Topological Relations in the World of Minimum Bounding Rectangles: A Study with R-trees

Size: px
Start display at page:

Download "Topological Relations in the World of Minimum Bounding Rectangles: A Study with R-trees"

Transcription

1 Toological Relations in the World of Minimm Bonding Rectangles: A Stdy with R-trees Dimitris Paadias Deartment of Comter Science and Engineering University of California, San Diego CA 0-0, USA dimitris@cs.csd.ed Timos Sellis Deartment of Electrical and Comter Engineering National Technical University of Athens GREECE timos@theseas.nta.gr Yannis Theodoridis Deartment of Electrical and Comter Engineering National Technical University of Athens GREECE theodor@theseas.nta.gr Max J. Egenhofer National Center for Geograhic Information and Analysis University of Maine, Orono ME 0-, USA max@mecan.maine.ed Abstract Recent develoments in satial relations have led to their se in nmeros alications involving satial databases. This aer is concerned with the retrieval of toological relations in Minimm Bonding Rectangle-based data strctres. We stdy the toological information that Minimm Bonding Rectangles convey abot the actal objects they enclose, sing the concet of rojections. Then we aly the reslts to R-trees and their variations, R + -trees and R*-trees in order to minimise disk accesses for eries involving toological relations. We also investigate eries that involve comlex satial conditions in the form of disjnctions and conjnctions and we discss ossible extensions.. INTRODUCTION The reresentation and rocessing of satial relations has gained mch attention in Satial Qery Langages (Paadias and Sellis, ; Egenhofer, ), Image and Mltimedia Databases (Sistla et al., ), Geograhic Alications (Frank, ), Satial Reasoning (Randell et al., ) and Cognitive Science (Glasgow and Paadias, ). Desite the attention that satial relations have attracted in other alication domains, they have not been extensively alied in satial data strctres. So far most of the work on satial access methods has concentrated on the relations disjoint and not_disjoint and on the retrieval of distance information (Rossoolos et al., ). This is not becase other satial relations are nimortant for ractical alications, bt mostly becase of the lack, ntil recently, of definitions for satial relations. In this aer we focs on the retrieval of toological relations, relations that stay invariant nder toological transformations sch as translation, rotation and scaling. In articlar we will deal with toological relations between contigos region objects (the term refers to homogeneosly -dimensional, connected objects with connected bondaries) as defined by the -intersection model (Egenhofer, ). According to this model, each object is reresented in R sace as a oint set which has an interior, a bondary and an exterior. The toological relation between any two objects and is described by the nine intersections of 's interior, bondary and exterior, with the interior, bondary and exterior of (based on the concets of oint-set toology). Ot of different relations that can be distingished by the model, only the following eight are meaningfl for region objects: disjoint, meet, eal, overla, contains (and the converse relation inside) and covers (and the converse covered_by)(see Figre ). disjoint(,) meet(,) overla(,) covered_by(,) covers(,) inside(,) contains(,) Fig. Toological relations eal(,) We call the revios set of toological relations m t in order to distingish them from the set m t ={disjoint, not_disjoint}. The meaning of disjoint is the same in both sets (it imlies that two objects have no common oints), while all the other relations of m t are refinements of not_disjoint. Sometimes in satial data strctres terminology, the term overla (instead of not_disjoint) is sed to denote any configration in which the objects are

