SOCIAL networking is among the fastest growing information

Size: px
Start display at page:

Download "SOCIAL networking is among the fastest growing information"

Transcription

1 IEEE TRANSACTIONS ON PARALLEL AND DISTRIBTED SYSTEMS, VOL. 26, NO. 5, MAY Effent Data Query n Intermttently-Conneted Moble Ad Ho Soal Networks Yang Lu, Student Member, IEEE, Yanyan Han, Zhpeng Yang, Student Member, IEEE, and Hongy Wu, Member, IEEE Abstrat Ths work addresses the problem of how to enable effent data query n a Moble Ad-ho SOal Network (MASON), formed by moble users who share smlar nterests and onnet wth one another by explotng Bluetooth and/or WF onnetons. The data query n MASONs faes several unque hallenges nludng opportunst lnk onnetvty, autonomous omputng and storage, and unknown or naurate data provders. Our goal s to determne an optmal transmsson strategy that supports the desred query rate wthn a delay budget and at the same tme mnmzes the total ommunaton ost. To ths end, we propose a entralzed optmzaton model that offers useful theoret nsghts and develop a dstrbuted data query protool for pratal applatons. To demonstrate the feasblty and effeny of the proposed sheme and to gan useful empral nsghts, we arry out a testbed experment by usng 25 off-the-shelf Dell Streak tablets for a perod of 15 days. Moreover, extensve smulatons are arred out to learn the performane trend under varous network settngs, whh are not pratal to buld and evaluate n laboratores. Index Terms Data query, moble ad ho soal networks, entralzed optmzaton model, dstrbuted protool, testbed experment, smulatons Ç 1 INTRODCTION SOCIAL networkng s among the fastest growng nformaton tehnologes, as evdened by the popularty of suh onlne soal network stes as Faebook, Twtter, LnkedIn and Google+ that ontnue to experene explosve growth. In ontrast to the popular web-based onlne soal networks that rely on the Internet nfrastruture (nludng ellular systems) for ommunaton, ths paper fouses on Moble Ad-ho SOal Network (MASON), an autonomous soal network formed by moble users who share smlar nterests and onnet wth one another by explotng the Bluetooth and/or WF onnetons of ther moble phones or portable tablets. A MASON s often reated for a loal ommunty where the partpants have frequent nteratons, e.g., people lvng n an urban neghborhood, students studyng n a ollege, or toursts vstng an arhaeologal ste. Its sze vares from a large group (for nstane, all the students n a unversty) to a small luster (suh as members of a shool band). It may serve a ommunty over a long span of years, or be temporary to last for as short as a few hours only (e.g., for soal networkng among a group of toursts). 1.1 System Overvew An ndvdual MASON s nomparable wth onlne soal networks n terms of the populaton of partpants, the The authors are wth the Center for Advaned Computer Studes, nversty of Lousana at Lafayette, Lafayette, LA E-mal: {yxl0782, wu}@as.lousana.edu, yanyan.ull@gmal.om, zhpengyang@hotmal.om. Manusrpt reeved 23 Jan. 2014; revsed 21 Apr. 2014; aepted 23 Apr Date of publaton 28 Apr. 2014; date of urrent verson 8 Apr Reommended for aeptane by J. Lloret. For nformaton on obtanng reprnts of ths artle, please send e-mal to: reprnts@eee.org, and referene the Dgtal Objet Identfer below. Dgtal Objet Identfer no /TPDS number of soal onnetons and the amount of soal meda. However MASONs gan sgnfant value by servng as a supplement and augment to onlne soal networks and by effetvely supportng loal ommunty-based ad-ho soal networkng. For example, t helps dsover and update soal lnks that are not aptured by onlne soal networks and allows a user to query loalzed data suh as loal knowledge, ontats and expertse, surroundng news and photos, or other nformaton that people usually annot or do not bother to report to onlne webstes but may temporarly keep on ther portable deves or generate upon a request. Ths work addresses the problem of how to enable effent data query n MASONs. Consder a MASON wth N nodes. Eah node an be a query ssuer or a data provder, or more ommonly at n both roles for dfferent query requests. The queres fall nto C ategores. Eah node has ertan expertse to answer a query. Let E denote the expertse matrx, where E ndates the expertse of Node to answer a query n Category,.e., the probablty that Node an provde a satsfatory answer to a query n Category. A query s reated by a query ssuer. It s delvered by the network toward the nodes that an suessfully provde an answer (.e., data provders). If a data provder reeves the query, t sends the query reply to the query ssuer. Our goal s to determne an optmal transmsson strategy that supports the desred query rate and at the same tme mnmzes the total ommunaton ost. The formal problem formulaton wll be gven n Seton nque Challenges The use of free, short-range rado s hghly desred for a dversty of MASON applatons. At the same tme, however, t results n a dstntve ommunaton paradgm haraterzed by ntermttent lnk onnetvty and ß 2014 IEEE. Personal use s permtted, but republaton/redstrbuton requres IEEE permsson. See for more nformaton.

2 1302 IEEE TRANSACTIONS ON PARALLEL AND DISTRIBTED SYSTEMS, VOL. 26, NO. 5, MAY 2015 autonomous omputng and storage. More spefally, the data query n MASONs faes the followng unque hallenges. (1) Opportunst lnk onnetvty. The onnetvty of MASONs s very low and ntermttent, formng a sparse network where a node s onneted to other nodes only oasonally. Ths s n a sharp ontrast to onlne soal networks, where users always have relable Internet onnetons. The data delvery delay n MASONs s potentally long, due to the loose onnetvty among nodes. Fortunately, suh delay, thought not desrable, s usually tolerable by many data query applatons n MASONs that are often delay nsenstve. (2) Autonomous omputng and storage. Central servers are employed to store and proess user data n onlne soal networks. Suh servers are, however, no longer avalable n MASONs, where ndvdual portable deves must perform dstrbuted data storage and omputaton. It s well known that portable deves have lmted omputng, storage and energy apaty. Nevertheless, suh onstrants are partularly dsadvantageous to MASONs, beause a node must proess data n a dstrbuted manner and store them loally for a muh longer tme before sendng them to another node, due to ntermttent onnetvty. (3) nknown or naurate expertse. When a node ssues a query, t s often unaware of the nodes that have suffent expertse to answer the query, sne the ost s prohbtvely hgh to onstrut a struture to ndex data and data provders lke P2P networks. It s obvously neffent ether to frequently flood queres, whh are expensve and often onsdered spams. Worse yet, n prate, a moble node hardly knows ts probablty to answer queres n eah ategory presely. It may ntally lam ts expertse based on the moble user s soal roles and avalable resoures. But suh ntally lamed expertse s often naurate. The above haratersts make the data query n MASONs a very unque, nterestng, and hallengng problem, renderng not only onventonal data query shemes for well-onneted omputer systems but also dstrbuted solutons for moble ad ho networks [1], [2], [3], [4], [5], [6] and moble (onlne) soal networks [7] napplable here. Only a handful of works have onsdered data query n opportunst network settngs. For example, Osmoss [8] employs an epdem approah to perform fle lookup n poket swthed networks. Whle t s smple and relable, the ommunaton overhead s very hgh due to the floodng-lke epdem routng. DelQue [9] ams to query geoloaton-based nformaton. It assumes eah node moves aordng to a gven shedule and adopts a sem-markov model to predt nodal meetng events, n order to dentfy a proper relay to arry the query to the target loaton and brng the nterested nformaton bak to the soure. Yang et al. [10] proposes a dstrbuted database query framework based on several ommunaton and omputng tehnques spefally talored for RFID networks. Nether of them effently support data queres n MASONs. On the other hand, although several routng algorthms have been proposed for opportunst networks by explotng soal relatons among moble users to aheve effent routng [11], [12], [13], [14], [15], [16], they are developed for gener ommunatons, wthout onsderaton of the unque needs and onstrants n data query. Among them, Zhu et al. [16] s the most reent one, where explots a dstrbuted ommunty parttonng algorthm to dvde a DTN nto smaller ommuntes. For ntra-ommunty ommunaton, a utlty funton onvolutng soal smlarty and soal entralty wth a deay fator s used to hoose relay nodes. For nterommunty ommunaton, the nodes movng frequently aross ommuntes are hosen as relays to arry data to destnaton effently. Although Zhu et al. [16] ntrodues a soluton for DTNs whh leverages soal propertes and moblty haratersts of users, t s not truly applable for the data query n MASONs, beause t does not apture the nherent features for the query delvery n MASONs, hene the nodes are not helpful for eah other by makng the orret desons to arry queres to satsfatory nodes. 1.3 Our Contrbuton In ths work, we frst formulate the optmzaton problem for the data query n MASONs. Then, we develop a statedagram-based analyt model to derve the ommunaton overhead and query rate as the funtons of transmsson strategy. A branh and bound algorthm s adopted to dsover the optmal transmsson strategy that mnmzes the total ommunaton ost whle ahevng a target query rate. The optmzaton model s entralzed, thus unpratal for real world mplementaton. However, t offers useful nsghts for the development of a dstrbuted data query protool. The proposed protool s based on two key tehnques. Frst, as motvated by the analyss, t employs reahable expertse as the routng metr to gude the transmsson of query requests. Seond, t explots the redundany n query transmsson, whh an effetvely mprove the query delvery rate n prate f t s properly ontrolled. To demonstrate the feasblty and effeny of the proposed data query protool and to gan useful empral nsghts, we have arred out a testbed experment usng off-the-shelf Dell Streak tablets. Our experment nvolves 25 volunteers and lasts for 15 days. Moreover, extensve smulatons (based on odes extrated from our prototype mplementaton) are arred out to learn the performane trend under varous network settngs, whh are not pratal to buld and evaluate n laboratores. The remander of ths paper s organzed as follows. Seton 2 ntrodues the proposed data query sheme, nludng a theoret optmzaton model and a pratal protool. Seton 3 presents a testbed experment and results. Seton 4 dsusses large-sale smulatons under real-world moblty traes and power-law moblty model. Fnally, Seton 5 onludes the paper. 2 PROPOSED DATA QERY SCHEME Whle MASONs offer nterestng opportuntes to support ad ho data query, ts apaty s unsurprsngly low ompared to many other data networks due to ts extremely lmted and nondetermnst ommunaton opportuntes. To learn the essene of optmal query delvery and to understand the performane upper bound, we frst arry out a prelmnary analyt study of the data query n MASONs before movng nto the detaled protool desgn and

