cabs more drectly. Copyrght 2010 ScRes.

10 900. WANG E AL. (a) = 100 s, = 5000 s When the value of s hgh ( 500 s ), the cluster heads are able to tolerate larger weghted watng delay before recrutng new pro-cabs, and less pro-cabs wll be hred n the MANEs. hs s the reason why n able 2, the effectve numbers of pro-cabs are much smaller when 500 s. On the other hand, f the value s low ( 100 s ), the cluster heads have to frequently hre new pro-cabs to keep the watng delay of the messages low. Comparng Fgure 4(a) wth Fgure 4(b), or Fgure 4(c) wth Fgure 4(d), t s easy to observe that the number of procabs sgnfcantly ncreases when the value of s low. In general, our smulaton results show that the DCD algorthm can effectvely adapts the number of cabs to the traffc volume of the network. We wll also show how t bounds the weghted watng delay below n Secton VI-C. Before that, the performance of the ACR algorthm wll be dscussed Performance of the ACR Algorthm (b) = 500 s, = 5000 s (c) = 100 s, = 200 s (d) = 500 s, = 200 s Fgure 4. Performance of the DCD algorthm. o evaluate the performance of the ACR algorthm, we assume that cabs are selected beforehand and are randomly deployed wthout the use of the DCD algorthm n ths secton. We compare the delay ncurred by the adaptve route constructed by the ACR algorthm wth the most commonly adopted soluton,.e. the SP route. It s well-known that SP s a NP-hard problem, where optmal soluton cannot be found wthn polynomal tme. We use brute force to fnd the optmal SP route n the smulaton. However t would only be feasble to use o address ths ssue, we also use the Improvng Search Algorthm [22] to fnd an approxmate soluton to SP as a feasble route. Both optmal and approxmate SP routes are statc and are used as benchmarks to compare wth the routes obtaned by the ACR algorthm. 10 clusters of hosts are conered for ths smulaton. he cabs are randomly allocated n dfferent clusters ntally, and move accordng to the routes generated by the ACR algorthm, optmal SP algorthm, or approxmate SP algorthm. he results are shown n Fgure 5. We can see that ncreasng the number of cabs reduces the average delay of the ACR routes, but no sgnfcant change can be observed for SP and approxmate SP routes. hs s because when the same statc route s used, the delay of a message only depends on the dstance between ts source and destnaton on the route, and s not affected by the number of cabs. herefore once the route s constructed, the delay of messages wll not be affected by the change n the number of cabs. On the other hand, when the ACR algorthm s used, each cab defnes ts own route based on the messages t Copyrght 2010 ScRes.

11 . WANG E AL. 901 Fgure 5. Performance of the ACR algorthm (travelng delay). s carryng. he delay of a message can be further shortened f there are fewer messages to be carred by a sngle cab. For example, f a cab only carres one message, t wll drectly move to the destnaton cluster of the message. A far amount of tme can be thus saved. As the number of cabs ncreases, fewer messages wll be carred by each of them, and the average delay s thus further decreased. Moreover, snce the ACR algorthm s fully dstrbuted, no collaboraton s needed among the cabs. Each of them can decde ts route locally. he ncrease n cab number does not affect the complexty of the scheme at all Overall Performance of the MCab Scheme In the MCab scheme, when both the DCD and ACR algorthms are appled for the cab deployment and route desgn, the average weghted delay s further reduced. We demonstrate ths result n Fgure 6 to Fgure 8 by showng how the MCab scheme perform when there are dfferent number of clusters n c. he three fgures dsplays the weghted average delay of the messages, the effectve number of pro-cabs and the percentage (n sze) of messages delvered by procabs, respectvely. In each fgure, two values of (5000 s and 200 s) are used n the smulaton. wo dfferent mplementatons of the DCD algorthm ( 100 s, 2000 s and 500 s, 200 s ) are demonstrated. he results for the other two cases ( 500 s, 2000 s and 100 s, 200 s ) exhbt smlar results, and are thus omtted n ths secton In Fgure 6 the smulaton results are shown as a group of columns. he upper part (above 0) of each column shows the duraton of average watng delay, and the lower part (below 0) shows the duraton of average travelng delay. As a result, the entre bar length shows the overall average delay of the correspondng scheme. We compare the results wth the case where a sngle helpng host s deployed n themf scheme, and moves on a SP route or an approxmate SP route. Moreover, we also plot the values of the average weghted delay when the ACR algorthm s appled to a sngle cab wthout the DCD algorthm n the fgure as a reference. It can be seen from Fgure 6 that the total delay s much lower wth the MCab scheme. Let s coner the case when there are 5 clusters. We can observe that for ths case the MF scheme wth approxmate SP ncurs the longest message delay. It s followed by the MF scheme wth optmal SP, because the optmal route of SP wll be shorter than an approxmate route. As we have shown n Secton VI-B, the ACR algorthm wll reduce the travelng delay by usng adaptve cab routes nstead of fxed ones. However, as the ACR algorthm calculates the weghts of the clusters based on the messages that have been stored n the cab, t does not guarantee a lower watng delay for those messages that have not yet been uploaded to the cabs. herefore, the watng delay s not sgnfcantly reduced by the ACR algorthm alone (wthout the DCD algorthm). Moreover, t can be observed n some partcular ncdences that t may ncur a slghtly hgher watng delay than the MF schemes (e.g s wth 15 clusters). When the DCD algorthm s also appled (.e. the MCab scheme), the watng delay s further reduced, because some of the messages can be carred by the temp-cabs and do not need to wat for the pro-cabs. he message delay reduces to the mum when the MCR scheme wth 100 s and 2000 s s adopted. hs s because when 100 s, more procabs are deployed than the case when 500 s as we dscussed n Secton VI-A. he same fact can also be observed from Fgure 7, where on average 4.5 pro-cabs are deployed to serve 5 clusters when 5000 s, 100 s as compare to 1.3 pro-cabs when 500 s. We note the other cluster szes n Fgure 6 also exhbt the same performance trend descrbed above. It should also be ponted out that when there are 5 clusters and 5000 s, the dfference between the results for MCab ( 500 s, 2000 s ) and ACR (wthout DCD) s not as sgnfcant as when there are more clusters, or when the value of s lower ( 200 s, n Fgure 6(b)). hs s because the watng delay n ths case s lower than 500 s even when there s only one cab. As a result, we note that the average number of cabs s also about 1 when the DCD algorthm s used, whch means t s not necessary to deploy more cabs to further reduce the average weghted watng delay n the MCab scheme. Moreover, snce the number of temp-cabs deployed n ths scenaro s also very few as s large, the performance s very smlar to the ACR (wthout DCD) case. Copyrght 2010 ScRes.