2 not disjoint. The relations of m t are airwise disjoint, and they rovide a comlete coverage. Randell et al. () reached the same set of airwise disjoint toological relations following an axiomization based on mereology and exressed in a many-sorted logic. Tests with hman sbjects have shown evidence that the -intersection model has otential for defining cognitively meaningfl satial redicates, a fact that makes the above relations a good candidate for commercial systems (Mark and Egenhofer, ). In fact, the -intersection model has been imlemented with Geograhical Information Systems (GIS); Hadzilacos and Tryfona (), for instance, sed it to exress geograhical constraints, and Mark and Xia () to determine satial relations in ARC/INFO. In addition there are imlementations in commercial systems sch as Intergrah (MGE, ) and Oracle MD (Keighan, ). Frthermore, the relations of m t have been inflential in assessing the consistency of toological information in satial databases (Egenhofer and Sharma, ), ery otimisation strategies (Clementini et al., ) and Satial Reasoning (Grigni et al., ). This aer is concerned with the retrieval of the toological relations of m t sing satial data strctres based on Minimm Bonding Rectangles (MBRs). In articlar we concentrate on R-trees and their variations. Section illstrates the ossible relations between MBRs and describes the corresonding toological relations. Section investigates the toological information that MBRs convey abot the actal objects they enclose with resect to the -intersection model. Section alies the reslts in R-tree-based data strctres and comares the retrieval times. Section is concerned with eries that involve comlex satial conditions. Section discsses extensions that deal with imrecision and Section concldes with comments on ftre work.. MINIMUM BOUNDING RECTANGLES It is a common strategy in satial access methods to store object aroximations and se these aroximations to index the data sace in order to efficiently retrieve the otential objects that satisfy the reslt of a ery. Deending on the alication domain there are several otions in choosing object aroximations. Brinkhoff et al. () comare the se of rotated minimm bonding rectangles, convex hlls, minimm bonding n-corner convexes etc. in satial access methods. In this aer we examine methods based on the traditional aroximation of Minimm Bonding Rectangles. MBRs have been sed extensively to aroximate objects in Satial Data Strctres and Satial Reasoning becase they need only two oints for their reresentation; in articlar, each object is reresented as an ordered air (' l,' ) of oints that corresond to the lower left and the er right oint of the MBR ' that covers (' l stands for the lower and ' for the er oint of the MBR). While MBRs demonstrate some disadvantages when aroximating non-convex or diagonal objects, they are the most commonly sed aroximations in satial alications. The term rimary object/mbr denotes the object/mbr to be located and the term reference object/mbr denotes the object in relation to which the rimary object/mbr is located (in the examles hereafter, the reference objects are grey, while the rimary objects are transarent). Let X and Y be fnctions that give the x and y coordinate of a oint resectively. When we se two oints for the reresentation of the reference object and for the case of one-dimensional sace, the axis is divided into artitions. Three artitions are oen line segments (the interior and the exteriors of line segment ' l ' ) and two artitions are the oints ' l and ' (see Figre ). R l R R l R R l R l R R l R l R0 R l R l R l ' l ' x Fig. Possible relations in D sace If the rimary object is also reresented by two oints ' l and ' ordered on the x axis (X(' l )< X(' )) then the nmber of airwise disjoint relations between the two objects in D sace is. These relations corresond to the relations between time intervals introdced in (Allen, ). The symbols ' l and ' in Figre denote the edge oints (lower and er) for the reference MBR and the characters l and the lower and the er oints of the rimary MBR. In order to stdy the corresondence between MBRs and actal objects with resect to the relations of m t we will aly the revios reslts in D sace sing rojections on the x and y axis. In this case, the constraint for the lower and the er oints of the bonding rectangle is: X(' l ) < X(' ) Y(' l ) < Y(' ) and the nmber of airwise disjoint relations is (the sare of the nmber of relations in D sace). These relations are illstrated in Figre ; they corresond to the highest accracy sing two oints er object, in the sense that they cannot be defined as disjnctions of more "restrictive" relations.

3 R i_ R i_ R i_ R i_ R i_ R i_ R i_ R i_ R i_ R i_0 R i_ R i_ R i_ R _j R _j R _j R _j R _j R _j R _j R _j R _j R 0_j R _j R _j R _j Figre illstrates the corresonding toological relation for the configrations of Figre. For instance, all the configrations of the first row (R _j where j can take any vale from to ) corresond to the disjoint relation, and the total nmber of configrations in which the MBRs are disjoint is. On the other hand, only relation R _ corresonds to the toological relation contains. Eal Contains Inside Covers CoveredBy Disjoint Meet 0 Overla 0 TOTAL 0 0 Fig. Toological relations between MBRs A nmber of satial data strctres based on MBRs have been develoed. The most romising gro incldes R-trees (Gttman, ) and their variations. The R-tree is a height-balanced tree, which consists of intermediate and leaf nodes (R-trees are direct extensions of B-trees in k- Fig. Possible relations between MBRs dimensions). The MBRs of the actal data objects are assmed to be stored in the leaf nodes of the tree. Intermediate nodes are bilt by groing rectangles at the lower level. An intermediate node is associated with some rectangle which encloses all rectangles that corresond to lower level nodes. Each air of nodes may satisfy any of the toological relations of m t. The fact that R-trees ermit overla among node entries sometimes leads to nsccessfl hits on the tree strctre. The R + -tree (Sellis et al., ) and the R * -tree (Beckmann et al., 0) methods have been roosed to address the roblem of erformance degradation cased by the overlaing regions or excessive dead-sace resectively. To avoid this roblem, the R + -tree achieves zero overla among intermediate node entries by allowing artitions to slit nodes. The trade-off is that more sace is reired becase of the dlicate entries and ths the height of the tree may be greater than the original R-tree. On the other hand, the R * -tree ermits overla among nodes, bt tries to minimise it by organising rectangles into nodes sing a more comlex algorithm than the one of the original R-tree.

