Irregular Cellular Learning Automata and Its Application to Clustering in Sensor Networks

Size: px
Start display at page:

Download "Irregular Cellular Learning Automata and Its Application to Clustering in Sensor Networks"

Transcription

1 Irregular Cellular Learnng Automata and Its Applcaton to Clusterng n Sensor Networks M. Esnaashar 1, M. R. Meybod 1,2 1 Soft Computng Laboratory, Computer Engneerng and Informaton Technology Department Amrkabr Unversty of Technology, Tehran, Iran (Esnaashar,mmeybod)@aut.ac.r 2 Insttute for Studes n Theoretcal Physcs and Mathematcs (IPM) School of Computer Scence, Tehran, Iran Abstract In the frst part of ths paper, we propose a generalzaton of cellular learnng automata (CLA) called rregular cellular learnng automata (ICLA) whch removes the restrcton of rectangular grd structure n tradtonal CLA. Ths generalzaton s expected because there are a number of applcatons whch cannot be adequately modeled wth rectangular grds. One category of such applcatons s n the area of wreless sensor networks. In these networks, nodes are usually scattered randomly throughout the envronment, so no regular structure can be assumed for modelng ther behavor. In the second part of the paper, based on the proposed model we desgn a clusterng algorthm for sensor networks. Smulaton results show that the proposed clusterng algorthm s very effcent and outperforms smlar exstng methods. Index Terms Irregular cellular learnng automata, Cellular learnng automata, Sensor networks, Clusterng algorthm 1. Introducton Cellular automata are mathematcal models for systems consstng of large number of smple dentcal components wth local nteractons. CA s a non-lnear dynamcal system n whch space and tme are dscrete. It s called cellular because t s made up of cells lke ponts n a lattce or lke squares of checker boards, and t s called automata because t follows a smple rule [2]. The smple components act together to produce complcated patterns of behavor. Cellular automata perform complex computatons wth a hgh degree of effcency and robustness. They are especally sutable for modelng natural systems that can be descrbed as massve collectons of smple objects nteractng locally wth each other [3, 4]. Informally, a d-dmensonal CA conssts of an nfnte d-dmensonal lattce of dentcal cells. Each cell can assume a state from a fnte set of states. The cells update ther states synchronously on dscrete steps accordng to a local rule. The new state of each cell depends on the prevous states of a set of cells, ncludng the cell tself, and consttutes ts neghborhood [5]. The state of all cells n the lattce s descrbed by a confguraton. A confguraton can be descrbed as the state of the whole lattce. The rule and the 14

2 ntal confguraton of the CA specfy the evoluton of CA that tells how each confguraton s changed n one step. Formally, a CA can be defned as follows: Defnton 1. A d-dmensonal cellular automata s a structure A = ( Z d, Φ, N, F ), where Z d s a lattce of d-tuples of nteger numbers. Each cell n the d-dmensonal lattce Z d s represented by a d-tuple (z 1, z 2,..., z d ). Φ = {1,..., m} s a fnte set of states. N = { x x,..., } s a fnte subset of Z d called the neghborhood vector, where d 1, 2 xm x Z. The Neghborhood vector determnes the relatve poston of the neghborng lattce cells from any gven cell u n the lattce Z d. The neghbors of a partcular cell u are the set of cells. We assume that there exsts a neghborhood functon { u + x = 1,2,... m} N (u) that maps a cell u to the set of ts neghbors, that s N ( u) = ( u + x1, u + x2,..., u + xm). For sake of smplcty, we assume that the frst element of the neghborhood vector (.e. x ) s 1 equal to d-tuple (0, 0,..., 0) or equvalently u + x1 = u. The neghborhood functon N (u) must satsfy the followng two condtons: - u N(u ) for all u Z d. - u1 N ( u2) u2 N ( u1) for all u u Z d 1, 2 F : Φ m Φ s the local rule of the cellular automata. It computes the new state for each cell from the current states of ts neghbors. On the other hand, learnng automaton s a smple entty operates n an unknown random envronment. In a smple form, the automaton has a fnte set of actons to choose from and at each stage, ts choce (acton) depends upon ts acton probablty vector. For each acton chosen by the automaton, the envronment gves a renforcement sgnal wth fxed unknown probablty dstrbuton. The automaton then updates ts acton probablty vector dependng upon the renforcement sgnal at that stage, and evolves to the some fnal desred behavor. A class of learnng automata s called varable structure learnng automata and are represented by quadruple { α, β, p, T} n whch α = { α 1, α 2, L, α r } represents the acton set of the automata, β = { β1, β 2, L, β r } represents the nput set, p = { p1, p2, L, pr} represents the acton probablty set, and fnally p( n + 1) = T[ α( n), β ( n), p( n)] represents the learnng algorthm. Let α be the acton chosen at tme n, then the recurrence equaton for updatng p s defned as for favorable responses, and p ( n + 1) = p ( n) + a.(1 p ( n)) p j ( n + 1) = p j ( n) a. p j ( n) p ( n + 1) = (1 b). p ( n) b p j ( n + 1) = + (1 b) p j ( n) r 1 j j j j for unfavorable ones. In these equatons, a and b are reward and penalty parameters respectvely. If a = b, learnng algorthm s called L 1 R P, f a << b, t s called L 2 R ε P, and f b = 0, t s called L 3. R I (1) (2) 1 Lnear Reward-Penalty 2 Lnear Reward epslon Penalty 3 Lnear Reward Inacton 15