12 902. WANG E AL. (a) = 5000 s Fgure 6. Delay of messages. (b) = 200 s (a) = 5000 s Fgure 7. Effectve number of pro-cabs. (b) = 200 s (a) = 5000 s (b) = 200 s Fgure 8. Percentage of messages delvered by pro-cabs. When the number of clusters ncreases, the weghted average watng delay wth the ACR algorthm (wthout DCD) ncreases as well. hs s due to the fact that the pro-cab needs to vst more clusters before delverng the message to ts destnaton. As a result, when the DCD algorthm s appled (.e. the MCab scheme), more procabs wll be recruted to lower the watng delay so that t does not exceeds the predetermned upper bound. Moreover, we can observe that under the control of the DCD algorthm, the weghted average watng delay of Copyrght 2010 ScRes.

13 . WANG E AL. 903 the messages are bounded by the values of, whch are ndcated by the respectve reference lnes (500 s and 100 s) n Fgure 6. he effectveness of deployng temp-cabs can also be demonstrated by comparng the heght of the columns representng MCab scheme ( 500 s, 200 s ) n Fgure 6(a) and Fgure 6(b), where the values of weghted average watng delay decrease wth the value of. As decreases from 5000 s to 200 s, more temp-cabs can be used for nter-cluster message delvery, thus less pro- cabs are recruted (as depcted n Fgure 7), and a lower percentage of messages wll be delvered by the pro-cabs (as depcted n Fgure 8). It thus reduces the watng delay of the messages, snce some of them are taken to ther respectve destnaton clusters by temp-cabs before the tmer expres. hs phenomenon cannot be observed when 100 s, 2000 s, and when there are 20 clusters for 500 s, 200 s, because n these cases addtonal pro-cabs are recruted n order to bound the watng delay below. Interestngly, the travelng delay s also slghtly reduced by usng more temp-cabs. hs s because the messages carred by a temp-cab are selected based on ther destnatons. he temp-cab takes them from the source cluster head and moves drectly to ther destnaton cluster, whle a pro-cab s shared by messages wth dfferent destnatons, and may have to vst several other clusters (as destnatons of other messages stored n the pro-cab) before eventually delver these messages. However, as we can see from Fgure 8, the majorty of messages are delvered by the procabs n most of the scenaros, and the weghted average travelng delay s stll domnated by the routes of pro-cabs,.e. the ACR algorthm. In concluson, smulaton valdates that both the DCD and ACR algorthms works as we expected, and the MCab scheme effectvely reduces the average weghted watng delay. 7. Conclusons In ths paper, we propose a scheme wth flexble helpng hosts, namely Message Cab (MCab) for message delvery n parttoned MANEs. It conssts of two algorthms, referred to as the Dynamc Cab Deployment (DCD) algorthm and the Adaptve Cab Route (ACR) algorthm. he DCD algorthm dynamcally selects hosts to become the helpng hosts (cabs) and move among the clusters to delver nter-cluster messages, and the ACR algorthm desgns the route of the cabs so that the delay of the messages can be reduced. Comparng wth the exstng schemes wth helpng hosts, we have shown that the MCab scheme effectvely shorten the delay of the messages wth the smulatons and s adaptve to the varyng traffc n the network, whch has not been dscussed n the exstng works. 