3 LI ET AL.: EFFICIENT DATA QERY IN INTERMITTENTLY-CONNECTED MOBILE AD HOC SOCIAL NETWORKS 1303 evaluaton. Based on the nsghts ganed from our analyss, a dstrbuted data query protool s proposed, amng to enable hghly effent ad ho query under pratal MASON settngs. 2.1 Prelmnary Analyss In ths seton, we frst formulate the problem, and then ntrodue an analyt model to obtan an optmal soluton, followed by dsussons on numer results and nsghts for developng a pratal dstrbuted data query protool Assumptons and Problem Formulaton The ommunaton n MASONs depends on nodal meetng events. The tme nterval between two onseutve meetng events between two nodes, e.g., Nodes and j, s denoted by a random varable T j. We onsder gener nodal moblty n ths work wthout assumng any spef dstrbuton of T j. 1 For analyt tratablty, we make the followng two assumptons n data transmsson. Assumpton 1. Gven the ntermttent network settng, we assume that the ommunaton apaty s ruled by nodal meetng opportuntes. In other words, we assume the ommunaton delay s domnated by nodal meetng ntervals. When two nodes meet, the hannel bandwdth s suffent for them to exhange data pakets wth neglgble delay. Assumpton 2. Multple opes of a query request may exst n the network, but we assume a node reeves and forwards the same request only one. For a query request, we defne a bnary transmsson matrx, X, where X j ¼ 1 ndates that Node sends the request to Node j, f the former holds the request when t meets the latter. If X j ¼ 0, then Node does not send the request to Node j even f t arres the request upon meetng Node j. Note that X j and X j an be dfferent. The transmsson of query reply follows standard oneto-one routng wth known soure and destnaton. Its expeted transmsson delay has been well studed [11], [12], [13], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37]. Therefore we fous on the delvery of query requests only n ths analyss (but a omplete protool s desgned and mplemented as to be dsussed n later setons). Our goal s to determne the optmal X suh that the total ommunaton ost, defned as CðXÞ, s mnmzed, whle at the same tme the probablty to delver the query to a data provder wthn a gven delay budget d (denoted as P ðxþ) reahes the desred threshold, b. More spefally, the problem s formulated as follows: Objetve : S:t: : mnðcðxþþ; P ðxþ b: nder the assumpton of suffent hannel bandwdth, dfferent query requests do not ontend for ommunaton 1. It s out the sope of ths researh to model user mobltes. The readers are referred to the lteratures [17], [18], [19] that have reported extensve studes on moblty modelng. (1) Fg. 1. The state dagram shows all possble transmssons of a query n a network wth four nodes. Wthout loss of generalty, we let Node 1 be the query ssuer. Sne we fous on a sngle query, the ategory of the query (.e., ) s not shown n the state dagram. resoure and thus an be treated ndependently. To ths end, we next smply analyze the delvery of a sngle query request to derve CðXÞ and PðXÞ, and to optmze X Analyt and Optmzaton Model Our analyss s based on a state dagram. Eah state s a vetor wth N elements,.e., S ¼½s 1 ;s 2 ;...;s N Š, where s ¼ 1 sgnfes Node has not reeved the query request, s ¼ 0 ndates Node s arryng the query request but has never transmtted t to another node, and s ¼ 1 means Node has reeved the query request and forwarded t one. Fg. 1 llustrates an example of the state dagram for a network wth four nodes. Sne we fous on a sngle query n ths analyss, the ategory of the query (.e., ) s not shown n the dagram. Wthout loss of generalty, we let Node 1 be the query ssuer. Thus the ntal state s S ¼½0; 1; 1;...; 1Š (see Fg. 1). The state transts, as depted by an arrow n the dagram, when the query request s transmtted from one node to another, e.g., from Node to Node j. Suh a state transton s denoted by L j. Sne we have assumed that a node transmts and reeves the same request only one, L j s possble only f s ¼ 0 and s j ¼ 1,.e., Node s arryng the query request but has never transmtted t whle Node j has not reeved the query request. Note that ths s the neessary ondton to enable a transmsson, but Node does not have to transmt the query to Node j f suh a transmsson s deemed neffent. As a matter of fat, t s the goal of our analyss and optmzaton to determne the most effent transmssons. The state dagram (nludng the states and transtons) forms a tree struture. The ntal state s the root of the tree. As shown n Fg. 1, a state may appear multple tmes due to dfferent transton paths to reah t. We treat suh dental states separately (as f they are dfferent states). A state an be n three status,.e., atve, negatvely termnated,orpostvely termnated, as defned below. Defnton 1. An atve state s a state that allows further transtons. The query request wll be further transmtted under an atve state. An atve state must have a 0 element,.e., s ¼ 0, whh s also alled an atve element. Defnton 2. Node s an atve element of an atve state f and only f s ¼ 0. Fg. 1 depts atve states only. Eah atve state may reman atve or be termnated n two possble ways, resultng n a negatvely termnated state or a postvely termnated state.

4 1304 IEEE TRANSACTIONS ON PARALLEL AND DISTRIBTED SYSTEMS, VOL. 26, NO. 5, MAY 2015 Defnton 3. An atve state beomes a postvely termnated state f the query request s answered by the atve element. pon reevng a query request n Category, Node has a probablty of E to answer the query. If the query s answered wthn ts delay budget, t wll no longer be forwarded t to another node. Ths essentally termnates the query proess gven the assumpton that a node forwards the same request only one. Otherwse, wth a probablty of 1 E, Node may reman atve and arry the request that s ready to be transmtted to another node. Defnton 4. An atve state beomes a negatvely termnated state f the query request s dropped by the atve element. A query request s dropped f ts delay budget expres. As a result, the state nludes no atve element and the query proess s termnated wthout a suessful reply. Fg. 1 depts all possble transmssons of a query n a network. Our goal s not to exeute all suh transmssons whh lead to hgh ommunaton overhead, but nstead to perform seletve transmssons to mnmze the ommunaton ost. To ths end, we have ntrodued the transmsson matrx, X, where X j ¼ 1 sgnfes that Node wll send the request to Node j, when the former meets the latter wth s ¼ 0 and s j ¼ 1; or the transmsson wll not be performed otherwse. Wth X as the varable to be optmzed, we now analyze the probablty to reah eah state. Sne the state dagram forms a tree struture, there s a unque path from the root (.e., the ntal state) to a gven state S ¼½s 1 ;s 2 ;...;s N Š. One agan, dental states may appear n the tree and they are treated as separate states. Let L S denote the path from the root to S, whh onssts of a sequene of transmssons fl S 1 2 S ;L S 2 3 S ;...;L S K 1 K S g, where f1 S, S 2 ;..., S K g are the set of nodes nvolved n the transmssons n sequene. For example, the path from the root to State S 14 ½1; 0; 1; 1Š nludes suh lnks as L 13, L 34, and L 42. Eah lnk ntrodues a transmsson delay. The total delay to reah the state s P K 1 ¼1 T S þ1. S We are nterested n postvely termnated states. A state S an be postvely termnated f and only f the followng three ondtons are satsfed. Frst, the transmsson matrx s onfgured suh that X S þ1 S ¼ 1, 81 K 1, formng a vald path from the root to State S. Seond, the total delay to reah the state s no greater than the delay budget d,.e., P K 1 ¼1 T S d. Note that d an be very large under S þ1 pratal settngs. As a matter of fat, MASON users are often not anxous to reeve prompt query reples and thus may tolerate long delay. In ths ase, the delay dstrbutons n the above formula an be redued to smple probabltes as to be dsussed n the next seton. Thrd, the last node on the path (.e., Node K S ) an answer the query whle others along the path (.e., Nodes S, 81 K 1) annot. Therefore the probablty that State S s postvely termnated s P S ðxþ ¼ YK 1 X S þ1 S ¼1 E S K YK 1 ¼1 ( K 1 X Pr T S þ1 S d ¼1 1 E S ) : (2) Sne the states are unorrelated, the total probablty to reah a postvely termnated state,.e., the probablty to delver the query request to a node that an answer the query s PðXÞ ¼ P S P SðXÞ. Wth proper manpulaton, we arrve at the followng formula: P ðxþ ¼ XN 1 X N X N X N Y 1 X kj k jþ1 k 6¼k 0 ;k 2 ;...;k 1 j¼0 ¼1 k 1 6¼k 0 k 2 6¼k 0 ;k 1 ( ) Y 1 X 1 Ek 1 E kj k jþ1 Pr T kj k jþ1 d : j¼0 Note that, the total number of add operatons n the above equaton equals the number of states, whh we have found to be N S ¼ P N 1 ¼1 A N 1, where A N 1 s the permutaton of out of N 1 elements,.e., A N 1 ¼!=ðN 1 Þ! The ommunaton ost n a wreless network s often proportonal to the number of transmssons. The more the transmssons, the hgher the onsumpton of energy and storage spae. It s out the sope of ths work to defne the best ost funton. We smply let CðXÞ be the total number of transmssons nvolved n the delvery of a query request. Let DðSÞ denote the depth of State S n the tree. Obvously, DðSÞ represents the number of transmssons (.e., the ost) needed to transt from the ntal state to S. Note that, the ost of DðSÞ s nurred as long as State S s reahed, no matter whether the query an be answered or not. Thus CðXÞ s gven below: CðXÞ ¼ X S j¼0 (3) DðSÞ P SðXÞ E : (4) K S P ðxþ and CðXÞ are obtaned va Eqs. (3) and (4), and then plugged nto Eq. (1). A branh and bound algorthm [38] s adopted here to dsover the optmal transmsson matrx X, n order to mnmze CðXÞ whle ensurng PðXÞ b Numer Results and Insghts Fg. 2 shows the numer results of the optmzaton model based on Haggle trae [39] (to be detaled n Seton 1). b s set to be 0:5. nderverysmalld, many queres annot be delvered wthn the delay budget, resultng n low delvery rate. When d < 30, the delvery rate s n fat lower than the expeted threshold (.e., b). By followng the optmzaton model, the nodes do not make unneessary attempts to transmt them, and aordngly the overhead s low. Wth a longer delay budget, more queres an reah the orrespondng data provders and thus the delvery rate naturally nreases. At the same tme, delay and overhead nrease too beause more transmssons wth longer delay are aggressvely attempted. However, when d s suffently large, many optons of routng paths beome avalable (that all satsfy the delay budget), allowng the optmzaton model to hoose the one wth the lowest overhead (.e., the one that nvolves the least transmssons). Ths explans why overhead dereases under large d. Suh a hoe often sarfes delay, as long as t does not exeed the allowed budget. Therefore, the average delay monotonally nreases wth d.