3 Cellular learnng automata [1], whch s a combnaton of cellular automata (CA) and learnng automata (LA), s a powerful mathematcal model for many decentralzed problems and phenomena. The basc dea of CLA, whch s a subclass of stochastc CA, s to use learnng automata to adjust the state transton probablty of stochastc CA. A CLA s a CA n whch a learnng automaton s assgned to every cell. The learnng automaton resdng n a partcular cell determnes ts acton (state) on the bass of ts acton probablty vector. Lke CA, there s a rule that the CLA operates under. The rule of the CLA and the actons selected by the neghborng LAs of any partcular LA determne the renforcement sgnal to the LA resdng n a cell. The neghborng LAs of any partcular LA consttute the local envronment of that cell. The local envronment of a cell s nonstatonary because the acton probablty vectors of the neghborng LAs vary durng evoluton of the CLA. The basc dea of CLA, whch s a subclass of stochastc CA, s to use learnng automata to adjust the state transton probablty of stochastc CA. A d- dmensonal CLA s formally defned below. Defnton 2. A d-dmensonal cellular learnng automata s a structure A = ( Z d, Φ, A, N, F ) where Z d s a lattce of d-tuples of nteger numbers. Φ s a fnte set of states. A s the set of LA each of whch s assgned to one cell of the CLA. N = { x x,..., } s a fnte subset of Z d called the neghborhood vector, where d 1, 2 xm x Z. F : Φ m β s the local rule of the cellular learnng automata, where β s the set of values that the renforcement sgnal can take. It computes the renforcement sgnal for each LA based on the actons selected by the neghborng LA. It computes the new state for each cell from the current states of ts neghbors. CLA has found many applcatons such as mage processng [6, 7, 8, 9], rumor dffuson [10], modelng of commerce networks [8], channel assgnment n cellular networks [11] and VLSI placement [12], to menton a few. In the frst part of ths paper, we propose a generalzaton of cellular learnng automata (CLA) called rregular cellular learnng automata (ICLA) whch removes the restrcton of rectangular grd structure n tradtonal CLA. Ths generalzaton s expected because there are applcatons whch cannot be adequately modeled wth rectangular grds. One category of such applcatons s n the area of wreless sensor networks. In these networks, nodes are usually scattered randomly throughout the envronment, so no regular structure can be assumed for modelng ther behavor. In the second part of the paper, based on the proposed model, we desgn a clusterng algorthm for sensor networks. The proposed clusterng algorthm tres to maxmze the probablty of selectng nodes wth hgher energy and hgher number of neghbors as cluster heads. To evaluate the performance of the proposed algorthm several experments have been conducted usng dfferent clusterng algorthms and the results are compared wth the proposed algorthm Smulaton results show that the proposed clusterng algorthm s very effcent n terms clusterng formaton n the network) and outperform smlar exstng methods. The rest of ths paper s organzed as follows. In secton 2 rregular cellular learnng automata s presented. Secton 3 gves an overvew on clusterng algorthms for wreless sensor networks. The proposed ICLA based clusterng algorthm s descrbed n secton 4. Smulaton results are gven n secton 5. Secton 6 s the concluson. 16

4 2. Irregular Cellular Learnng Automata The dea of rregular cellular automata was suggested n md 80s [13], but due to the computatonally ntensve operatons requred to search rregular neghborhood, t has been receved less attenton snce then. In an nformal way, ICA s a confguraton of ponts n the space wth no pror restrcton. Each pont has a number of other ponts as ts neghbors. The few examples of ICA all use Vorono polygons or the related Delaunay trangulaton to dvde space and determne the neghbors of each pont [14]. Vorono polygons dvde space nto regons surroundng objects such that any pont n an object's polygon s closer to that object than to any other object, whle Delaunay trangulaton s a trangulaton of the ponts n a Vorono dagram where the crcumcrcle of each trangle s an empty trangle. Irregular cellular learnng automata (Fgure 1) s a combnaton of rregular cellular automata (ICA) and learnng automata. We defne ICLA as an undrected graph n whch, each vertex represents a cell whch s equpped wth a learnng automaton. The learnng automaton resdng n a partcular cell determnes ts state (acton) on the bass of ts acton probablty vector. Lke CLA, there s a rule that the ICLA operate under. The rule of the CLA and the actons selected by the neghborng LAs of any partcular LA determne the renforcement sgnal to the LA resdng n a cell. The neghborng LAs of any partcular LA consttute the local envronment of that cell. The local envronment of a cell s non-statonary because the acton probablty vectors of the neghborng LAs vary durng evoluton of the ICLA. An ICLA s formally defned below., where A = ( G < E, V >, Φ, A, F) Defnton 3. An rregular cellular learnng automata s a structure G s an undrected graph, wth V as the set of vertces and E as the set of edges. Φ s a fnte set of states. A s the set of LA each of whch s assgned to one cell of the ICLA. s the local rule of the rregular cellular learnng automata n each vertex j, where F : Φ j β Φ j = { Φ (, j) E} + { Φ j} s the set of states of all neghbors of j and β s the set of values that the renforcement sgnal can take. β computes the renforcement sgnal for LA j based on the actons selected by the neghborng LA. Fgure 1. Irregular cellular learnng automata Note that n the defnton of ICLA, no explct defnton of neghborhood of each cell s gven. Ths s because neghborhood n ICLA s mplctly defned n defnton of the graph G. 17

5 In what follows, we consder ICLA wth n cells. The learnng automaton A whch has a fnte acton set α s assocated to cell (for =1,2,,,,n) of the ICLA. Let the cardnalty of α be m. The state of the ICLA represented by p= (p 1, p 2,..., p n ), where s the acton p = ( p1, p2,..., pm ) probablty vector of A. The operaton of the ICLA takes place as the followng teratons. At teraton k, each learnng automaton chooses an acton. Let α be the acton chosen by A. Then all learnng automata α receve a renforcement sgnal. Let be the renforcement sgnal receved by A β β. Ths renforcement sgnal s produced by the applcaton of local rule. Fnally, each LA F( Φ ) β updates ts acton probablty vector on the bass of the suppled renforcement sgnal and the chosen acton by the cell. Ths process contnues untl the desred result s obtaned. 3. Clusterng Algorthms In ths secton, we wll brefly overvew some of the exstng clusterng algorthms for wreless sensor networks. One of the frst clusterng algorthms ntroduced for sensor networks s the low energy adaptve clusterng herarchy (LEACH) algorthm [15]. The operaton of LEACH s separated nto two phases: the setup phase and the steady state phase. Durng the setup phase, a predetermned fracton of nodes p, elect themselves as cluster heads by comparng a chosen random number wth a predefned threshold. After the cluster heads have been elected, they broadcast an advertsement message to the rest of the nodes n the network that they are the new cluster heads. Upon recevng ths advertsement, all the noncluster head nodes decde on the cluster to whch they want to belong, based on the sgnal strength of the advertsement. The noncluster head nodes nform the approprate cluster heads that they wll be members of the cluster. There are algorthms whch elect cluster heads based on a certan crteron such as number of neghbor nodes [16, 17], or remanng energy of the nodes [18]. In the algorthm presented n [19], each node wats for a random duraton. After ths perod, f no message s receved from a certan cluster head, the node states tself as a new cluster head. In some other methods, such as the one n [20], cluster heads are specfed pror to network deployment. These specfed nodes are placed n certan postons and other nodes are scattered around them. After the network deployment, these nodes try to form clusters havng balanced traffc load. Ths means that fewer nodes are assgned to the clusters closer to the snk node and more nodes are assgned to the clusters far from the snk. In the sensor network consdered n [21], two classes of nodes are avalable; sensor nodes and aggregator nodes. Aggregator nodes are placed n certan poston and play the role of cluster heads. In ths paper three dfferent algorthms are presented for scatterng sensor nodes around the aggregators. In [22], snk node s assumed to be a moble node whch queres data from dfferent part of the network based on ts dstance from dfferent nodes. For ths reason, t s requred to have a dynamc clusterng algorthm whch can adapt tself to the changng poston of the moble snk, consderng the overall energy consumpton n the network. A number of dfferent routng algorthms for sensor networks are surveyed n [23, 24]. One major class of these algorthms s the herarchcal routng. In herarchcal routng algorthms, at 18