8. References [1] S. Corson and J. Macker, Moble Ad Hoc Networkng (MANE): Routng Protocol Performance Issues and Evaluaton Coneratons, Unted States, [2] W. Zhao and M. Ammar, Message Ferryng: Proactve Routng n Hghly-Parttoned Wreless Ad Hoc Networks, Proceedngs of the 9th IEEE Workshop on Future rends of Dstrbuted Computng Systems (FDCS 03), Washngton, DC, 2003, pp [3] W. Zhao, M. Ammar and E. Zegura, A Message Ferryng Approach for Data Delvery n Sparse Moble Ad hoc Networks, Proceedngs of the 5th ACM Internatonal Symposum on Moble Ad Hoc Networkng and Computng (MobHoc 04), New York, 2004, pp [4] W. Zhao, M. Ammar and E. Zegura, Controllng the Moblty of Multple Data ransport Ferres n a Delay-olerant Network, Proceedngs of the 24th Annual Jont Conference of the IEEE Computer and Communcatons Socetes (INFOCOM 05), Mam, Vol. 2, 2005, pp [5] M. M. B. arq, M. Ammar and E. Zegura, Message Ferry Route Desgn for Sparse Ad Hoc Networks wth Moble Nodes, Proceedngs of the 7th ACM nternatonal symposum on Moble Ad Hoc Networkng and Computng (MobHoc 06), New York, NY, 2006, pp [6] M. Ye, X. ang and D. L. Lee, Far Delay olerant Moble Data Ferryng, Proceedngs of the 10th Internatonal Conference on Moble Data Management: Systems, Servces and Mddle-ware (MDM 09). Washngton, DC, 2009, pp [7] M. H. Ammar, D. Chakrabarty, A. D. Sarma, S. Kalyanasundaram and R. J. Lpton, Algorthms for Message Ferryng on Moble Ad Hoc Networks, Proceedngs of the IARCS Annual Conference on Foundatons of Software echnology and heoretcal Computer Scence (FSCS 09), II Kanpur, 2009, pp [8] Q. L and D. Rus, Communcaton n Dsconnected Ad Hoc Networks Usng Message Relay, Journal of Parallel and Dstrbuted Computng, Vol. 63, No. 1, 2003, pp [9] S.-H. Chung, K.-F. Ssu, C.-H. Chou, and H. C. Jau, Improvng Data ransmsson wth Helpng Nodes for Geographcal Ad Hoc Routng, Computer Networks, Vol. 51, 2007, pp [10] C.-H. Ou, K.-F. Ssu and H. C. Jau, Connectng Network Parttons wth Locaton-Asssted Forwardng Nodes n Moble Ad Hoc Envroents, Proceedngs of the 10th IEEE Pacfc Rm Internatonal Symposum on Dependable Computng (PRDC 04), Washngton, DC, 2004, pp [11] D. Borsett, C. Casett, C.-F. Chassern, M. Fore and J. M. Barcel o-ordnas, Vrtual Data Mules for Data Col- Copyrght 2010 ScRes.

An Optimal Algorithm for Prufer Codes *

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

More information

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

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

More information

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

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

More information

Parallelism for Nested Loops with Non-uniform and Flow Dependences

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

More information

Load Balancing for Hex-Cell Interconnection Network

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

More information

S1 Note. Basis functions.

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

More information

Reducing Frame Rate for Object Tracking

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

More information

Cluster Analysis of Electrical Behavior

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

More information

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

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

More information

Efficient Distributed File System (EDFS)

Efficient Distributed File System (EDFS) Effcent Dstrbuted Fle System (EDFS) (Sem-Centralzed) Debessay(Debsh) Fesehaye, Rahul Malk & Klara Naherstedt Unversty of Illnos-Urbana Champagn Contents Problem Statement, Related Work, EDFS Desgn Rate

More information

MobileGrid: Capacity-aware Topology Control in Mobile Ad Hoc Networks

MobileGrid: Capacity-aware Topology Control in Mobile Ad Hoc Networks MobleGrd: Capacty-aware Topology Control n Moble Ad Hoc Networks Jle Lu, Baochun L Department of Electrcal and Computer Engneerng Unversty of Toronto {jenne,bl}@eecg.toronto.edu Abstract Snce wreless moble

More information

DESIGNING TRANSMISSION SCHEDULES FOR WIRELESS AD HOC NETWORKS TO MAXIMIZE NETWORK THROUGHPUT

DESIGNING TRANSMISSION SCHEDULES FOR WIRELESS AD HOC NETWORKS TO MAXIMIZE NETWORK THROUGHPUT DESIGNING TRANSMISSION SCHEDULES FOR WIRELESS AD HOC NETWORKS TO MAXIMIZE NETWORK THROUGHPUT Bran J. Wolf, Joseph L. Hammond, and Harlan B. Russell Dept. of Electrcal and Computer Engneerng, Clemson Unversty,