4 If the only toological relations of interest are the relations of m t, then when the MBRs of two objects are disjoint we can conclde that the objects that they reresent are also disjoint. If the MBRs however share common oints, no conclsion can be drawn abot the toological relation between the objects. For this reason, satial eries involve the following two ste strategy: First a filter ste based on MBRs is sed to raidly eliminate MBRs of objects that cold not ossibly satisfy the ery and select a set of otential candidates. Then dring a refinement ste each candidate is examined (by sing comtational geometry technies) and false hits are detected and eliminated. Brinkhoff et al. () extended the above strategy to inclde a second filter ste with finer aroximations than MBR (e.g. convex hlls) in order to exclde some false hits from the set of candidates. The above technies seed- the retrieval of relations of m t sing R-trees and variations. The resent work bilds on the original two-ste strategy to exlore a larger, and more detailed set of toological relations. Sch an investigation is imortant becase satial eries freently reire the kind of alitative resoltion distingished by m t. We first stdy how MBR aroximations can be sed for the retrieval of toological relations of m t between actal objects and then we aly the reslts in actal imlementations.. TOPOLOGICAL RELATIONS THAT MBRS CONVEY ABOUT THE ACTUAL OBJECTS Since the MBRs are only aroximations of the actal objects, the toological relation between MBRs does not necessarily coincide with the toological relation between the objects. In most cases the MBRs of objects that satisfy a given relation, shold satisfy a nmber of ossible relations with resect to the MBR of the reference object. We will start with relations that involve the retrieval of a small nmber of MBRs and we will gradally move to relations for which a large nmber of MBRs shold be retrieved. In order to answer the ery "find all objects eal to object " we need to retrieve all MBRs that are eal to ', that is all MBRs that satisfy the relation R - with resect to '. Only these MBRs may enclose objects that satisfy the ery. On the other hand, the retrieved MBRs may also enclose objects that satisfy the relations overla, covered_by, covers or meet with resect to (see Figre ). As in the case of m t, a refinement ste is needed when dealing with the relations of m t if the MBRs of the retrieved objects are not disjoint. In a revios aer we have shown how the strategy can be alied for the retrieval of direction relations between extended objects (Paadias et al., ). Fig. Possible relations for objects when the MBRs are eal The relations contains and inside also involve the retrieval of MBRs that satisfy nie configrations. In articlar, the objects that contain can only be in MBRs that contain ' (MBRs that satisfy the relation R - with resect to '), while the objects inside, can only be in MBRs that are inside ' (MBRs that satisfy the relation R _ with resect to '). As in the case of eal, a refinement ste is needed becase: contains(',') disjoint(,) meet(,) overla(,) contains(,) covers(,) and inside(',') disjoint(,) meet(,) overla(,) inside(,) covered_by(,) The rest of the relations involve the retrieval of more than one MBR configrations. For instance, in order to answer the ery "find all objects that cover a given object" we need to retrieve the MBRs that satisfy the relations covers, contains, or eal with resect to the MBR of the reference object. As Figre illstrates, these MBRs satisfy the rojection relations R i-j where i and j in {,,,}. Similarly the retrieval of covered_by involves all MBRs that satisfy the relations R i-j where i,j in {,,,0} with resect to '. R _j R _j R _j R _j R i_ R i_ R i_ R i_ Fig. Configrations of MBRs that yield the relation covers between actal objects The remaining relations involve the retrieval of a large nmber of MBRs. In the case of disjoint, for instance, all the MBRs may enclose objects that are disjoint with the reference object, excet for those that satisfy the relation R i-j where i, in {,,,} and j in {,,,0}, or i