6 frst a clusterng scheme s appled to form a knd of herarchy and afterward, the process of routng s dvded nto nter and ntra-cluster phases. Some examples of these algorthms are "Fxed sze cluster routng", TEEN 4, and APTEEN 5. AIMRP 6 [25] assumes that the snk node s placed n the center of the network. In the setup phase, snk ntates a Tre message to the nodes n the radus of a around tself. These nodes consttute the frst level. These nodes rebroadcast the Ter message to the nodes n the radus of a around themselves. Ths process contnues untl all nodes n the network fnd ther level n the herarchy to the snk node. In [26] an algorthm called Hybrd Energy-Effcent Approach (HEED) 7 has been gven. Ths algorthm has four prmary goals: () prolongng network lfetme, () termnatng the clusterng process wthn a constant number of teratons/steps, () mnmzng control overhead (to be lnear n the number of nodes), and (v) producng well-dstrbuted cluster heads and compact clusters. Before a node starts executng HEED, t sets ts probablty of becomng a cluster head, CH based on ts resdual energy. Usng ths probablty, each node decdes to become a prob tentatve cluster head or not. Tentatve cluster heads, declare themselves to ther neghbors usng a cluster-head message, and double ther CH probablty. Once prob CH n a node reaches 1, prob that node becomes a fnal cluster head. Nodes whch are not cluster head (tentatve or fnal), upon recevng a cluster-head message, jon to the declared cluster. Nodes whch do not receve ant cluster-head message after a whle, decde to become cluster head. In [27] an enhancement on HEED algorthm n whch, only nodes wth resdual energy hgher than a specfed threshold can become cluster head s presented. Also at the end of the algorthm, when some nodes do not jon to any cluster yet, unlke HEED n whch all these nodes become cluster head, HEED algorthm s executed agan for electng cluster heads among them. 4. The Proposed Algorthm Three man objectves of a good clusterng method for sensor networks are ) producng welldstrbuted cluster heads, ) maxmzng the electon probablty of hgh energy level nodes as cluster heads and ) mnmzng the number of sparse clusters. In ths secton an attempt has been made to desgn a clusterng algorthm to satsfy all three objectves. The frst objectve s mplctly satsfed by the use of ICLA for the clusterng. Ths s because n each neghborhood of the ICLA, at least one cluster head wll be elected and ths wll guarantee the well-dstrbuton of cluster heads among the network. The other two objectves wll be satsfed through a proper selecton of local rule for ICLA. Fgures 2 and 3 gve the pseudo code for the proposed clusterng algorthm. Based on the senor network topology an ICLA s created, that s an ICLA s mapped nto the network n such a way that each node n the sensor network be mapped nto a node n the ICLA. Any two nodes n the sensor networks that are close enough to hear each other s sgnal have an edge n the correspondng graph G of the ICLA. 4 Threshold-Senstve Energy-Effcent Protocol 5 Adaptve Perodc TEEN 6 Address-lght Integrated MAC, and Reportng Protocol 7 Hybrd Energy-Effcent Approach 19

7 Fgure 2. Pseudo code for the ICLA-Clusterng algorthm. Each node follows ths code separately The proposed algorthm has two phases. In the ntal phase, snk node ntates 'Start- Clusterng' message throughout the network. Each node, upon recevng ths message, rebroadcasts t to ts neghbors and then enters the clusterng phase whch s performed usng ICLA. LA n each node has two dfferent actons a 0 =0 and a 1 =1. a 1 ndcates that the node should declare tself as the cluster head, and a 0 ndcates the opposte decson. At the start of the clusterng phase, each node sets the ntal probablty of these two actons to.5 and enters the clusterng teratons. At each teraton k, the LA n each node chooses one of ts actons a 0 or a 1. Each node then forwards the selected acton of ts LA to ts neghbors. Afterward, node, wats for a specfed duraton to receve ts neghbors messages ndcatng ther selected actons. At the end of ths duraton, each node calculates the response β based on the local rule of the ICLA. Ths local rule s desgned n such a way that to both maxmzng the electon probablty of hgh energy level nodes as cluster heads and mnmzng the number of sparse clusters. The local rule for node can be defned nformally as follows: 20

8 Fgure 3. Pseudo code for the man loop each node follows n the proposed clusterng method - If selected acton of LA n node s a 1 (node declares tself as a cluster head) then o If no neghbor of node declares tself as a cluster head, a favorable response wll be generated. o If some neghbors of node declare themselves as cluster head, but node has more resdual energy and more number of neghbors among them, a favorable response wll be generated. o Otherwse, an unfavorable response wll be generated. - If selected acton of LA n node s a 0 (node doesn't declare tself as a cluster head) then o If no neghbor of node declares tself as a cluster head, au unfavorable response wll be generated. o If some neghbors of node declare themselves as cluster head, but node has more resdual energy and more number of neghbors among them, an unfavorable response wll be generated. o Otherwse, a favorable response wll be generated. The local rule can be defned formally as follows. Let a be the acton selected by node and a j be the acton selected by the neghbor node j (whch can be ether a 0 or a 1 ). Equaton (3) gves the local rule of the ICLA. 0; β = 1; N a ( P a j ) + (1 a )) + j= 1 N (1 a ).( P a j ) (1 a ) 0 j= 1 Otherwse (3) 21