5 LI ET AL.: EFFICIENT DATA QERY IN INTERMITTENTLY-CONNECTED MOBILE AD HOC SOCIAL NETWORKS 1305 Fg. 2. Numer results of optmzaton. Whle the above optmzaton model an be mplemented by eah ndvdual node, t s ntrnsally entralzed (requrng global nformaton), and thus unpratal for real world mplementaton. However, t offers useful nsghts for the development of a dstrbuted data query protool. In partular, the essene of the optmzaton s to frst learn the probabltes to delver the query along dfferent paths to dfferent nodes, and then dede the optmal paths by strkng the balane between ost and delvery probablty. Ths nsght stmulates us to develop a dstrbuted sheme to learn reahable expertse, whh an effetvely gude the transmsson of a query to the node(s) wth suffent expertse to answer t. 2.2 Protool Desgn In ths seton, we ntrodue a dstrbuted protool for the data query n MASONs. It s based on two key tehnques. Frst, as motvated by our analyt and optmzaton model dsussed above, t employs reahable expertse as the routng metr to gude the transmsson of query requests. Seond, t explots the redundany n query transmsson. Redundany s not onsdered n the analyss due to ts ntratablty, but an effetvely mprove the query delvery rate n prate f t s properly ontrolled Routng Metr The delvery of query depends on a routng metr, whh s updated routnely and mantaned separately from the routng algorthm tself. We frst ntrodue suh a metr,.e., reahable expertse, that gudes query transmsson. As ntrodued n Seton 2.1.2, eah node has ertan expertse to answer a query. We let E denote the expertse of Node to answer a query n Category. We have assumed E s known n our analyss. In prate, however, t s nontrval to properly defne the expertse, beause a moble node hardly knows presely ts probablty to answer queres n eah ategory. It may ntally lam ts expertse based on the moble user s soal roles (e.g., professons), nterests, and avalable resoures. But suh ntally lamed expertse s often naurate. Therefore, after ntalzaton, the expertse should be updated aordng to the feedbaks from other nodes, espeally the query ssuers. In ths researh, we adopt the exponentally weghted movng average (EWMA) to mantan and update expertse. More spefally, we have E ð1 mþ E þ mf ; (5) where 0 m 1 s a onstant weght to keep partal memory of hstor status, ½E Š s the expertse before t s updated, and F s the feedbak sore for queres that Node has answered n Category. Varous feedbak ratng shemes an be adopted to determne F. In ths researh, we employ a smple sheme, whh supports quk onvergene as to be dsussed n Seton 3 and shown n Fg. 3h. The expertse ndates the apablty of a node to answer queres, but tself s nsuffent to gude query transmsson. For example, a node may have hgh expertse, but s not reahable by the query ssuer and thus beomes less helpful to answer the query. To ths end, we defne k-hop reahable expertse. As dsussed earler, the MASON users an often tolerate long delay. Thus, the delay random varables n Eq. (3),.e., T S þ1, S may be redued to smple nodal ontat probabltes. Let p j denote the probablty that Nodes and j meet. The mantenane and update of p j have been dsussed extensvely n DTN networks [20], [21], [24], [25]. The k-hop reahable expertse s alulated as follows: E ðkþ ¼1 Y j2f 1 pj E j ðk 1Þ ; (6) where F s the set of nodes that Node meets frequently. E ðkþ ndates the probablty that Node an delver the query wthn k hops to a node that an answer the query. Clearly, E ð0þ ¼E and E ð1þ ¼1 Q j2f ð1 p j Ej Þ. Node ollets fej ðk 1Þj8 j 2 F and 0 <kng whenever t meets other nodes, and perodally makes an update on E ðkþ aordng to Eq. (6). Based on the k-hop reahable expertse, we defne the aggregated reahable expertse, AE ¼ 1 YN k¼1 1 E ðkþ ; (7) whh ndates the overall ablty of Node to help a query n Category to be answered. It serves as the routng metr to gude query delvery as to be dsussed next Routng wth Dynam Redundany Control Based on the routng metr,.e., reahable expertse, we now ntrodue the routng algorthm. The delvery of a query s guded by the aggregated reahable expertse, where the query s generally forwarded from the node wth a lower aggregated reahable expertse to the node wth a hgher

6 1306 IEEE TRANSACTIONS ON PARALLEL AND DISTRIBTED SYSTEMS, VOL. 26, NO. 5, MAY 2015 Fg. 3. Expermental results.

7 LI ET AL.: EFFICIENT DATA QERY IN INTERMITTENTLY-CONNECTED MOBILE AD HOC SOCIAL NETWORKS 1307 one. In ontrast to the onventonal store-and-forward data transmsson where a sngle opy of data s transmtted aross the network, redundany s often employed n opportunst networks. Whle redundany s not addressed n the analyss due to ts ntratablty, t s mportant n prate to aheve the desred query delvery rate. Generally speakng, the hgher the redundany, the hgher probablty the query s answered suessfully. However, redundany must be properly ontrolled as exessve redundany may exhaust network apaty and thus degrade the performane. A nave approah s to reate a fxed amount of redundany for eah query. For example, a predetermned number of opes of the query an be reated and dstrbuted to other nodes n the network. Ths approah, however, s often neffent, beause the effetveness of redundany depends on the nodes that reeve, arry and forward the query. In an extreme ase, all redundant opes of the query may be transmtted and arred by the nodes that have lttle hane to meet the node(s) that an answer the query and thus beome neffetve. As a matter of fat, the effetveness of redundany hghly depends on the reahable expertse of the nodes that arry the redundant opes. To ths end, we ntrodue a parameter to dynamally reflet the effetve redundany. More spefally, the proposed routng algorthm wth dynam redundany ontrol s outlned below. Let R q denote the redundany of Query q as observed by Node. The parameter s the estmated probablty that at least one opy of the query s answered by any other nodes n the network. It s mantaned and updated n a dstrbuted way. Assume Query q s n Category. R q s ntalzed to zero when the query s reated and subsequently updated durng ts transmsson. Sne ommunaton opportunty s low, transmsson s often between two nodes only. If more than two nodes are wthn ommunaton range, we assume an underlyng medum aess ontrol protool that randomly selets one node as the sender. Therefore we onsder a general senaro where Node meets Node j n the followng dsussons. Frst, Nodes and j exhange ther k-hop reahable expertse and update ther aggregated reahable expertse aordng to Eqs. (6) and (7). Then, Node fethes the query wth the lowest redundany n ts queue. The queue holds the queres that Node reates or reeves from other nodes. It s sorted aordng to the redundany level suh that the query wth the lowest redundany (denoted as Query q n Category ) s at the head of the queue. If Node j has a hgh expertse for queres n Category (.e., Ej a where a s a predefned onstant), t dretly answers the query by reatng and sendng a query reply to the query ssuer. Sne the destnaton of the query reply (.e., the query ssuer) s known, t an be delvered va any exstng routng protool for opportunst networks [11], [12], [13], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [40], [41], [42], [43]. We adopt the sheme proposed n [24] n our mplementaton. Note that the answer of Node j s not always satsfatory. It has a probablty of Ej to be satsfed by the query ssuer. Therefore, Node removes the query from ts queue wth a probablty of Ej. Otherwse, f Node j annot answer the query dretly (.e., Ej < a), Node heks the redundany of Query q. If R q b where b s the desred query delvery probablty, t mples that hgh enough redundany has been reated for thequery.thusnoatonwllbetaken.nodesmply holds the query untl t meets a node that an dretly answer the query or the query must be dropped due to queue overflow. An overflow happens when a new query s added nto the queue whh s already full. In ths ase, thequerywththehghestredundany(.e.,attheendof the queue) s dropped. If R q < b, the query should be further propagated. But t s transmtted to Node j only f AE <AEj (.e., the latter has a better hane to delver the query). Ths transmsson reates two opes of the query, eah sharng partal responsblty to get the answer. The redundanes for the two opes are assgned as follows: R q j 1 1 R q 1 E ð0þ ; (8) R q 1 1 R q 1 E j ð0þ : (9) In both formulas, ð1 R q Þ denotes the estmated probablty that none of other nodes (exept Nodes and j) an get the answer for Query q, and ð1 E ð0þþ and ð1 E jð0þþ gve the probablty that Node and Node j annot dretly answer the query. Therefore the updated R q (or R q j ) ndates the probablty that at least one opy of the query an be answered by other nodes exept Node (or Node j). The transmsson of Query q ontnues upon future ommunaton opportuntes untl, as dsussed earler, the query s answered by a node or dropped due to queue overflow. pon reevng the query reply, the query ssuer evaluates t and onstruts a feedbak paket, whh s delvered to the node that answers the query, agan, va an exstng routng protool for opportunst networks. The latter then updates ts expertse aordng to Eq. (5). 3 PROTOTYPE AND EXPERIMENT To demonstrate the feasblty and effeny of the proposed data query protool and to gan useful empral nsghts, we have arred out a testbed experment usng off-the-shelf Dell Streak tablets. In ths seton, we frst ntrodue our testbed setup and then present expermental results. 3.1 Prototype and Testbed Setup We have developed a prototype system by usng Dell Streak 5 and 7 tablets that are of the smartphone/tablet PC hybrd operatng on Androd 2:2. The ommunaton between the tablets s enabled va Bluetooth. A Streak tablet has 16 GB nternal storage adequate to keep large amounts of expermental data. We have mplemented our proposed data query protool by usng standard Androd APIs, losely followng the desrpton n the prevous seton. In order to save power, eah node ntates neghbor dsovery one every a random nterval (between 5 to 10 mnutes). Our experment nvolves twenty fve volunteers nludng faulty members and students. They are marked as Node 0 to 24. In the experment, we defne three ategores