More information

Real-Time Guarantees. Traffic Characteristics. Flow Control

Real-Time Guarantees. Traffic Characteristics. Flow Control Real-Tme Guarantees Requrements on RT communcaton protocols: delay (response s) small jtter small throughput hgh error detecton at recever (and sender) small error detecton latency no thrashng under peak

More information

Load-Balanced Anycast Routing

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

More information

Related-Mode Attacks on CTR Encryption Mode

Related-Mode Attacks on CTR Encryption Mode Internatonal Journal of Network Securty, Vol.4, No.3, PP.282 287, May 2007 282 Related-Mode Attacks on CTR Encrypton Mode Dayn Wang, Dongda Ln, and Wenlng Wu (Correspondng author: Dayn Wang) Key Laboratory

More information

DEAR: A DEVICE AND ENERGY AWARE ROUTING PROTOCOL FOR MOBILE AD HOC NETWORKS

DEAR: A DEVICE AND ENERGY AWARE ROUTING PROTOCOL FOR MOBILE AD HOC NETWORKS DEAR: A DEVICE AND ENERGY AWARE ROUTING PROTOCOL FOR MOBILE AD HOC NETWORKS Arun Avudanayagam Yuguang Fang Wenjng Lou Department of Electrcal and Computer Engneerng Unversty of Florda Ganesvlle, FL 3261

More information

Simulation Based Analysis of FAST TCP using OMNET++

Simulation Based Analysis of FAST TCP using OMNET++ Smulaton Based Analyss of FAST TCP usng OMNET++ Umar ul Hassan 04030038@lums.edu.pk Md Term Report CS678 Topcs n Internet Research Sprng, 2006 Introducton Internet traffc s doublng roughly every 3 months

More information

Routing in Degree-constrained FSO Mesh Networks

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

More information

CMPS 10 Introduction to Computer Science Lecture Notes

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

More information

Constructing Minimum Connected Dominating Set: Algorithmic approach

Constructing Minimum Connected Dominating Set: Algorithmic approach Constructng Mnmum Connected Domnatng Set: Algorthmc approach G.N. Puroht and Usha Sharma Centre for Mathematcal Scences, Banasthal Unversty, Rajasthan 304022 usha.sharma94@yahoo.com Abstract: Connected

More information

A Load-balancing and Energy-aware Clustering Algorithm in Wireless Ad-hoc Networks

A Load-balancing and Energy-aware Clustering Algorithm in Wireless Ad-hoc Networks A Load-balancng and Energy-aware Clusterng Algorthm n Wreless Ad-hoc Networks Wang Jn, Shu Le, Jnsung Cho, Young-Koo Lee, Sungyoung Lee, Yonl Zhong Department of Computer Engneerng Kyung Hee Unversty,

More information

Evaluation of an Enhanced Scheme for High-level Nested Network Mobility

Evaluation of an Enhanced Scheme for High-level Nested Network Mobility IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.15 No.10, October 2015 1 Evaluaton of an Enhanced Scheme for Hgh-level Nested Network Moblty Mohammed Babker Al Mohammed, Asha Hassan.

More information

A Binarization Algorithm specialized on Document Images and Photos

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

More information

Virtual Memory. Background. No. 10. Virtual Memory: concept. Logical Memory Space (review) Demand Paging(1) Virtual Memory

Virtual Memory. Background. No. 10. Virtual Memory: concept. Logical Memory Space (review) Demand Paging(1) Virtual Memory Background EECS. Operatng System Fundamentals No. Vrtual Memory Prof. Hu Jang Department of Electrcal Engneerng and Computer Scence, York Unversty Memory-management methods normally requres the entre process

More information

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS

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

More information

Video Proxy System for a Large-scale VOD System (DINA)

Video Proxy System for a Large-scale VOD System (DINA) Vdeo Proxy System for a Large-scale VOD System (DINA) KWUN-CHUNG CHAN #, KWOK-WAI CHEUNG *# #Department of Informaton Engneerng *Centre of Innovaton and Technology The Chnese Unversty of Hong Kong SHATIN,

More information

Mathematics 256 a course in differential equations for engineering students

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

More information

Module Management Tool in Software Development Organizations

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

More information

Analysis of Continuous Beams in General

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

More information

Optimizing Energy-Latency Trade-off in Sensor Networks with Controlled Mobility

Optimizing Energy-Latency Trade-off in Sensor Networks with Controlled Mobility Optmzng Energy-Latency Trade-off n Sensor Networks wth Controlled Moblty Ryo Sughara Rajesh K. Gupta Computer Scence and Engneerng Department Unversty of Calforna, San Dego La Jolla, Calforna 9293 Emal:

More information

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

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

More information

Wishing you all a Total Quality New Year!

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

More information

General Share-A-Ride Problem

General Share-A-Ride Problem General Share-A-Rde Problem Sesya Sr Purwant Department of Industral Engneerng, Insttut Tenolog Bandung, Bandung, Indonesa Department of Industral Management, Natonal Tawan Unversty of Scence and Technology,

More information

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

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

More information

Private Information Retrieval (PIR)

Private Information Retrieval (PIR) 2 Levente Buttyán Problem formulaton Alce wants to obtan nformaton from a database, but she does not want the database to learn whch nformaton she wanted e.g., Alce s an nvestor queryng a stock-market

More information

Assignment # 2. Farrukh Jabeen Algorithms 510 Assignment #2 Due Date: June 15, 2009.