9 In (3), N s the number of neghbors of node and P s a parameter whch specfes the sutablty of node for becomng cluster head among ts neghbors. P s defned usng (4) gven below. P = α [ EnergyLevel EnergyLevel Max ] + (4) (1 α) [ N N Max ] K In equaton (4), EnergyLevel s the remanng energy level of the node. α s a coeffcent whch controls the effect of energy level and number of neghbors on the selected cluster heads. Hgher value of α ndcates cluster heads wth more energy level, and lower value of t ndcates cluster heads wth hgher number of neghbors. K s a constant whch controls the sutablty of node for becomng cluster head among ts neghbors. Large values of K ndcate more sutablty whch results n few clusters. Ths may leave some parts of the network wth no cluster. On the other hand, choosng small values of K results n too many clusters whch leads to more number of sparse clusters. Fnally, EnergyLevel Max and N Max are defned usng equatons (5) and (6) respectvely. EnergyLevel Max = Max ( EnergyLevel j ) 1 j N (5) N Max = Max ( N j ) 1 j N In these equatons, EnergyLevel j s the remanng energy level of the neghbor j and N j s the number of neghbors of the node j. Ths nformaton about neghbor nodes are exchanged between nodes perodcally va smple announcement messages. For postve values of P, node would be a better choce than ts neghbors for becomng cluster head, but negatve values of P ndcates that some neghbors are better choces than for ths purpose. Based on the response specfed by the equaton (3), LA resdes n node rewards (penalzes) ts selected acton a usng equatons (7)((8)), respectvely. (6) p ( n + 1) = p ( n) + a.(1 p ( n)) p ( n + 1) = 1 p ( n + 1) j (7) p ( n + 1) = (1 b) p ( n) p ( n + 1) = 1 p (n+ 1) j (8) In the above two equatons, a s the reward rate and b s the punshment rate. In the smulatons conducted, equatons (9) and (10) are used to compute a and b, respectvely. a = a [ Alpha ( N. EnergyLeve l )] + N (1 a ) [ a j ( N EnergyLeve l )] j= 1 (9) 22

10 N b = a [ a j j = 1 ( N EnergyLeve l )] + (1 a ) [ Alpha ( N. EnergyLeve l )] In equatons (9) and (10), Alpha s a parameter whch s defned through equaton (11). Ths parameter s used to control the value of a and b so that they wll be always less than 1. (10) Alpha= mn( EnergyLeve l,1 ( N EnergyLeve l )) (11) Ths process contnues untl the acton probablty of one of the actons of the LA n each node passes a certan threshold near to 1. Nodes for whch, the probablty of selectng acton a 1 of ther LA passes ths threshold, become cluster heads and broadcast messages ndcatng that they are fnal cluster heads. Other nodes wat for the fnal cluster head messages from ther neghbors and jon to the one whch has hgher energy level and hgher number of neghbors. 5. Expermental Results To evaluate the performance of the proposed method several experments have been conducted. In these experments, the proposed method s compared wth the basc HEED clusterng method proposed n [26] and ts extenson gven n [27]. Also, we study the convergence behavor of learnng automata n dfferent nodes of the network. All smulatons have been mplemented usng NS2 smulator. We use IEEE as the MAC layer protocol. Nodes of the network are placed randomly on a 2 dmensonal grd. In all experments, α s set to.3 and K s set to 1.2. Smulatons are performed for 50, 100, 200, 300, 400, and 500 nodes. The results are averaged over 20 runs. Experment 1: In ths experment, we study the cluster formaton behavor of the proposed method n comparson wth two other methods mentoned earler. For ths purpose, we make use of the followng metrcs: total number of clusters formed n the network, number of sparse clusters (whch s defned as the clusters havng less than 3 nodes) and mean number of nodes n each cluster. Fgure 4 depcts the number of clusters formed n the network, number of sparse clusters s depcted n fgure 5 and fgure 6 depcts the mean number of nodes n each cluster. These fgures show that the cluster formaton n the proposed method s better than the other two methods. Ths s because t forms less number of clusters and more nodes n each cluster on average. Also, number of sparse clusters n the proposed method s less than that n two other methods. Ths means that the proposed method can better dstrbute cluster heads among the sensor network. Experment 2: In ths experment, mean energy level of cluster heads at the end of clusterng algorthm and pror to any data transton s studed. Intal energy level of each node s selected randomly between 60% and 100% of the maxmum energy level (full charged). For estmatng the energy consumpton of each node, we assume that sendng a sngle packet wll consume.005% of the maxmum energy level and recevng a sngle packet wll consume.001% of the maxmum energy level. Energy consumpton n dle state s assumed to be 0. Fgure 7 depcts the results for the proposed method n comparson to other two methods. Ths fgure shows that the proposed method outperforms HEED and ts extenson n terms of energy level of cluster heads wth the factor of approxmately

11 Number of Clusters HEED ExtendedHEED ICLA Number of Nodes Fgure 4. Total number of clusters formed n the network for the proposed method and other methods Number of Sparse Clusters HEED ExtendedHEED ICLA Number of Nodes Fgure 5. Number of sparse clusters formed n the network for the proposed method and other methods Experment 3: Ths experment s conducted to better understand the convergence behavor of learnng automata resdng n the nodes of the network. The sensor network used n ths experment s composed of 50 nodes whch are scattered randomly throughout a 2 dmensonal grd of 1000x1000 meter. Fgure 8 depcts ths network and fgure 9 depcts clusters formed n t usng the proposed algorthm. Mean Number of Nodes n Each Cluster HEED ExtendedHEED ICLA Number of Nodes Fgure 6. Mean number of nodes n each cluster for the proposed method and other methods 24