8 1308 IEEE TRANSACTIONS ON PARALLEL AND DISTRIBTED SYSTEMS, VOL. 26, NO. 5, MAY 2015 TABLE 1 Results under Experment of queres,.e., hstory, sene and arts (whh are named Category 0, 1 and 2, respetvely). Eah partpant has a lamed ntal expertse for answerng queres n eah ategory and generates twelve queres per day n randomly hosen ategores. Note that the ntal expertse s not aurate. The true expertse s arbtrarly set by lettng a small set of nodes to have an expertse of 1 to answer queres n eah ategory. More spefally, Nodes 3, 13, and 19 an answer queres n Category 0; Nodes 4, 14, and 20 an answer queres n Category 2; and Node 1 an answer queres n any ategory. Other nodes ntalze expertse to be ð0 1Þ randomly and learn and update ther aggregated expertse durng the experment. The experment had run for ffteen days, startng from Monday 16:00 p.m. n the frst week to Monday 17:00 p.m n the thrd week. We ompare several varatons of the proposed sheme and related shemes. In the followng dsussons, zerohop, one-hop, two-hop, and three-hop stand for the proposed sheme that allows up to zero hop, one hop, two hops, and three hops relayng, respetvely; No Feedbak means the proposed sheme (wth two-hop relay) but wthout the feedbak mehansm to retfy expertse; Floodng s the smple floodng sheme for query delvery; Gossp [44] onsders multple ategores and assgns the queres n eah ategory a transmsson probablty for data transmsson; Wllngness [45] s a sheme that a query s delvered based on wllngness, whh s the degree to whh a node atvely engages n tryng to re-transmt a query; Spray and Wat [33] s onsdered as a baselne opportunst delvery protool; Soal-based [16] s a soal-based routng sheme. We are prmarly nterested n two parameters: (1) the suess query rate,.e., the rato of suessfully answered queres to the total generated queres, and (2) the query delay whh s the perod from the tme when a node generates the query to the tme when t reeves the answer. 3.2 Expermental Results Table 1 shows the overall performane of dfferent shemes. The two-hop sheme aheves the hghest query rate. It s not surprsng to fnd the one-hop sheme wth a lower query rate sne a node merely tres to answer the queres va up to one hop relay. The zero-hop sheme has the lowest query rate as a query an be answered only when the query ssuer meets the data provder dretly. On the other hand, t seems ant-ntutve that allowng a longer relay path (e.g., the three-hop sheme) leads to a negatve gan. But ths s reasonable beause exessve redundany s reated when too many nodes are nvolved n relayng queres, subsequently overloadng the network and resultng n degraded performane. For a smlar reason, the Floodng sheme has a even lower query rate gven ts extremely hgh redundany. nder the No Feedbak sheme, the naurate expertse s not retfed, resultng n msleadng reahable expertse and thus lower query rate. Gossp [44] onsders multple ategores and assgns the queres n eah ategory a transmsson probablty for data transmsson. However, asagosspngapproah,ts data transmsson s randomzed. Therefore a query s often answered and arred by nodes wth nsuffent expertse, thus ndung many non-satsfatory reples. Wllngness [45] s a sheme that a query s delvered based on wllngness, whh s the degree to whh a node atvely engages n tryng to re-transmt a query. The wllngness does not reflet the expertse based on whh a node reples queres, therefore the nodes are not helpful for eah other to arry queres to nodes wth suffent expertse. We also ompare wth Spray and Wat [33] whh s onsdered as a baselne opportunst delvery protool. [33] fxes the number of opes for eah query whh lmts the queres to go through orret paths to be repled by nodes wth suffent expertse, makng query rate even lower. Soal-based [16] explots a dstrbuted ommunty parttonng algorthm to dvde a DTN nto smaller ommuntes. For ntra-ommunty ommunaton, a utlty funton onvolutng soal smlarty and soal entralty wth a deay fator s used to hoose relay nodes. For nter-ommunty ommunaton, the nodes movng frequently aross ommuntes are hosen as relays to arry data to destnaton effently. Although Zhu et al. [16] ntrodues a soluton for DTNs whh leverages soal propertes and moblty haratersts of users, t s not truly applable for the data query n MASONs, beause when a node ssues a query, t s often unaware of the nodes that have suffent expertse to answer the query. The ost s prohbtvely hgh to onstrut a struture to ndex data and data provders lke P2P networks. It s obvously neffent ether to frequently flood queres, whh are expensve and often onsdered spams. It s not surprsng that our proposed sheme has better performane than Soal-based, sne Soal-based does not apture the nherent features for the query delvery n MASONs, hene the nodes are not helpful for eah

9 LI ET AL.: EFFICIENT DATA QERY IN INTERMITTENTLY-CONNECTED MOBILE AD HOC SOCIAL NETWORKS 1309 other by makng the orret desons to arry queres to satsfatory nodes. In general, when more hops are allowed n relayng queres, the overhead nreases, beause a query s more aggressvely propagated. As a result, more opes of the query are transmtted n the network and the query ssuer often reeves more reples. At the same tme, sne a query may potentally go through a longer path to reah the data provder, the average delay also nreases. Compared wth the two-hop sheme, No Feedbak has longer delay and more number of reples beause norret expertse often leads the queres to wrong routes. More than 96 perent queres are answered suessfully. The unanswered queres are all generated durng the fnal hours of the experment. Fg. 3a llustrates the number of queres answered on eah day of the experment. As an be seen, the results vary among days, refletng the movng patterns of the partpants. More queres are answered durng weekdays than weekends due to the lower nteratve atvtes of students and faulty on Saturday and Sunday. In fat, many queres annot be answered durng weekends and have to wat untl Monday of the next week. Ths explans the peak on the seond Monday. It s worth mentonng that the frst and the thrd Monday are not the whole days, hene the number of answered queres s less than the seond Monday. The atvty pattern s also evdened by the delay varaton shown n Fg. 3b. Queres generated n weekends have longer delay ompared wth those n weekdays. The delay of queres generated on Frday s also hgh beause no lasses are sheduled on Frday afternoon and many offes are losed after 1:00 p.m.. Fg. 3b also shows the total traff n the network, whh follows a smlar pattern of nodal atvtes. Fgs. 3 and 3d further zoom n to show the results n eah hour of a day. The data are averaged over 15 days. Both the network traff and the number of answered queres are hgh durng daytme and low at nght, whh agan shows the query heavly depends on the atvty of students and faulty who arry the tablets. Lkewse, we expet lower delay durng daytme. However, the results are just the opposte (as depted n Fg. 3d). Suh ant-ntutve observaton s due to the queres from a few nodes, whh experened extremely long delay that domnates the overall performane. In fat, most queres generated durng daytme ndeed have short delay. But a set of nodes (nludng Nodes 3-12) rely on asnglenode(node2) to arry ther queres n Category 2 to orrespondng data provders. Suh delvery happens around 9:00 a.m. daly. The queres generated after 9:00 a.m. must wat untl the next day, thus ndung unusually long lateny that sgnfantly elevates the overall average. If we exlude suh queres (see the lower purple bars n Fg. 3d), the average delay beomes muh lower, and the daytme delay s generally shorter than that durng nght. The average delay and traff of dfferent nodes are llustrated n Fg. 3e. In general, delay and traff vary among dfferent nodes due to the randomness n nodal moblty and query generaton and transmsson. Node 0 has extremely poor onnetvty (ether dretly or ndretly) to the nodes wth hgh expertse, resultng n very long delay ompared wth other nodes. Contrarly, sne Node 1 s able to answer all the queres of three ategores, t has the mnmum delay. In addton, Node 2 has the heavest traff load beause t frequently meets other nodes, whle Node 0 arres the least traff due to few nteratons between t and other nodes. The delay dstrbuton s shown n Fg. 3f. More than 65 perent queres are answered wthn two hours. The queres wth longer delays are ether generated by Nodes 3-12 as dsussed above, or generated durng weekends and thus annot be repled untl the next Monday. Fg. 3g llustrates the dstrbuton of path length. All queres are answered wthn three hops. Fg. 3h shows the onvergene of the lamed expertse to the ground truth. A node s hosen as an example, whle smlar results are observed n other nodes as well. As we an see, the feedbak mehansm effetvely adjusts the node s expertse, gradually approahng to the true value wthn a few hours. 4 SIMLATION RESLTS Besdes the experment dsussed above, extensve smulatons are arred out to learn the performane trend of the proposed data query algorthm under varous network settngs, whh are not pratal to evaluate by usng lab equpments. The smulaton odes are extrated from our prototype mplementaton, and the smulaton results are obtaned under real-world traes and powerlaw moblty model. Eah node mantans a maxmum queueszeof1; Smulaton under Haggle Trae We have evaluated our proposed sheme under several real-world traes avalable at CRAWDAD [46]. Table 2 shows the results based on Haggle trae [39], whh nludes 98 partpants arryng small deves (Motes) durng Infoom We run the smulaton for a perod of 342; 916 seonds (or about 4 days). Eah node generates 1:08 queres per hour. The queres fall nto fve ategores, and eah ategory s assoated wth three expert nodes that an provde satsfatory reples. Smlar to the results n Table 1, Table 2 shows the results under Haggle trae. The dstrbutons of query rate and delay are llustrated n Fg. 4. About 90 perent of nodes an aheve a query rate of 80 perent or hgher under the proposed sheme. At the same tme, more than half of the queres are answered wthn an hour. 4.2 Smulaton under Power-Law Moblty Model Besdes the above results based on Haggle trae, we have arred out smulatons under power-law moblty model, whh enables onvenent study of performane trend wth the varaton of several network parameters. More spefally, we smulate an area that s parttoned nto a grd of ells. Eah node s assoated wth a randomly-hosen home ell, n whh t ntally resdes. In a tme slot, t may move n one of the four dretons,.e., up, down, left and rght, or stay n ts home ell. Let P ðxþ denote the probablty for Node to be at Cell x. P ðxþ ¼k ð 1 d ðxþ Þs where k s