Assignment # 2. Farrukh Jabeen Algorithms 510 Assignment #2 Due Date: June 15, 2009. Farrukh Jabeen Algorthms 51 Assgnment #2 Due Date: June 15, 29. Assgnment # 2 Chapter 3 Dscrete Fourer Transforms Implement the FFT for the DFT. Descrbed n sectons 3.1 and 3.2. Delverables: 1. Concse descrpton

More information

Network Coding as a Dynamical System

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

More information

AADL : about scheduling analysis

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

More information

Online Policies for Opportunistic Virtual MISO Routing in Wireless Ad Hoc Networks

Online Policies for Opportunistic Virtual MISO Routing in Wireless Ad Hoc Networks 12 IEEE Wreless Communcatons and Networkng Conference: Moble and Wreless Networks Onlne Polces for Opportunstc Vrtual MISO Routng n Wreless Ad Hoc Networks Crstano Tapparello, Stefano Tomasn and Mchele

More information

Dynamic Bandwidth Provisioning with Fairness and Revenue Considerations for Broadband Wireless Communication

Dynamic Bandwidth Provisioning with Fairness and Revenue Considerations for Broadband Wireless Communication Ths full text paper was peer revewed at the drecton of IEEE Communcatons Socety subject matter experts for publcaton n the ICC 008 proceedngs. Dynamc Bandwdth Provsonng wth Farness and Revenue Consderatons

More information

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

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

More information

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

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

More information

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration Improvement of Spatal Resoluton Usng BlockMatchng Based Moton Estmaton and Frame Integraton Danya Suga and Takayuk Hamamoto Graduate School of Engneerng, Tokyo Unversty of Scence, 6-3-1, Nuku, Katsuska-ku,

More information

CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION

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

More information

A Low-Overhead Routing Protocol for Ad Hoc Networks with selfish nodes

A Low-Overhead Routing Protocol for Ad Hoc Networks with selfish nodes A Low-Oerhead Routng Protocol for Ad Hoc Networks wth selfsh nodes Dongbn Wang 1, Xaofeng Wang 2, Xangzhan Yu 3, Kacheng Q 1, Zhbn Xa 1 1 School of Software Engneerng, Bejng Unersty of Posts and Telecommuncatons,100876,

More information

y and the total sum of

y and the total sum of Lnear regresson Testng for non-lnearty In analytcal chemstry, lnear regresson s commonly used n the constructon of calbraton functons requred for analytcal technques such as gas chromatography, atomc absorpton

More information

On the Exact Analysis of Bluetooth Scheduling Algorithms

On the Exact Analysis of Bluetooth Scheduling Algorithms On the Exact Analyss of Bluetooth Schedulng Algorth Gl Zussman Dept. of Electrcal Engneerng Technon IIT Hafa 3000, Israel glz@tx.technon.ac.l Ur Yechal Dept. of Statstcs and Operatons Research School of

More information

Extending Network Life by Using Mobile Actors in Cluster-based Wireless Sensor and Actor Networks

Extending Network Life by Using Mobile Actors in Cluster-based Wireless Sensor and Actor Networks Extendng Networ Lfe by Usng Moble Actors n Cluster-based Wreless Sensor and Actor Networs Nauman Aslam, Wllam Phllps, Wllam Robertson and S. Svaumar Department of Engneerng Mathematcs & Internetworng Dalhouse

More information

Sorting Review. Sorting. Comparison Sorting. CSE 680 Prof. Roger Crawfis. Assumptions

Sorting Review. Sorting. Comparison Sorting. CSE 680 Prof. Roger Crawfis. Assumptions Sortng Revew Introducton to Algorthms Qucksort CSE 680 Prof. Roger Crawfs Inserton Sort T(n) = Θ(n 2 ) In-place Merge Sort T(n) = Θ(n lg(n)) Not n-place Selecton Sort (from homework) T(n) = Θ(n 2 ) In-place

More information

3. CR parameters and Multi-Objective Fitness Function

3. CR parameters and Multi-Objective Fitness Function 3 CR parameters and Mult-objectve Ftness Functon 41 3. CR parameters and Mult-Objectve Ftness Functon 3.1. Introducton Cogntve rados dynamcally confgure the wreless communcaton system, whch takes beneft

More information

Report on On-line Graph Coloring

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

More information

Learning the Kernel Parameters in Kernel Minimum Distance Classifier

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

More information

Explicit Formulas and Efficient Algorithm for Moment Computation of Coupled RC Trees with Lumped and Distributed Elements

Explicit Formulas and Efficient Algorithm for Moment Computation of Coupled RC Trees with Lumped and Distributed Elements Explct Formulas and Effcent Algorthm for Moment Computaton of Coupled RC Trees wth Lumped and Dstrbuted Elements Qngan Yu and Ernest S.Kuh Electroncs Research Lab. Unv. of Calforna at Berkeley Berkeley