12 Mean Remanng Energy Level n Cluster Heads HEED ExtendedHEED ICLA Number of Nodes Fgure 7. Mean remanng energy level n cluster heads for the proposed method and other methods Fgure 8. Sensor network of 50 nodes n a grd of 1000x1000 meter Fgure 9. Sensor network of fgure 8 whch s clustered usng proposed method We study the acton probabltes of two automata resdng n two nodes; one s a cluster head and the other s a cluster member. Fgure 10 shows convergence of the acton probablty of acton "Be Head" to 1 n the learnng automaton resdng n the cluster head node to the acton "Be Head". Fgure 11 shows convergence of the acton probablty of acton "Be Member" to 1 n the learnng automaton resdng n the other node. 25

13 1.2 1 Be Head Be Member Probablty Tme Fgure 10. Acton probabltes of the learnng automaton n the cluster head node Probablty Be Head Be Member Tme Fgure 11. Acton probabltes of the learnng automaton n the cluster member node 6. Concluson In ths paper we propose rregular cellular learnng automata as a generalzaton to the cellular learnng automata whch removes the restrcton of rectangular grd structure n tradtonal CLA. There are a number of problems and applcatons whch cannot be adequately modeled wth rectangular grds, one of whch s the area of wreless sensor networks. In these networks, nodes are usually scattered randomly throughout the envronment, so no regular structure can be assumed for modelng ther behavor. Based on the proposed ICLA model, we present a novel clusterng algorthm for sensor networks n whch, each node n the network s a cell n the ICLA, and hence equpped wth an LA. Any two nodes that are close enough together to hear each other s sgnal have a common edge n the graph G of the ICLA. LA n each node has two actons one of whch ndcates that the node should declare tself as a cluster head and the other one ndcates the opposte decson. Usng local rule of the ICLA, acton probablty of one of these two actons converges to 1 n each of these LA. Nodes for whch probablty of acton "Be Head" of ther LA converges to 1, becomes cluster head and other nodes become cluster members. It has been shown through smulatons that the proposed clusterng algorthm outperforms other smlar methods n terms of cluster formaton as well as remanng energy level of cluster heads at the end of the clusterng. References [1] H. Beyg, M. R. Meybod, "A Mathematcal Framework for Cellular Learnng Automata", Advances n Complex Systems, Vol. 7, Nos. 3 & 4, pp , September & December

14 [2] E. Fredkn, "Dgtal machne: A nformatonal process based on reversble cellular automata", Physca D45, pp , [3] M. Mtchell, "Computaton n cellular automata: A selected revew", Techncal report, Santa Fe Insttute, Santa Fe, NM, USA, September [4] N. H. Packard, S. Wolfram, "Two-dmensonal cellular automata", Journal of Stat. Phys. 38, pp , [5] J. Kar, "Reversblty of 2D cellular automata s undecdable", Physca D45, pp , [6] M. R. Kharazm, M. R. Meybod, "Image Segmentaton Usng Cellular Learnng Automata", n Proc. of 10th Iranan Conf. on Electrcal Engneerng, ICEE-96, Tabrz, Iran, May [7] M. R. Kharazm, M. R. Meybod, "Image Restoraton Usng Cellular Learnng Automata", n Proc. of 2nd Iranan Conf. on Machne Vson, Image Processng and Applcatons, Tehran, Iran, pp , [8] M. R. Meybod, H. Beyg, M. Taherkhan, "Cellular Learnng Automata and ts Applcatons", Journal of Scence Tech., Sharf (Sharf Unversty of Technology, Tehran, Iran), pp , [9] M. R. Meybod, M. R. Khojaste, "Applcaton of Cellular Learnng Automata n Modelng of Commerce Networks", n Proc. of 6th Annual Intl. Computer Socety of Iran Computer Conf., CSICC- 2001, Isfehan, Iran, pp , February [10] M. R. Meybod, M. Taherkhan, "Applcaton of Cellular Learnng Automata n Modelng of Rumor Dffuson", n Proc. of 9th Conf. on Electrcal Engneerng, Power and Water Insttute of Technology, Tehran, Iran, pp , May [11] H. Beyg, M. R. Meybod, "A Self-Organzng Channel Assgnment Algorthm: A Cellular Learnng Automata Approach", Sprnger-Verlag Lecture Notes n Computer Scence, Vol. 2690, pp , [12] H. Beyg, M. R. Meybod, "Call Admsson n Cellular Networks: A Learnng Automata Approach", Sprnger-Verlag Lecture Notes n Computer Scence, Vol. 2510, pp , [13] H. Couclels, "Cellular Worlds - a framework for modelng mcro-macro dynamcs", Envronment and Plannng, pp , [14] D. Stevens, "Integraton of an Irregular Cellular Automata Approach and Geographc Informaton Systems for Hgh-Resoluton Modelng of Urban Growth", Msc. Thess, Department of Geography, Unversty of Toronto, [15] W. Henzelman, A. Chandrakasan, H. Balakrshnan, Energy Effcent Communcaton Protocol for Wreless Mcrosensor Networks, Intl. Conf. on System Scences, Hawa, January [16] C. Y. Wen, W. A. Sethares, Automatc Decentralzed Clusterng for Wreless Sensor Networks, EURASIP Journal on Wreless Communcatons and Networkng, pp , March [17] K. Shn, A. Abraham, S. Y. Han, Self Organzng Sensor Networks Usng Intellgent Clusterng, Intl. Conf. on Computatonal Scence and Applcatons, UK, M. Gavrlova et al. (Eds.), Lecture Notes n Computer Scence (LNCS 3983), Sprnger Verlag, Germany, pp , [18] M. Ye, Ch. L, G. Chen, J. Wu, EECS: An Energy Effcent Clusterng Scheme n Wreless Sensor Networks, In Proc. of the IEEE Intl. Workshop on Strateges for Energy Effcency n Ad Hoc and Sensor Networks (IWSEEASN'05), Aprl [19] V. Mttal, M. Demrbas, A. Arora, LOCI: Local Clusterng Servce for Large Scale Wreless Sensor Networks, Techncal Report OSU-CISRC-2/03-TR07, Oho State Unversty, [20] S. Soro, W. B. Henzelman, Prolongng the Lfetme of Wreless Sensor Networks va Unequal Clusterng, 5th Intl. IEEE Workshop on Algorthms for Wreless, Moble, Ad Hoc, and Sensor Networks (WMAN05), Aprl [21] Y. Guo, J. MeNar, Cost Effcent Cluster Formaton for Wreless Sensor Networks, In Proc. of the Intl. Conf. on Cybernetcs and Informaton Technologes, Systems, and Applcatons (CITSA), Orlando, Florda, July [22] M. Lotfnezhad, B. Lang, Energy Effcent Clusterng n Sensor Networks wth Moble Agents, In Proc. of the IEEE Wreless Communcatons and Networkng Conference (WCNC), New Orleans, Lousana, March