10 1310 IEEE TRANSACTIONS ON PARALLEL AND DISTRIBTED SYSTEMS, VOL. 26, NO. 5, MAY 2015 TABLE 2 Results under Haggle Trae a onstant, s s the exponent of the power-law dstrbuton and d ðxþ denotes the dstane between Cell x and Node s home ell. The default network parameters nlude a network of 100 nodes, a s of 2, 10 ategores of queres, fve experts per ategory, and a generaton rate of 0:02 queres per tme unt per node. In an opportunst network, the ommunaton apaty hghly depends on the meetng opportuntes among moble nodes. As shown n Fg. 5a, query reply rate grows wth the nrease of the network densty, beause the nodes have more opportuntes to meet eah other and exhange ther queres. Fg. 5b depts the mpat of the power-law fator s. When s s large, the nodes tend to stay n ther home ells,.e., have low moblty, resultng n small probabltes to meet eah other and onsequently small network apaty. Therefore the query reply rate s low. When s s extremely large, the query reply rate may approah as low as zero. On the other hand, when s s small, the nodes have unform moblty,.e., smlar probabltes to vst all ells and aordngly smlar routng metr (.e., k-hop reahable expertse), renderng routng neffetve. nder the smulated network settng, s ¼ 2 results n the best performane. The mpat of traff load s llustrated n Fg. 5. Whle the query reply rate keeps stable at the begnnng under all shemes, t starts to drop when the generaton rate exeeds 0:03. In general, wth a hgher query generaton rate, the overall traff load nreases, resultng n more frequent queue overflow and onsequently lower query reply rate. As dsussed n Seton 2, a threshold b s employed for dynam redundany ontrol. A larger b allows more redundany to be reated, amng to aheve a hgher query reply rate. However, f b s too large, the exessve redundany degrades the utlzaton of ommunaton and storage resoures and lowers the overall performane aordngly (see the Fg. 5d). Fg. 5e shows that a hgher query reply rate s aheved wth the nrease of queue sze, beause more queres and reples an be kept n the queue untl they are delvered. The number of experts for eah ategory s also studed n ths work. As shown n Fg. 5f, more experts for a ategory result n hgher query reply rate beause more nodes an answer the queres n ths ategory. 5 CONCLSION We have studed the problem of data query n a Moble Adho SOal Network, amng to determne an optmal transmsson strategy that supports the desred query rate wthn Fg. 4. Dstrbuton under Haggle trae. Fg. 5. Performane trend under power-law moblty model.

11 LI ET AL.: EFFICIENT DATA QERY IN INTERMITTENTLY-CONNECTED MOBILE AD HOC SOCIAL NETWORKS 1311 a delay budget and at the same tme mnmzes the total ommunaton ost. We have proposed a entralzed optmzaton model that offers useful theoret nsghts and developed a dstrbuted data query protool for pratal applatons. To demonstrate the feasblty and effeny of the proposed sheme and to gan useful empral nsghts, we have arred out a testbed experment by usng 25 off-the-shelf Dell Streak tablets for a perod of 15 days. Moreover, we have run extensve smulatons to learn the performane trend under varous network settngs, whh are not pratal to buld and evaluate n laboratores. REFERENCES [1] B. Yang and A. Hurson, A ontent-aware multmeda aessng model n ad ho networks, n Pro. 11th Int. Conf. Parallel Dstrb. Syst., 2005, pp [2] C. Avn and C. Brto, Effent and robust query proessng n dynam envronments usng random walk tehnques, n Pro. 3rd Int. Symp. Inf. Proess. Sens. Netw., 2004, pp [3] N. Chang and M. Lu, Controlled floodng searh n a large network, IEEE/ACM Trans. Netw., vol. 15, no. 2, pp , Apr [4] T. Hara, Effetve repla alloaton n ad ho networks for mprovng data aessblty, n Pro. IEEE 20th Annu. Jont Conf. Comput. Commun., 2001, pp [5] W. W. Terpstra, J. Kangasharju, C. Leng, and A. P. Buhmann, BubbleStorm: Reslent, probablst, and exhaustve peer-topeer searh, n Pro. ACM Conf. Appl., Tehnol., Arht., Protools Comput. Commun., 2007, pp [6] A. N. Shferaw, V.-M. Sutur, and L. Brune, Interest-awareness for nformaton sharng n MANETs, n Pro. 11th Int. Conf. Moble Data Manage., 2010, pp [7] Z. L, H. Shen, G. Lu, and J. L, SOS: A dstrbuted moble QA system based on soal networks, n Pro. IEEE 32nd Int. Conf. Dstrb. Comput. Syst., 2012, pp [8] P. Hu, J. Leguay, J. Crowroft, J. Sott, T. Fredman, and V. Conan, Osmoss n poket swthed networks, n Pro. 1st Int. Conf. Commun. Netw. Chna, 2006, pp [9] J. Fan, J. Chen, Y. Du, P. Wang, and Y. Sun, DelQue: A soally aware delegaton query sheme n delay-tolerant networks, IEEE Trans. Vehular Tehnol., vol. 60, no. 5, pp , Jun [10] Z. Yang, T. Nng, and H. Wu, Dstrbuted data query n ntermttently onneted passve RFID networks, IEEE Trans. Parallel Dstrb. Syst., vol. 24, no. 10, pp , Ot [11] A. Me, G. Morabto, P. Sant, and J. Stefa, Soal-aware stateless forwardng n poket swthed networks, n Pro. IEEE Conf. Comput. Commun., 2011, pp [12] W. Gao and G. Cao, ser-entr data dssemnaton n dsrupton tolerant networks, n Pro. IEEE Conf. Comput. Commun., 2011, pp [13] Z. L and H. Shen, SEDM: Explotng soal networks n utltybased dstrbuted routng for DTNs, IEEE Trans. Comput., vol. 62, no. 1, pp , Jan [14] E. M. Daly and M. Haahr. (2007). Soal network analyss for routng n dsonneted delay-tolerant MANETs Pro. 8th ACM Int. Symp. Moble Ad Ho Netw. Comput., pp [Onlne]. Avalable: [15] P. Hu, J. Crowroft, and E. Yonek. (2009)., Bubble Rap: Soalbased forwardng n delay tolerant networks n Pro. 9th ACM Int. Symp. Moble Ad Ho Netw. Comput., pp [Onlne]. Avalable: [16] K. Zhu, W. L, and X. Fu, SMART: A soal and moble aware routng strategy for dsrupton tolerant networks, IEEE Trans. Vehular Teh., vol. PP, no. 99, pp. 1 1, [17] A. Chantreau, P. Hu, J. Crowroft, C. Dot, R. Gass, and J. Sott, Impat of human moblty on the desgn of opportunst forwardngalgorthms, npro.ieeeconf.comput.commun.,2006,pp [18] M. Km, D. Kotz, and S. Km, Extratng a moblty model from real user traes, n Pro. IEEE Conf. Comput. Commun., 2006, pp [19] T. Spyropoulos, K. Psouns, and C. Raghavendra, Performane analyss of moblty-asssted routng, n Pro. 7th ACM Int. Symp. Moble Ad Ho Netw. Comput., 2006, pp [20] K. Fall, A delay-tolerant network arhteture for hallenged nternets, n Pro. ACM Conf. Appl., Tehnol., Arht., Protools Comput. Commun., 2003, pp [21] A. Balasubramanan, B. N. Levne, and A. Venkataraman, DTN routng as a resoure alloaton problem, n Pro. ACM Conf. Appl., Tehnol., Arht., Protools Comput. Commun., 2007, pp [22] D. Gunawardena, T. Karaganns, A. Proutere, E. Santos-Neto, and M. Vojnov, Soop: Deentralzed and opportunst multastng of nformaton streams, n Pro. 11th ACM Int. Conf. Moble Comput. Netw., 2011, pp [23] T. Spyropoulos, K. Psouns, and C. S. Raghavendra, Effent routng n ntermttently onneted moble networks: The sngleopy ase, IEEE/ACM Trans. Netw., vol. 16, no. 1, pp , Feb [24] P. Juang, H. Ok, Y. Wang, M. Martonos, L.-S. Peh, and D. Rubensten, Energy-effent omputng for wldlfe trakng: Desgn tradeoffs and early experenes wth Zebranet, n Pro. Annu. Conf. Arht. Support Program Languages Operatng Syst., 2002, pp [25] T. Small and Z. J. Haas, The shared wreless nfostaton model A new ad ho networkng paradgm, n Pro. 4th ACM Int. Symp. Moble Ad Ho Netw. Comput., 2003, pp [26] B. Burns, O. B. Bran, and N. Levne, MV routng and apaty buldng n dsrupton tolerant networks, n Pro. IEEE Conf. Comput. Commun., 2005, pp [27] N. Banerjee, M. D. Corner, D. Towsley, and B. N. Levne, Relays, base statons, and meshes: Enhanng moble networks wth nfrastruture, n Pro. 14th ACM Int. Conf. Moble Comput. Netw, 2008, pp [28] P. Hu, A. Chantreau, J. Sott, R. Gass, J. Crowroft, and C. Dot, Poket swthed networks and human moblty n onferene envronments, n Pro. ACM SIGCOMM Workshop Delay-Tolerant Netw., 2005, pp [29] S. B. Esenman, E. Mluzzo, N. D. Lane, R. A. Peterson, G.-S. Ahn, and A. T. Campbell, The BkeNet moble sensng system for ylst experene mappng, n Pro. 5th Int. Conf. Embedded Netw. Sens. Syst., 2007, pp [30] P. Klasnja, B. L. Harrson, L. LeGrand, A. LaMara, J. Froehlh, and S. E. Hudson, sng wearable sensors and real tme nferene to understand human reall of routne atvtes, n Pro. 10th Int. Conf. bqutous Comput., 2008, pp [31] E. Mluzzo, N. D. Lane, K. Fodor, R. A. Peterson, H. Lu, M. Musoles, S. B. Esenman, X. Zheng, and A. T. Campbell, Sensng meets moble soal networks: The desgn, mplementaton and evaluaton of the CeneMe applaton, n Pro. 6th ACM Conf. Embedded Netw. Sens. Syst., 2008, pp [32] P. Mohan, V. Padmanabhan, and R. Ramjee, Nerell: Rh montorng of road and traff ondtons usng moble smartphones, n Pro. 6th ACM Conf. Embedded Netw. Sens. Syst., 2008, pp [33] T. Spyropoulos, K. Psouns, and C. S. Raghavendra, Spray and wat: An effent routng sheme for ntermttently onneted moble networks, n Pro. ACM SIGCOMM Workshop Delay-Tolerant Netw., 2005, pp [34] A. Lndgren, A. Dora, and O. Shelen, Probablst routng n ntermttently onneted networks, ACM SIGMOBILE Moble Comput. Commun. Rev., vol. 7, no. 3, pp , [35] S. Jan, M. Demmer, R. Patra, and K. Fall, sng redundany to ope wth falures n a delay tolerant network, n Pro. ACM Conf. Appl., Tehnol., Arht., Protools Comput. Commun., 2005, pp [36] A. A. Hanbal, P. Nan, and E. Altman, Performane of ad ho networks wth two hop relay routng and lmted paket lfetme, Performane Eval., vol. 65, no. 6, pp , [37] X. Te, A. Venkataraman, and A. Balasubramanan, R3: Robust replaton routng n wreless networks wth dverse onnetvty haratersts, n Pro. 17th Annu. Int. Conf. Moble Comput. Netw., 2011, pp [38] A. H. Land and A. G. Dog, An automat method of solvng dsrete programmng problems, Eonometra, vol. 28, no. 3, pp , [39] J. Sott, R. Gass, J. Crowroft, P. Hu, C. Dot, and A. Chantreau. (2009). CRAWDAD trae ambrdge/haggle/mote/nfoom2006 (v ), [Onlne]. Avalable: dartmouth.edu/