More information

IP mobility support is becoming very important as the

IP mobility support is becoming very important as the 706 IEEE TRANSACTIONS ON COMPUTERS, VOL. 52, NO. 6, JUNE 2003 A New Scheme for Reducng Lnk and Sgnalng Costs n Moble IP Young J. Lee and Ian F. Akyldz, Fellow, IEEE Abstract IP moblty support s provded

More information

Research of Dynamic Access to Cloud Database Based on Improved Pheromone Algorithm

Research of Dynamic Access to Cloud Database Based on Improved Pheromone Algorithm , pp.197-202 http://dx.do.org/10.14257/dta.2016.9.5.20 Research of Dynamc Access to Cloud Database Based on Improved Pheromone Algorthm Yongqang L 1 and Jn Pan 2 1 (Software Technology Vocatonal College,

More information

F Geometric Mean Graphs

F Geometric Mean Graphs Avalable at http://pvamu.edu/aam Appl. Appl. Math. ISSN: 1932-9466 Vol. 10, Issue 2 (December 2015), pp. 937-952 Applcatons and Appled Mathematcs: An Internatonal Journal (AAM) F Geometrc Mean Graphs A.

More information

RAP. Speed/RAP/CODA. Real-time Systems. Modeling the sensor networks. Real-time Systems. Modeling the sensor networks. Real-time systems:

RAP. Speed/RAP/CODA. Real-time Systems. Modeling the sensor networks. Real-time Systems. Modeling the sensor networks. Real-time systems: Speed/RAP/CODA Presented by Octav Chpara Real-tme Systems Many wreless sensor network applcatons requre real-tme support Survellance and trackng Border patrol Fre fghtng Real-tme systems: Hard real-tme:

More information

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

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

More information

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data A Fast Content-Based Multmeda Retreval Technque Usng Compressed Data Borko Furht and Pornvt Saksobhavvat NSF Multmeda Laboratory Florda Atlantc Unversty, Boca Raton, Florda 3343 ABSTRACT In ths paper,

More information

Fast Retransmission of Real-Time Traffic in HIPERLAN/2 Systems

Fast Retransmission of Real-Time Traffic in HIPERLAN/2 Systems Fast Retransmsson of Real-Tme Traffc n HIPERLAN/ Systems José A Afonso and Joaqum E Neves Department of Industral Electroncs Unversty of Mnho, Campus de Azurém 4800-058 Gumarães, Portugal {joseafonso,

More information

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

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

More information

Towards Mobility as a Network Control Primitive

Towards Mobility as a Network Control Primitive Towards Moblty as a Network Control Prmtve Davd K. Goldenberg Dept. Comp. Sc. Yale Unversty New Haven, CT 6517 davd.goldenberg@yale.edu Brad E. Rosen Dept. Comp. Sc. Yale Unversty New Haven, CT 6517 brad.rosen@yale.edu

More information

IJCTA Nov-Dec 2016 Available

IJCTA Nov-Dec 2016 Available Dr K Santh et al, Internatonal Journal of Computer Technology & Applcatons,Vol 7(6),773-779 Optmzed Route Technque for DSR Routng Protocol n MANET Dr.K.Santh, Assocate Professor, Dept. of Computer Scence,

More information

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision

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

More information

The Codesign Challenge

The Codesign Challenge ECE 4530 Codesgn Challenge Fall 2007 Hardware/Software Codesgn The Codesgn Challenge Objectves In the codesgn challenge, your task s to accelerate a gven software reference mplementaton as fast as possble.

More information

Shared Running Buffer Based Proxy Caching of Streaming Sessions

Shared Running Buffer Based Proxy Caching of Streaming Sessions Shared Runnng Buffer Based Proxy Cachng of Streamng Sessons Songqng Chen, Bo Shen, Yong Yan, Sujoy Basu Moble and Meda Systems Laboratory HP Laboratores Palo Alto HPL-23-47 March th, 23* E-mal: sqchen@cs.wm.edu,

More information

Using Particle Swarm Optimization for Enhancing the Hierarchical Cell Relay Routing Protocol

Using Particle Swarm Optimization for Enhancing the Hierarchical Cell Relay Routing Protocol 2012 Thrd Internatonal Conference on Networkng and Computng Usng Partcle Swarm Optmzaton for Enhancng the Herarchcal Cell Relay Routng Protocol Hung-Y Ch Department of Electrcal Engneerng Natonal Sun Yat-Sen

More information

A fair buffer allocation scheme

A fair buffer allocation scheme A far buffer allocaton scheme Juha Henanen and Kalev Klkk Telecom Fnland P.O. Box 228, SF-330 Tampere, Fnland E-mal: juha.henanen@tele.f Abstract An approprate servce for data traffc n ATM networks requres