15 [23] K. Akkaya, M. Youns, "A survey on routng protocols for wreless sensor networks", Elsever Ad Hoc Network Journal, pp , [24] M. Ilyas, I. Mahgoub, "Handbook of Sensor Networks: Compact Wreless and Wred Sensng Systems", CRC Press, London, Washngton, D.C., [25] S. Kulkarn, A. Iyer, C. Rosenberg, "An Address-Lght, Integrated MAC and Routng Protocol for Wreless Sensor Networks", to appear n IEEE/ACM Transactons on Networkng, accepted August [26] O. Youns and S. Fahmy, "Dstrbuted Clusterng n Ad-hoc Sensor Networks: A Hybrd, Energy- Effcent Approach", In Proc. of IEEE INFOCOM, volume 1, pp , March [27] H. Huang, J. Wu, A Probablstc Clusterng Algorthm n Wreless Sensor Networks, In Proc. of IEEE 62nd Semannual Vehcular Technology Conference (VTC), Sept

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

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

A Binarization Algorithm specialized on Document Images and Photos

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

More information

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

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

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

More information

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

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

More information

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data

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

More information

Learning the Kernel Parameters in Kernel Minimum Distance Classifier

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

More information

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

Cluster Analysis of Electrical Behavior

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

More information

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

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching A Fast Vsual Trackng Algorthm Based on Crcle Pxels Matchng Zhqang Hou hou_zhq@sohu.com Chongzhao Han czhan@mal.xjtu.edu.cn Ln Zheng Abstract: A fast vsual trackng algorthm based on crcle pxels matchng

More information

Classifier Selection Based on Data Complexity Measures *

Classifier Selection Based on Data Complexity Measures * Classfer Selecton Based on Data Complexty Measures * Edth Hernández-Reyes, J.A. Carrasco-Ochoa, and J.Fco. Martínez-Trndad Natonal Insttute for Astrophyscs, Optcs and Electroncs, Lus Enrque Erro No.1 Sta.

More information

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

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

More information

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

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

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

High-Boost Mesh Filtering for 3-D Shape Enhancement

High-Boost Mesh Filtering for 3-D Shape Enhancement Hgh-Boost Mesh Flterng for 3-D Shape Enhancement Hrokazu Yagou Λ Alexander Belyaev y Damng We z Λ y z ; ; Shape Modelng Laboratory, Unversty of Azu, Azu-Wakamatsu 965-8580 Japan y Computer Graphcs Group,

More information

An Algorithm for Weighted Positive Influence Dominating Set Based on Learning Automata

An Algorithm for Weighted Positive Influence Dominating Set Based on Learning Automata 4 th Internatonal Conference on Knowledge-Based Engneerng and Innovaton (KBEI-2017) Dec. 22 th, 2017 (Iran Unversty of Scence and Technology) Tehran, Iran An Algorthm for Weghted Postve Influence Domnatng

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

Edge Detection in Noisy Images Using the Support Vector Machines

Edge Detection in Noisy Images Using the Support Vector Machines Edge Detecton n Nosy Images Usng the Support Vector Machnes Hlaro Gómez-Moreno, Saturnno Maldonado-Bascón, Francsco López-Ferreras Sgnal Theory and Communcatons Department. Unversty of Alcalá Crta. Madrd-Barcelona

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

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

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

Applying Continuous Action Reinforcement Learning Automata(CARLA) to Global Training of Hidden Markov Models

Applying Continuous Action Reinforcement Learning Automata(CARLA) to Global Training of Hidden Markov Models Applyng Contnuous Acton Renforcement Learnng Automata(CARLA to Global Tranng of Hdden Markov Models Jahanshah Kabudan, Mohammad Reza Meybod, and Mohammad Mehd Homayounpour Department of Computer Engneerng

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

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

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

More information

Outline. Self-Organizing Maps (SOM) US Hebbian Learning, Cntd. The learning rule is Hebbian like:

Outline. Self-Organizing Maps (SOM) US Hebbian Learning, Cntd. The learning rule is Hebbian like: Self-Organzng Maps (SOM) Turgay İBRİKÇİ, PhD. Outlne Introducton Structures of SOM SOM Archtecture Neghborhoods SOM Algorthm Examples Summary 1 2 Unsupervsed Hebban Learnng US Hebban Learnng, Cntd 3 A

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

GSLM Operations Research II Fall 13/14

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

More information

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

S1 Note. Basis functions.

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

More information

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

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

More information

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

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

More information

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

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

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

More information

Adaptive Energy and Location Aware Routing in Wireless Sensor Network

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

More information

Support Vector Machines

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

More information

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

Optimal Workload-based Weighted Wavelet Synopses

Optimal Workload-based Weighted Wavelet Synopses Optmal Workload-based Weghted Wavelet Synopses Yoss Matas School of Computer Scence Tel Avv Unversty Tel Avv 69978, Israel matas@tau.ac.l Danel Urel School of Computer Scence Tel Avv Unversty Tel Avv 69978,

More information

The Research of Support Vector Machine in Agricultural Data Classification

The Research of Support Vector Machine in Agricultural Data Classification The Research of Support Vector Machne n Agrcultural Data Classfcaton Le Sh, Qguo Duan, Xnmng Ma, Me Weng College of Informaton and Management Scence, HeNan Agrcultural Unversty, Zhengzhou 45000 Chna Zhengzhou