12 1312 IEEE TRANSACTIONS ON PARALLEL AND DISTRIBTED SYSTEMS, VOL. 26, NO. 5, MAY 2015 [40] Y. Wang and H. Wu, DFT-MSN: The delay fault tolerant moble sensor network for pervasve nformaton gatherng, n Pro. IEEE Conf. Comput. Commun., 2006, pp [41] Y. Wang and H. Wu, Delay/fault-tolerant moble sensor network (DFT-MSN): A new paradgm for pervasve nformaton gatherng, IEEE Trans. Moble Comput., vol. 6, no. 9, pp , Sep [42] Y. Wang, H. Wu, F. Ln, and N.-F. Tzeng, Cross-layer protool desgn and optmzaton for delay/fault-tolerant moble sensor networks, IEEE J. Sel. Areas Commun., vol. 26, no. 5, pp , Jun [43] H. Dang and H. Wu, Clusterng and luster-based routng protool for delay-tolerant moble networks, IEEE Trans. Wreless Commun., vol. 9, no. 6, pp , Jun [44] Z. J. Haas, J. Y. Halpern, and L. L, Gossp-based ad ho routng, IEEE/ACM Trans. Netw., vol. 14, no. 3, pp , Jun [45] K. A. Harras and K. C. Almeroth, Controlled floodng n dsonneted sparse moble networks, Wreless Commun. Moble Comput., vol. 9, no. 1, pp , [46] [Onlne]. Avalable: Yang Lu (S 11) reeved the BE degree n eletral engneerng and ts automaton, and the ME degree n ontrol theory and ontrol engneerng from Harbn Engneerng nversty, Chna, n 2008 and 2010, respetvely. He has been workng toward the PhD degree n omputer engneerng at the Center for Advaned Computer Studes, nversty of Lousana at Lafayette, sne Hs urrent researh nterests nlude delay-tolerant networks, rado frequeny dentfaton (RFID) systems, and wreless sensor networks. He s a student member of the IEEE. Yanyan Han reeved the BS degree n eletron nformaton engneerng from Shandong nversty, Jnan, Chna, n 2008, and has been workng toward the PhD degree n omputer sene at the Center for Advaned Computer Studes (CACS), nversty of Lousana at Lafayette (L Lafayette), sne Her urrent researh nterests nlude moble ad ho networks and delay-tolerant networks. Zhpeng Yang (S 07) reeved the BS and MS degrees n omputer sene from Tanjn nversty, Chna, n 2001 and 2004, respetvely. He has been workng toward the PhD degree n omputer sene at the Center for Advaned Computer Studes (CACS), nversty of Lousana at Lafayette (L Lafayette), sne From 2004 to 2006, he was a software engneer n Nortel and Luent Chna. Hs urrent researh nterests nlude delay-tolerant networks, rado frequeny dentfaton (RFID) systems, and wreless sensor networks. He s a student member of the IEEE. Hongy Wu (M 02) reeved the BS degree n sentf nstruments from Zhejang nversty, Hangzhou, Chna, n 1996, and the MS degree n eletral engneerng and the PhD degree n omputer sene from the State nversty of New York (SNY) at Buffalo n 2000 and 2002, respetvely. Sne then, he has been wth the Center for Advaned Computer Studes (CACS), nversty of Lousana at Lafayette (L Lafayette), where he s urrently a professor and hold the Alfred and Helen Lamson Endowed professorshp n omputer sene. Hs researh nterests nlude delay-tolerant networks, rado frequeny dentfaton (RFID) systems, wreless sensor networks, and ntegrated heterogeneous wreless systems. He reeved the S Natonal Sene Foundaton (NSF) CAREER Award n 2004 and the L Lafayette Dstngushed Professor Award n He s a member of the IEEE. " For more nformaton on ths or any other omputng top, please vst our Dgtal Lbrary at

On the End-to-end Call Acceptance and the Possibility of Deterministic QoS Guarantees in Ad hoc Wireless Networks

On the End-to-end Call Acceptance and the Possibility of Deterministic QoS Guarantees in Ad hoc Wireless Networks On the End-to-end Call Aeptane and the Possblty of Determnst QoS Guarantees n Ad ho Wreless Networks S. Srram T. heemarjuna Reddy Dept. of Computer Sene Dept. of Computer Sene and Engneerng Unversty of

More information

Session 4.2. Switching planning. Switching/Routing planning

Session 4.2. Switching planning. Switching/Routing planning ITU Semnar Warsaw Poland 6-0 Otober 2003 Sesson 4.2 Swthng/Routng plannng Network Plannng Strategy for evolvng Network Arhtetures Sesson 4.2- Swthng plannng Loaton problem : Optmal plaement of exhanges

More information

Matrix-Matrix Multiplication Using Systolic Array Architecture in Bluespec

Matrix-Matrix Multiplication Using Systolic Array Architecture in Bluespec Matrx-Matrx Multplaton Usng Systol Array Arhteture n Bluespe Team SegFault Chatanya Peddawad (EEB096), Aman Goel (EEB087), heera B (EEB090) Ot. 25, 205 Theoretal Bakground. Matrx-Matrx Multplaton on Hardware

More information

Time Synchronization in WSN: A survey Vikram Singh, Satyendra Sharma, Dr. T. P. Sharma NIT Hamirpur, India

Time Synchronization in WSN: A survey Vikram Singh, Satyendra Sharma, Dr. T. P. Sharma NIT Hamirpur, India Internatonal Journal of Enhaned Researh n Sene Tehnology & Engneerng, ISSN: 2319-7463 Vol. 2 Issue 5, May-2013, pp: (61-67), Avalable onlne at: www.erpublatons.om Tme Synhronzaton n WSN: A survey Vkram

More information

Semi-analytic Evaluation of Quality of Service Parameters in Multihop Networks

Semi-analytic Evaluation of Quality of Service Parameters in Multihop Networks U J.T. (4): -4 (pr. 8) Sem-analyt Evaluaton of Qualty of Serve arameters n Multhop etworks Dobr tanassov Batovsk Faulty of Sene and Tehnology, ssumpton Unversty, Bangkok, Thaland bstrat

More information

Research on Neural Network Model Based on Subtraction Clustering and Its Applications

Research on Neural Network Model Based on Subtraction Clustering and Its Applications Avalable onlne at www.senedret.om Physs Proeda 5 (01 ) 164 1647 01 Internatonal Conferene on Sold State Deves and Materals Sene Researh on Neural Networ Model Based on Subtraton Clusterng and Its Applatons

More information

Minimize Congestion for Random-Walks in Networks via Local Adaptive Congestion Control

Minimize Congestion for Random-Walks in Networks via Local Adaptive Congestion Control Journal of Communatons Vol. 11, No. 6, June 2016 Mnmze Congeston for Random-Walks n Networks va Loal Adaptve Congeston Control Yang Lu, Y Shen, and Le Dng College of Informaton Sene and Tehnology, Nanjng

More information

Connectivity in Fuzzy Soft graph and its Complement

Connectivity in Fuzzy Soft graph and its Complement IOSR Journal of Mathemats (IOSR-JM) e-issn: 2278-5728, p-issn: 2319-765X. Volume 1 Issue 5 Ver. IV (Sep. - Ot.2016), PP 95-99 www.osrjournals.org Connetvty n Fuzzy Soft graph and ts Complement Shashkala

More information

Bit-level Arithmetic Optimization for Carry-Save Additions

Bit-level Arithmetic Optimization for Carry-Save Additions Bt-leel Arthmet Optmzaton for Carry-Sae s Ke-Yong Khoo, Zhan Yu and Alan N. Wllson, Jr. Integrated Cruts and Systems Laboratory Unersty of Calforna, Los Angeles, CA 995 khoo, zhanyu, wllson @sl.ula.edu

More information

Microprocessors and Microsystems

Microprocessors and Microsystems Mroproessors and Mrosystems 36 (2012) 96 109 Contents lsts avalable at SeneDret Mroproessors and Mrosystems journal homepage: www.elsever.om/loate/mpro Hardware aelerator arhteture for smultaneous short-read

More information

A Novel Dynamic and Scalable Caching Algorithm of Proxy Server for Multimedia Objects