More information

Quantifying Performance Models

Quantifying Performance Models Quantfyng Performance Models Prof. Danel A. Menascé Department of Computer Scence George Mason Unversty www.cs.gmu.edu/faculty/menasce.html 1 Copyrght Notce Most of the fgures n ths set of sldes come from

More information

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

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

More information

Feature Reduction and Selection

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

More information

Meta-heuristics for Multidimensional Knapsack Problems

Meta-heuristics for Multidimensional Knapsack Problems 2012 4th Internatonal Conference on Computer Research and Development IPCSIT vol.39 (2012) (2012) IACSIT Press, Sngapore Meta-heurstcs for Multdmensonal Knapsack Problems Zhbao Man + Computer Scence Department,

More information

Quantifying Responsiveness of TCP Aggregates by Using Direct Sequence Spread Spectrum CDMA and Its Application in Congestion Control

Quantifying Responsiveness of TCP Aggregates by Using Direct Sequence Spread Spectrum CDMA and Its Application in Congestion Control Quantfyng Responsveness of TCP Aggregates by Usng Drect Sequence Spread Spectrum CDMA and Its Applcaton n Congeston Control Mehd Kalantar Department of Electrcal and Computer Engneerng Unversty of Maryland,

More information

CSCI 104 Sorting Algorithms. Mark Redekopp David Kempe

CSCI 104 Sorting Algorithms. Mark Redekopp David Kempe CSCI 104 Sortng Algorthms Mark Redekopp Davd Kempe Algorthm Effcency SORTING 2 Sortng If we have an unordered lst, sequental search becomes our only choce If we wll perform a lot of searches t may be benefcal

More information

Performance Improvement of Direct Diffusion Algorithm in Sensor Networks

Performance Improvement of Direct Diffusion Algorithm in Sensor Networks Mddle-East Journal of Scentfc Research 2 (): 566-574, 202 ISSN 990-9233 IDOSI Publcatons, 202 DOI: 0.5829/dos.mejsr.202.2..43 Performance Improvement of Drect Dffuson Algorthm n Sensor Networks Akbar Bemana

More information

Parallel matrix-vector multiplication

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

More information

Avoiding congestion through dynamic load control

Avoiding congestion through dynamic load control Avodng congeston through dynamc load control Vasl Hnatyshn, Adarshpal S. Seth Department of Computer and Informaton Scences, Unversty of Delaware, Newark, DE 976 ABSTRACT The current best effort approach

More information

Adaptive Energy and Location Aware Routing in Wireless Sensor Network

Adaptive Energy and Location Aware Routing in Wireless Sensor Network Adaptve Energy and Locaton Aware Routng n Wreless Sensor Network Hong Fu 1,1, Xaomng Wang 1, Yngshu L 1 Department of Computer Scence, Shaanx Normal Unversty, X an, Chna, 71006 fuhong433@gmal.com {wangxmsnnu@hotmal.cn}

More information

Lecture 5: Multilayer Perceptrons

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

More information

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

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

More information

SRB: Shared Running Buffers in Proxy to Exploit Memory Locality of Multiple Streaming Media Sessions

SRB: Shared Running Buffers in Proxy to Exploit Memory Locality of Multiple Streaming Media Sessions SRB: Shared Runnng Buffers n Proxy to Explot Memory Localty of Multple Streamng Meda Sessons Songqng Chen,BoShen, Yong Yan, Sujoy Basu, and Xaodong Zhang Department of Computer Scence Moble and Meda System

More information

An Efficient Genetic Algorithm with Fuzzy c-means Clustering for Traveling Salesman Problem

An Efficient Genetic Algorithm with Fuzzy c-means Clustering for Traveling Salesman Problem An Effcent Genetc Algorthm wth Fuzzy c-means Clusterng for Travelng Salesman Problem Jong-Won Yoon and Sung-Bae Cho Dept. of Computer Scence Yonse Unversty Seoul, Korea jwyoon@sclab.yonse.ac.r, sbcho@cs.yonse.ac.r

More information

Mobility Based Routing Protocol with MAC Collision Improvement in Vehicular Ad Hoc Networks

Mobility Based Routing Protocol with MAC Collision Improvement in Vehicular Ad Hoc Networks Moblty Based Routng Protocol wth MAC Collson Improvement n Vehcular Ad Hoc Networks Zhhao Dng, Pny Ren, Qnghe Du Shaanx Smart Networks and Ubqutous Access Rearch Center School of Electronc and Informaton

More information

USING GRAPHING SKILLS

USING GRAPHING SKILLS Name: BOLOGY: Date: _ Class: USNG GRAPHNG SKLLS NTRODUCTON: Recorded data can be plotted on a graph. A graph s a pctoral representaton of nformaton recorded n a data table. t s used to show a relatonshp

More information

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

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

More information

Connection-information-based connection rerouting for connection-oriented mobile communication networks

Connection-information-based connection rerouting for connection-oriented mobile communication networks Dstrb. Syst. Engng 5 (1998) 47 65. Prnted n the UK PII: S0967-1846(98)90513-7 Connecton-nformaton-based connecton reroutng for connecton-orented moble communcaton networks Mnho Song, Yanghee Cho and Chongsang