More information

IMPROVEMENT of MULTIPLE ROUTING BASED on FUZZY CLUSTERING and PSO ALGORITHM IN WSNS TO REDUCE ENERGY CONSUMPTION

IMPROVEMENT of MULTIPLE ROUTING BASED on FUZZY CLUSTERING and PSO ALGORITHM IN WSNS TO REDUCE ENERGY CONSUMPTION IMPROVEMENT of MULTIPLE ROUTING BASED on FUZZY CLUSTERING and PSO ALGORITHM IN WSNS TO REDUCE ENERGY CONSUMPTION Gholamreza Farahan Department of Electrcal Engneerng and Informaton Technology, Iranan Research

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

A Deflected Grid-based Algorithm for Clustering Analysis

A Deflected Grid-based Algorithm for Clustering Analysis A Deflected Grd-based Algorthm for Clusterng Analyss NANCY P. LIN, CHUNG-I CHANG, HAO-EN CHUEH, HUNG-JEN CHEN, WEI-HUA HAO Department of Computer Scence and Informaton Engneerng Tamkang Unversty 5 Yng-chuan

More information

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

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

More information

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

User Authentication Based On Behavioral Mouse Dynamics Biometrics

User Authentication Based On Behavioral Mouse Dynamics Biometrics User Authentcaton Based On Behavoral Mouse Dynamcs Bometrcs Chee-Hyung Yoon Danel Donghyun Km Department of Computer Scence Department of Computer Scence Stanford Unversty Stanford Unversty Stanford, CA

More information

TN348: Openlab Module - Colocalization

TN348: Openlab Module - Colocalization TN348: Openlab Module - Colocalzaton Topc The Colocalzaton module provdes the faclty to vsualze and quantfy colocalzaton between pars of mages. The Colocalzaton wndow contans a prevew of the two mages

More information

Parallel matrix-vector multiplication

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

More information

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

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

More information

Lecture 5: Multilayer Perceptrons

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

More information

CMPS 10 Introduction to Computer Science Lecture Notes

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

More information

Unsupervised Learning and Clustering

Unsupervised Learning and Clustering Unsupervsed Learnng and Clusterng Why consder unlabeled samples?. Collectng and labelng large set of samples s costly Gettng recorded speech s free, labelng s tme consumng 2. Classfer could be desgned

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

AMath 483/583 Lecture 21 May 13, Notes: Notes: Jacobi iteration. Notes: Jacobi with OpenMP coarse grain

AMath 483/583 Lecture 21 May 13, Notes: Notes: Jacobi iteration. Notes: Jacobi with OpenMP coarse grain AMath 483/583 Lecture 21 May 13, 2011 Today: OpenMP and MPI versons of Jacob teraton Gauss-Sedel and SOR teratve methods Next week: More MPI Debuggng and totalvew GPU computng Read: Class notes and references

More information

F Geometric Mean Graphs

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

More information

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

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

More information

A KIND OF ROUTING MODEL IN PEER-TO-PEER NETWORK BASED ON SUCCESSFUL ACCESSING RATE

A KIND OF ROUTING MODEL IN PEER-TO-PEER NETWORK BASED ON SUCCESSFUL ACCESSING RATE A KIND OF ROUTING MODEL IN PEER-TO-PEER NETWORK BASED ON SUCCESSFUL ACCESSING RATE 1 TAO LIU, 2 JI-JUN XU 1 College of Informaton Scence and Technology, Zhengzhou Normal Unversty, Chna 2 School of Mathematcs

More information

Solving two-person zero-sum game by Matlab

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

More information

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

TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS. Muradaliyev A.Z.

TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS. Muradaliyev A.Z. TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS Muradalyev AZ Azerbajan Scentfc-Research and Desgn-Prospectng Insttute of Energetc AZ1012, Ave HZardab-94 E-mal:aydn_murad@yahoocom Importance of

More information

Optimization Methods: Integer Programming Integer Linear Programming 1. Module 7 Lecture Notes 1. Integer Linear Programming

Optimization Methods: Integer Programming Integer Linear Programming 1. Module 7 Lecture Notes 1. Integer Linear Programming Optzaton Methods: Integer Prograng Integer Lnear Prograng Module Lecture Notes Integer Lnear Prograng Introducton In all the prevous lectures n lnear prograng dscussed so far, the desgn varables consdered

More information

A Decentralized Lifetime Maximization Algorithm for Distributed Applications in Wireless Sensor Networks

A Decentralized Lifetime Maximization Algorithm for Distributed Applications in Wireless Sensor Networks A Decentralzed Lfetme Maxmzaton Algorthm for Dstrbuted Applcatons n Wreless Sensor Networks Vrgna Pllon, Mauro Franceschell, Lug Atzor, Alessandro Gua Dept. of Electrcal and Electronc Engneerng, Unversty

More information

Classification Method in Integrated Information Network Using Vector Image Comparison

Classification Method in Integrated Information Network Using Vector Image Comparison Sensors & Transducers 2014 by IFSA Publshng, S. L. http://www.sensorsportal.com Classfcaton Method n Integrated Informaton Network Usng Vector Image Comparson Zhou Yuan Guangdong Polytechnc Normal Unversty

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

Machine Learning. Topic 6: Clustering

Machine Learning. Topic 6: Clustering Machne Learnng Topc 6: lusterng lusterng Groupng data nto (hopefully useful) sets. Thngs on the left Thngs on the rght Applcatons of lusterng Hypothess Generaton lusters mght suggest natural groups. Hypothess

More information

Mathematics 256 a course in differential equations for engineering students

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

More information

Wavefront Reconstructor

Wavefront Reconstructor A Dstrbuted Smplex B-Splne Based Wavefront Reconstructor Coen de Vsser and Mchel Verhaegen 14-12-201212 2012 Delft Unversty of Technology Contents Introducton Wavefront reconstructon usng Smplex B-Splnes

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

Positive Semi-definite Programming Localization in Wireless Sensor Networks

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

More information

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

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

More information

Related-Mode Attacks on CTR Encryption Mode

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

More information

Hierarchical clustering for gene expression data analysis

Hierarchical clustering for gene expression data analysis Herarchcal clusterng for gene expresson data analyss Gorgo Valentn e-mal: valentn@ds.unm.t Clusterng of Mcroarray Data. Clusterng of gene expresson profles (rows) => dscovery of co-regulated and functonally