A Novel Dynamic and Scalable Caching Algorithm of Proxy Server for Multimedia Objects Journal of VLSI Sgnal Proessng 2007 * 2007 Sprnger Sene + Busness Meda, LLC. Manufatured n The Unted States. DOI: 10.1007/s11265-006-0024-7 A Novel Dynam and Salable Cahng Algorthm of Proxy Server for

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

Significance of Eigenvector Centrality for Routing in a Delay Tolerant Network

Significance of Eigenvector Centrality for Routing in a Delay Tolerant Network Journal of Computatons & Modellng, vol.1, no.1, 2011, 91-100 ISSN: 1792-7625 (prnt), 1792-8850 (onlne) Internatonal Scentfc Press, 2011 Sgnfcance of Egenvector Centralty for Routng n a Delay Tolerant Network

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

ABHELSINKI UNIVERSITY OF TECHNOLOGY Networking Laboratory

ABHELSINKI UNIVERSITY OF TECHNOLOGY Networking Laboratory ABHELSINKI UNIVERSITY OF TECHNOLOGY Networkng Laboratory Load Balanng n Cellular Networks Usng Frst Poly Iteraton Johan an Leeuwaarden Samul Aalto & Jorma Vrtamo Networkng Laboratory Helsnk Unersty of

More information

Outline. Type of Machine Learning. Examples of Application. Unsupervised Learning

Outline. Type of Machine Learning. Examples of Application. Unsupervised Learning Outlne Artfcal Intellgence and ts applcatons Lecture 8 Unsupervsed Learnng Professor Danel Yeung danyeung@eee.org Dr. Patrck Chan patrckchan@eee.org South Chna Unversty of Technology, Chna Introducton

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

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

Interval uncertain optimization of structures using Chebyshev meta-models

Interval uncertain optimization of structures using Chebyshev meta-models 0 th World Congress on Strutural and Multdsplnary Optmzaton May 9-24, 203, Orlando, Florda, USA Interval unertan optmzaton of strutures usng Chebyshev meta-models Jngla Wu, Zhen Luo, Nong Zhang (Tmes New

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

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

Measurement and Calibration of High Accuracy Spherical Joints

Measurement and Calibration of High Accuracy Spherical Joints 1. Introduton easurement and Calbraton of Hgh Auray Spheral Jonts Ale Robertson, Adam Rzepnewsk, Alexander Sloum assahusetts Insttute of Tehnolog Cambrdge, A Hgh auray robot manpulators are requred for

More information

A New Token Allocation Algorithm for TCP Traffic in Diffserv Network

A New Token Allocation Algorithm for TCP Traffic in Diffserv Network A New Token Allocaton Algorthm for TCP Traffc n Dffserv Network A New Token Allocaton Algorthm for TCP Traffc n Dffserv Network S. Sudha and N. Ammasagounden Natonal Insttute of Technology, Truchrappall,

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

Color Texture Classification using Modified Local Binary Patterns based on Intensity and Color Information

Color Texture Classification using Modified Local Binary Patterns based on Intensity and Color Information Color Texture Classfaton usng Modfed Loal Bnary Patterns based on Intensty and Color Informaton Shvashankar S. Department of Computer Sene Karnatak Unversty, Dharwad-580003 Karnataka,Inda shvashankars@kud.a.n

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

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

Link Graph Analysis for Adult Images Classification

Link Graph Analysis for Adult Images Classification Lnk Graph Analyss for Adult Images Classfaton Evgeny Khartonov Insttute of Physs and Tehnology, Yandex LLC 90, 6 Lev Tolstoy st., khartonov@yandex-team.ru Anton Slesarev Insttute of Physs and Tehnology,

More information

Performance Evaluation of TreeQ and LVQ Classifiers for Music Information Retrieval

Performance Evaluation of TreeQ and LVQ Classifiers for Music Information Retrieval Performane Evaluaton of TreeQ and LVQ Classfers for Mus Informaton Retreval Matna Charam, Ram Halloush, Sofa Tsekerdou Athens Informaton Tehnology (AIT) 0.8 km Markopoulo Ave. GR - 19002 Peana, Athens,

More information

Unsupervised Learning

Unsupervised Learning Pattern Recognton Lecture 8 Outlne Introducton Unsupervsed Learnng Parametrc VS Non-Parametrc Approach Mxture of Denstes Maxmum-Lkelhood Estmates Clusterng Prof. Danel Yeung School of Computer Scence and

More information

Cluster ( Vehicle Example. Cluster analysis ( Terminology. Vehicle Clusters. Why cluster?

Cluster (  Vehicle Example. Cluster analysis (  Terminology. Vehicle Clusters. Why cluster? Why luster? referene funton R R Although R and R both somewhat orrelated wth the referene funton, they are unorrelated wth eah other Cluster (www.m-w.om) A number of smlar ndvduals that our together as

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

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

Optimal shape and location of piezoelectric materials for topology optimization of flextensional actuators

Optimal shape and location of piezoelectric materials for topology optimization of flextensional actuators Optmal shape and loaton of pezoeletr materals for topology optmzaton of flextensonal atuators ng L 1 Xueme Xn 2 Noboru Kkuh 1 Kazuhro Satou 1 1 Department of Mehanal Engneerng, Unversty of Mhgan, Ann Arbor,

More information

An Optimal Algorithm for Prufer Codes *

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

More information

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS

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

More information

Performance Evaluation of Information Retrieval Systems

Performance Evaluation of Information Retrieval Systems Why System Evaluaton? Performance Evaluaton of Informaton Retreval Systems Many sldes n ths secton are adapted from Prof. Joydeep Ghosh (UT ECE) who n turn adapted them from Prof. Dk Lee (Unv. of Scence

More information

A Fast Way to Produce Optimal Fixed-Depth Decision Trees

A Fast Way to Produce Optimal Fixed-Depth Decision Trees A Fast Way to Produe Optmal Fxed-Depth Deson Trees Alreza Farhangfar, Russell Grener and Martn Znkevh Dept of Computng Sene Unversty of Alberta Edmonton, Alberta T6G 2E8 Canada {farhang, grener, maz}@s.ualberta.a

More information

A Real-Time Detecting Algorithm for Tracking Community Structure of Dynamic Networks

A Real-Time Detecting Algorithm for Tracking Community Structure of Dynamic Networks A Real-Tme Detetng Algorthm for Trakng Communty Struture of Dynam Networks Jaxng Shang*, Lanhen Lu*, Feng Xe, Zhen Chen, Jaa Mao, Xueln Fang, Cheng Wu* Department of Automaton, Tsnghua Unversty, Beng,,

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

Scalable Parametric Runtime Monitoring

Scalable Parametric Runtime Monitoring Salable Parametr Runtme Montorng Dongyun Jn Patrk O Nel Meredth Grgore Roşu Department of Computer Sene Unversty of Illnos at Urbana Champagn Urbana, IL, U.S.A. {djn3, pmeredt, grosu}@s.llnos.edu Abstrat

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

Multilabel Classification with Meta-level Features

Multilabel Classification with Meta-level Features Multlabel Classfaton wth Meta-level Features Sddharth Gopal Carnege Mellon Unversty Pttsburgh PA 523 sgopal@andrew.mu.edu Ymng Yang Carnege Mellon Unversty Pttsburgh PA 523 ymng@s.mu.edu ABSTRACT Effetve

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

The Simulation of Electromagnetic Suspension System Based on the Finite Element Analysis

The Simulation of Electromagnetic Suspension System Based on the Finite Element Analysis 308 JOURNAL OF COMPUTERS, VOL. 8, NO., FEBRUARY 03 The Smulaton of Suspenson System Based on the Fnte Element Analyss Zhengfeng Mng Shool of Eletron & Mahanal Engneerng, Xdan Unversty, X an, Chna Emal:

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

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

Steps for Computing the Dissimilarity, Entropy, Herfindahl-Hirschman and. Accessibility (Gravity with Competition) Indices

Steps for Computing the Dissimilarity, Entropy, Herfindahl-Hirschman and. Accessibility (Gravity with Competition) Indices Steps for Computng the Dssmlarty, Entropy, Herfndahl-Hrschman and Accessblty (Gravty wth Competton) Indces I. Dssmlarty Index Measurement: The followng formula can be used to measure the evenness between

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

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

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

More information

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

Efficient Routing Algorithms Combining History and Social Predictors in Mobile Social Networks

Efficient Routing Algorithms Combining History and Social Predictors in Mobile Social Networks Effcent Routng Algorthms Combnng and Socal Predctors n Moble Socal Networks Xao Chen 1, Chengyn Lu 2, Cong Lu 2 1 Department of Computer Scence, Texas State Unversty, San Marcos, TX USA 2 Department of

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

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

Design Level Performance Modeling of Component-based Applications. Yan Liu, Alan Fekete School of Information Technologies University of Sydney

Design Level Performance Modeling of Component-based Applications. Yan Liu, Alan Fekete School of Information Technologies University of Sydney Desgn Level Performane Modelng of Component-based Applatons Tehnal Report umber 543 ovember, 003 Yan Lu, Alan Fekete Shool of Informaton Tehnologes Unversty of Sydney Ian Gorton Paf orthwest atonal Laboratory

More information

Cognitive Radio Resource Management Using Multi-Agent Systems

Cognitive Radio Resource Management Using Multi-Agent Systems Cogntve Rado Resource Management Usng Mult- Systems Jang Xe, Ivan Howtt, and Anta Raja Department of Electrcal and Computer Engneerng Department of Software and Informaton Systems The Unversty of North

More information

ARTICLE IN PRESS. Signal Processing: Image Communication

ARTICLE IN PRESS. Signal Processing: Image Communication Sgnal Processng: Image Communcaton 23 (2008) 754 768 Contents lsts avalable at ScenceDrect Sgnal Processng: Image Communcaton journal homepage: www.elsever.com/locate/mage Dstrbuted meda rate allocaton