More information

Distributed Grid based Robust Clustering Protocol for Mobile Sensor Networks

Distributed Grid based Robust Clustering Protocol for Mobile Sensor Networks 414 The Internatonal Arab Journal of Informaton Technology, Vol. 8, No. 4, October 011 Dstrbuted Grd based Robust Clusterng Protocol for Moble Sensor Networks Shahzad Al and Sajjad Madan Department of

More information

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers IOSR Journal of Electroncs and Communcaton Engneerng (IOSR-JECE) e-issn: 78-834,p- ISSN: 78-8735.Volume 9, Issue, Ver. IV (Mar - Apr. 04), PP 0-07 Content Based Image Retreval Usng -D Dscrete Wavelet wth

More information

sensors ISSN

sensors ISSN Sensors 2011, 11, 5383-5401; do:10.3390/s110505383 OPEN ACCESS sensors ISSN 1424-8220 www.mdp.com/journal/sensors Artcle FRCA: A Fuzzy Relevance-Based Cluster Head Selecton Algorthm for Wreless Moble Ad-Hoc

More information

Support Vector Machines

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

More information

Efficient Content Distribution in Wireless P2P Networks

Efficient Content Distribution in Wireless P2P Networks Effcent Content Dstrbuton n Wreless P2P Networs Qong Sun, Vctor O. K. L, and Ka-Cheong Leung Department of Electrcal and Electronc Engneerng The Unversty of Hong Kong Pofulam Road, Hong Kong, Chna {oansun,

More information

The Hybrid Mobile Wireless Sensor Networks for Data Gathering

The Hybrid Mobile Wireless Sensor Networks for Data Gathering The Hybrd Moble Wreless Sensor Networks for Data Gatherng Bao Ren Bejng Unv. of Posts & Telecommuncatons, Bejng 876, Chna +86--62287 renb@bupt.edu.cn Jan Ma Noka Research Center, Bejng 3, Chna +86--65392828-2883

More information

Problem Set 3 Solutions

Problem Set 3 Solutions Introducton to Algorthms October 4, 2002 Massachusetts Insttute of Technology 6046J/18410J Professors Erk Demane and Shaf Goldwasser Handout 14 Problem Set 3 Solutons (Exercses were not to be turned n,

More information

An Entropy-Based Approach to Integrated Information Needs Assessment

An Entropy-Based Approach to Integrated Information Needs Assessment Dstrbuton Statement A: Approved for publc release; dstrbuton s unlmted. An Entropy-Based Approach to ntegrated nformaton Needs Assessment June 8, 2004 Wllam J. Farrell Lockheed Martn Advanced Technology

More information

Analysis of Collaborative Distributed Admission Control in x Networks

Analysis of Collaborative Distributed Admission Control in x Networks 1 Analyss of Collaboratve Dstrbuted Admsson Control n 82.11x Networks Thnh Nguyen, Member, IEEE, Ken Nguyen, Member, IEEE, Lnha He, Member, IEEE, Abstract Wth the recent surge of wreless home networks,

More information

Efficient Load-Balanced IP Routing Scheme Based on Shortest Paths in Hose Model. Eiji Oki May 28, 2009 The University of Electro-Communications

Efficient Load-Balanced IP Routing Scheme Based on Shortest Paths in Hose Model. Eiji Oki May 28, 2009 The University of Electro-Communications Effcent Loa-Balance IP Routng Scheme Base on Shortest Paths n Hose Moel E Ok May 28, 2009 The Unversty of Electro-Communcatons Ok Lab. Semnar, May 28, 2009 1 Outlne Backgroun on IP routng IP routng strategy

More information

Computer Communications

Computer Communications Computer Communcatons 3 (22) 3 48 Contents lsts avalable at ScVerse ScenceDrect Computer Communcatons journal homepage: www.elsever.com/locate/comcom On the queueng behavor of nter-flow asynchronous network

More information

Distributed Resource Scheduling in Grid Computing Using Fuzzy Approach

Distributed Resource Scheduling in Grid Computing Using Fuzzy Approach Dstrbuted Resource Schedulng n Grd Computng Usng Fuzzy Approach Shahram Amn, Mohammad Ahmad Computer Engneerng Department Islamc Azad Unversty branch Mahallat, Iran Islamc Azad Unversty branch khomen,

More information

OPTIMAL CONFIGURATION FOR NODES IN MIXED CELLULAR AND MOBILE AD HOC NETWORK FOR INET

OPTIMAL CONFIGURATION FOR NODES IN MIXED CELLULAR AND MOBILE AD HOC NETWORK FOR INET OPTIMAL CONFIGURATION FOR NODE IN MIED CELLULAR AND MOBILE AD HOC NETWORK FOR INET Olusola Babalola D.E. Department of Electrcal and Computer Engneerng Morgan tate Unversty Dr. Rchard Dean Faculty Advsor

More information

GSLM Operations Research II Fall 13/14

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

More information