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

Support Vector Machines

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

More information

NOVEL CONSTRUCTION OF SHORT LENGTH LDPC CODES FOR SIMPLE DECODING

NOVEL CONSTRUCTION OF SHORT LENGTH LDPC CODES FOR SIMPLE DECODING Journal of Theoretcal and Appled Informaton Technology 27 JATIT. All rghts reserved. www.jatt.org NOVEL CONSTRUCTION OF SHORT LENGTH LDPC CODES FOR SIMPLE DECODING Fatma A. Newagy, Yasmne A. Fahmy, and

More information

Smoothing Spline ANOVA for variable screening

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

More information

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

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

Study of Data Stream Clustering Based on Bio-inspired Model

Study of Data Stream Clustering Based on Bio-inspired Model , pp.412-418 http://dx.do.org/10.14257/astl.2014.53.86 Study of Data Stream lusterng Based on Bo-nspred Model Yngme L, Mn L, Jngbo Shao, Gaoyang Wang ollege of omputer Scence and Informaton Engneerng,

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

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

A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION

A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION 1 THE PUBLISHING HOUSE PROCEEDINGS OF THE ROMANIAN ACADEMY, Seres A, OF THE ROMANIAN ACADEMY Volume 4, Number 2/2003, pp.000-000 A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION Tudor BARBU Insttute

More information

Active Contours/Snakes

Active Contours/Snakes Actve Contours/Snakes Erkut Erdem Acknowledgement: The sldes are adapted from the sldes prepared by K. Grauman of Unversty of Texas at Austn Fttng: Edges vs. boundares Edges useful sgnal to ndcate occludng

More information

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

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

More information

Mobile Adaptive Distributed Clustering Algorithm for Wireless Sensor Networks

Mobile Adaptive Distributed Clustering Algorithm for Wireless Sensor Networks Internatonal Journal of Computer Applcatons (975 8887) Volume No.7, Aprl Moble Adaptve Dstrbuted Clusterng Algorthm for Wreless Sensor Networks S.V.Mansekaran Department of Informaton Technology Anna Unversty

More information

XV International PhD Workshop OWD 2013, October Machine Learning for the Efficient Control of a Multi-Wheeled Mobile Robot

XV International PhD Workshop OWD 2013, October Machine Learning for the Efficient Control of a Multi-Wheeled Mobile Robot XV Internatonal PhD Workshop OWD 203, 9 22 October 203 Machne Learnng for the Effcent Control of a Mult-Wheeled Moble Robot Uladzmr Dzomn, Brest State Techncal Unversty (prof. Vladmr Golovko, Brest State

More information

Term Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task

Term Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task Proceedngs of NTCIR-6 Workshop Meetng, May 15-18, 2007, Tokyo, Japan Term Weghtng Classfcaton System Usng the Ch-square Statstc for the Classfcaton Subtask at NTCIR-6 Patent Retreval Task Kotaro Hashmoto

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

COMPARISON OF TWO MODELS FOR HUMAN EVACUATING SIMULATION IN LARGE BUILDING SPACES. University, Beijing , China

COMPARISON OF TWO MODELS FOR HUMAN EVACUATING SIMULATION IN LARGE BUILDING SPACES. University, Beijing , China COMPARISON OF TWO MODELS FOR HUMAN EVACUATING SIMULATION IN LARGE BUILDING SPACES Bn Zhao 1, 2, He Xao 1, Yue Wang 1, Yuebao Wang 1 1 Department of Buldng Scence and Technology, Tsnghua Unversty, Bejng

More information

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

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

More information

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

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

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

More information

Structure Formation of Social Network

Structure Formation of Social Network Structure Formaton of Socal Network DU Nan 1, FENG Hu 2, HUANG Zgang 3, Sally MAKI 4, WANG Ru(Ruby) 5, and ZHAO Hongxa (Melssa) 6 1 Bejng Unversty of Posts and Telecommuncatons, Chna 2 Fudan Unversty,

More information

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

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

More information

VRT012 User s guide V0.1. Address: Žirmūnų g. 27, Vilnius LT-09105, Phone: (370-5) , Fax: (370-5) ,

VRT012 User s guide V0.1. Address: Žirmūnų g. 27, Vilnius LT-09105, Phone: (370-5) , Fax: (370-5) , VRT012 User s gude V0.1 Thank you for purchasng our product. We hope ths user-frendly devce wll be helpful n realsng your deas and brngng comfort to your lfe. Please take few mnutes to read ths manual

More information

Brave New World Pseudocode Reference

Brave New World Pseudocode Reference Brave New World Pseudocode Reference Pseudocode s a way to descrbe how to accomplsh tasks usng basc steps lke those a computer mght perform. In ths week s lab, you'll see how a form of pseudocode can be

More information

Clustering in Wireless Multimedia Sensor Networks Using Spectral Graph Partitioning

Clustering in Wireless Multimedia Sensor Networks Using Spectral Graph Partitioning Int. J. Communcatons, Network and System Scences, 2013, 6, 128-133 http://dx.do.org/10.4236/jcns.2013.63015 Publshed Onlne March 2013 (http://www.scrp.org/journal/jcns) Clusterng n Wreless Multmeda Sensor

More information

Querying by sketch geographical databases. Yu Han 1, a *

Querying by sketch geographical databases. Yu Han 1, a * 4th Internatonal Conference on Sensors, Measurement and Intellgent Materals (ICSMIM 2015) Queryng by sketch geographcal databases Yu Han 1, a * 1 Department of Basc Courses, Shenyang Insttute of Artllery,

More information

An Improved Image Segmentation Algorithm Based on the Otsu Method

An Improved Image Segmentation Algorithm Based on the Otsu Method 3th ACIS Internatonal Conference on Software Engneerng, Artfcal Intellgence, Networkng arallel/dstrbuted Computng An Improved Image Segmentaton Algorthm Based on the Otsu Method Mengxng Huang, enjao Yu,

More information