More information

TCP Performance over Current Cellular Access: A Comprehensive Analysis

TCP Performance over Current Cellular Access: A Comprehensive Analysis TCP Performane over Current Cellular Aess: A Comprehensve Analyss Eneko Atxuteg 1(B), Åke Arvdsson 2, Fdel Lberal 1, Karl-Johan Grnnemo 3, and Anna Brunstrom 3 1 Unversty of the Basque Country (UPV/EHU),

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

Efficient QoS Provisioning at the MAC Layer in Heterogeneous Wireless Sensor Networks

Efficient QoS Provisioning at the MAC Layer in Heterogeneous Wireless Sensor Networks Effcent QoS Provsonng at the MAC Layer n Heterogeneous Wreless Sensor Networks M.Soul a,, A.Bouabdallah a, A.E.Kamal b a UMR CNRS 7253 HeuDaSyC, Unversté de Technologe de Compègne, Compègne Cedex F-625,

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

A Topology-aware Random Walk

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

More information

Hermite Splines in Lie Groups as Products of Geodesics

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

More information

MULTIHOP wireless networks are a paradigm in wireless

MULTIHOP wireless networks are a paradigm in wireless 400 IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 17, NO. 1, JANUARY 2018 Toward Optmal Dstrbuted Node Schedulng n a Multhop Wreless Network Through Local Votng Dmtros J. Vergados, Member, IEEE, Natala

More information

A MPAA-Based Iterative Clustering Algorithm Augmented by Nearest Neighbors Search for Time-Series Data Streams

A MPAA-Based Iterative Clustering Algorithm Augmented by Nearest Neighbors Search for Time-Series Data Streams A MPAA-Based Iteratve Clusterng Algorthm Augmented by Nearest Neghbors Searh for Tme-Seres Data Streams Jessa Ln 1, Mha Vlahos 1, Eamonn Keogh 1, Dmtros Gunopulos 1, Janwe Lu 2, Shouan Yu 2, and Jan Le

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

LOCAL BINARY PATTERNS AND ITS VARIANTS FOR FACE RECOGNITION

LOCAL BINARY PATTERNS AND ITS VARIANTS FOR FACE RECOGNITION IEEE-Internatonal Conferene on Reent Trends n Informaton Tehnology, ICRTIT 211 MIT, Anna Unversty, Chenna. June 3-5, 211 LOCAL BINARY PATTERNS AND ITS VARIANTS FOR FACE RECOGNITION K.Meena #1, Dr.A.Suruland

More information

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization

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

More information

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

with `ook-ahead for Broadcast WDM Networks TR May 14, 1996 Abstract

with `ook-ahead for Broadcast WDM Networks TR May 14, 1996 Abstract HPeR-`: A Hgh Performance Reservaton Protocol wth `ook-ahead for Broadcast WDM Networks Vjay Svaraman George N. Rouskas TR-96-06 May 14, 1996 Abstract We consder the problem of coordnatng access to the

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

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

Determining the Optimal Bandwidth Based on Multi-criterion Fusion

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

More information

A Semi-Distributed Load Balancing Architecture and Algorithm for Heterogeneous Wireless Networks

A Semi-Distributed Load Balancing Architecture and Algorithm for Heterogeneous Wireless Networks A Sem-Dstrbuted oad Balancng Archtecture and Algorthm for Heterogeneous reless Networks Md. Golam Rabul Ala Choong Seon Hong * Kyung Hee Unversty, Korea rob@networkng.khu.ac.kr, cshong@khu.ac.kr Abstract

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

On Maximizing the Lifetime of Delay-Sensitive Wireless Sensor Networks with Anycast

On Maximizing the Lifetime of Delay-Sensitive Wireless Sensor Networks with Anycast On Maxmzng the Lfetme of Delay-Senstve Wreless Sensor Networks wth Anycast Joohwan Km, Xaojun Ln, Ness B. Shroff, and Prasun Snha School of Electrcal and Computer Engneerng, Purdue Unversty Departments

More information

CS 534: Computer Vision Model Fitting

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

More information

Improving Low Density Parity Check Codes Over the Erasure Channel. The Nelder Mead Downhill Simplex Method. Scott Stransky

Improving Low Density Parity Check Codes Over the Erasure Channel. The Nelder Mead Downhill Simplex Method. Scott Stransky Improvng Low Densty Party Check Codes Over the Erasure Channel The Nelder Mead Downhll Smplex Method Scott Stransky Programmng n conjuncton wth: Bors Cukalovc 18.413 Fnal Project Sprng 2004 Page 1 Abstract

More information

DUE to the recent popularization of hand-held mobile

DUE to the recent popularization of hand-held mobile IEEE TRANSACTIONS ON MOBILE COMPUTING, VOL. 15, NO. 4, APRIL 2016 803 Contact Duraton Aware Data Replcaton n DTNs wth Lcensed and Unlcensed Spectrum Jng Zhao, Student Member, IEEE, Xuejun Zhuo, Qnghua

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

Pattern Classification: An Improvement Using Combination of VQ and PCA Based Techniques

Pattern Classification: An Improvement Using Combination of VQ and PCA Based Techniques Ameran Journal of Appled Senes (0): 445-455, 005 ISSN 546-939 005 Sene Publatons Pattern Classfaton: An Improvement Usng Combnaton of VQ and PCA Based Tehnques Alok Sharma, Kuldp K. Palwal and Godfrey

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

Virtual Machine Migration based on Trust Measurement of Computer Node

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

More information

TAR based shape features in unconstrained handwritten digit recognition

TAR based shape features in unconstrained handwritten digit recognition TAR based shape features n unonstraned handwrtten dgt reognton P. AHAMED AND YOUSEF AL-OHALI Department of Computer Sene Kng Saud Unversty P.O.B. 578, Ryadh 543 SAUDI ARABIA shamapervez@gmal.om, yousef@s.edu.sa

More information

Routing in Delay Tolerant Networks Comprising Heterogeneous Node Populations

Routing in Delay Tolerant Networks Comprising Heterogeneous Node Populations Routng n Delay Tolerant Networks Comprsng Heterogeneous Node Populatons Thrasyvoulos Spyropoulos, Member, IEEE, Therry Turlett, Member, IEEE, and Kata Obraczka, Member, IEEE Abstract Communcaton networks

More information

Comparison of Heuristics for Scheduling Independent Tasks on Heterogeneous Distributed Environments

Comparison of Heuristics for Scheduling Independent Tasks on Heterogeneous Distributed Environments Comparson of Heurstcs for Schedulng Independent Tasks on Heterogeneous Dstrbuted Envronments Hesam Izakan¹, Ath Abraham², Senor Member, IEEE, Václav Snášel³ ¹ Islamc Azad Unversty, Ramsar Branch, Ramsar,

More information

5 The Primal-Dual Method

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

More information

Concurrent Apriori Data Mining Algorithms

Concurrent Apriori Data Mining Algorithms Concurrent Apror Data Mnng Algorthms Vassl Halatchev Department of Electrcal Engneerng and Computer Scence York Unversty, Toronto October 8, 2015 Outlne Why t s mportant Introducton to Assocaton Rule Mnng

More information

Routing in Degree-constrained FSO Mesh Networks

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

More information

A 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

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

Bottom-Up Fuzzy Partitioning in Fuzzy Decision Trees

Bottom-Up Fuzzy Partitioning in Fuzzy Decision Trees Bottom-Up Fuzzy arttonng n Fuzzy eson Trees Maej Fajfer ept. of Mathemats and Computer Sene Unversty of Mssour St. Lous St. Lous, Mssour 63121 maejf@me.pl Cezary Z. Janow ept. of Mathemats and Computer

More information

Progressive scan conversion based on edge-dependent interpolation using fuzzy logic

Progressive scan conversion based on edge-dependent interpolation using fuzzy logic Progressve san onverson based on edge-dependent nterpolaton usng fuzzy log P. Brox brox@mse.nm.es I. Baturone lum@mse.nm.es Insttuto de Mroeletróna de Sevlla, Centro Naonal de Mroeletróna Avda. Rena Meredes

More information

Fuzzy Modeling for Multi-Label Text Classification Supported by Classification Algorithms

Fuzzy Modeling for Multi-Label Text Classification Supported by Classification Algorithms Journal of Computer Senes Orgnal Researh Paper Fuzzy Modelng for Mult-Label Text Classfaton Supported by Classfaton Algorthms 1 Beatrz Wlges, 2 Gustavo Mateus, 2 Slva Nassar, 2 Renato Cslagh and 3 Rogéro

More information

A Clustering Algorithm for Chinese Adjectives and Nouns 1

A Clustering Algorithm for Chinese Adjectives and Nouns 1 Clusterng lgorthm for Chnese dectves and ouns Yang Wen, Chunfa Yuan, Changnng Huang 2 State Key aboratory of Intellgent Technology and System Deptartment of Computer Scence & Technology, Tsnghua Unversty,

More information

FUZZY SEGMENTATION IN IMAGE PROCESSING

FUZZY SEGMENTATION IN IMAGE PROCESSING FUZZY SEGMENTATION IN IMAGE PROESSING uevas J. Er,, Zaldívar N. Danel,, Roas Raúl Free Unverstät Berln, Insttut für Inforat Tausstr. 9, D-495 Berln, Gerany. Tel. 0049-030-8385485, Fax. 0049-030-8387509

More information

A Semi-parametric Approach for Analyzing Longitudinal Measurements with Non-ignorable Missingness Using Regression Spline

A Semi-parametric Approach for Analyzing Longitudinal Measurements with Non-ignorable Missingness Using Regression Spline Avalable at http://pvamu.edu/aam Appl. Appl. Math. ISSN: 93-9466 Vol., Issue (June 5), pp. 95 - Applatons and Appled Mathemats: An Internatonal Journal (AAM) A Sem-parametr Approah for Analyzng Longtudnal

More information

Steganalysis of DCT-Embedding Based Adaptive Steganography and YASS

Steganalysis of DCT-Embedding Based Adaptive Steganography and YASS Steganalyss of DCT-Embeddng Based Adaptve Steganography and YASS Qngzhong Lu Department of Computer Sene Sam Houston State Unversty Huntsvlle, TX 77341, U.S.A. lu@shsu.edu ABSTRACT Reently well-desgned

More information

Programming in Fortran 90 : 2017/2018

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

More information

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