5 R i_ R i_ R i_ R i_ R i_ R i_ R i_ R i_ R i_ R i_0 R i_ R i_ R i_ R _j R _j R _j R _j R _j R _j R _j R _j R _j R 0_j R _j R _j R _j Fig. Configrations of MBRs that yield the relation disjoint between actal objects in{,,,0} and j in {,,,}. Figre illstrates all configrations of MBRs that enclose objects otentially disjoint with the reference object. The blank configrations in Figre are not to be retrieved when we deal with contigos objects. If, for instance, the MBRs are related by the relation R _, then the MBRs overla and the actal objects also necessarily overla. Figre illstrates sch a configration; in this case a refinement ste is not needed. We followed the same rocedre for the relations meet and overla. Table shows the conclsions that can be drawn if we se the concet of rojections to stdy toological relations. For each relation, the Table illstrates the sbset of the MBR configrations that cold contain objects that satisfy the relation. Fig. Overlaing MBRs for which the actal objects necessarily overla

6 eal(,) contains(,) inside(,) covers(,) covered_by(,) disjoint(,) meet(,) overla(,) Table Toological relations imlemented For some cases the refinement ste can be avoided, that is, the relation between MBRs nambigosly determines the toological relation between the actal objects (e.g., the configration of Figre ). Figre illstrates the configrations for which the refinement ste is not needed (this haens only in some of the cases for disjoint and overla) disjoint overla Fig. Configrations for which a refinement ste is not needed Clementini et al. () stdied the se of MBR aroximations in ery rocessing involving toological relations. Frthermore, they defined a minimal sbset of the nine intersections that can otimally determine the relation between the actal objects taking into accont the freency of the relations. Their findings can be sed dring the refinement ste in order to minimise the cost of the comtation of intersections between otential candidates. In the next section we aly the reslts of sections and to actal imlementations based on R- trees.. IMPLEMENTATION OF TOPOLOGICAL RELATIONS IN R-TREES In order to retrieve the toological relations of m t sing R-trees one needs to define more general relations that will be sed for roagation in the intermediate nodes of the tree strctre. For instance, the intermediate nodes P that cold enclose MBRs ' that meet the MBR ' of the reference object, shold satisfy the more general constraint meet(p,') verla(p,') covers(p,') contains(p,'). Figre 0 illstrates examles of sch configrations. P P ' meet(p,') ' ' covers(p,') ' P ' ' overla(p,') P ' ' contains(p,') Fig. 0 Intermediate nodes that may contain MBRs that meet the MBR of the reference object Following this strategy, the search sace is rned by exclding the intermediate nodes P that do not satisfy the revios constraint. Table resents the relations that shold be satisfied between an intermediate node P and the MBR ' of the reference object, so that the node will be selected for roagation. Notice that the same relation between intermediate nodes and the reference MBR exists for all the levels of the tree strctre. For instance, the intermediate nodes that cold enclose other intermediate nodes P that satisfy the general constraint meet(p,') verla(p,') covers(p,') contains (P,') shold also satisfy the same constraint. This conclsion can be easily extracted from the above table and is alicable to all the toological relations of Table.

7 Relation between MBRs eal(',') contains(',') inside(',') covers(',') Relation between intermediate node P (that may enclose MBRs ') and reference MBR eal(p,') covers(p,') contains(p,') contains(p,') overla(p,') covered_by(p,') inside(p,') eal(p,') covers(p,') contains(p,') covers(p,') contains(p,') covered_by(',') overla(p,') covered_by(p,') eal(p,') covers(p,') contains(p,') disjoint(',') meet(',') disjoint(p,') meet(p,') verla(p,') covers(p,') contains(p,') meet(p,') verla(p,') covers(p,') contains(p,') overla(',') overla(p,') covers(p,') contains(p,') Table Relations for the intermediate nodes Smmarising, the rocessing of a ery of the form "find all objects that satisfy a given toological relation with resect to object " in R-tree-based data strctres involves the following stes:. Comte the MBRs ' that cold enclose objects that satisfy the ery. This maing involves Table.. Find the toological relations that the MBRs ' of the first ste may satisfy with resect to '. This rocedre involves Figre.. Starting from the to node, exclde the intermediate nodes P which cold not enclose MBRs that satisfy the toological relations of the second ste and recrsively search the remaining nodes. This rocedre involves Table.. Follow a refinement ste for the retrieved MBRs, excet for the cases illstrated in Figre. In order to exerimentally antify the erformance of the above algorithm, we created tree strctres by inserting 0,000 MBRs randomly generated. We tested three data files: - the first file contains small MBRs: the size of each rectangle is at most 0,0% of the global area - the second file contains medim MBRs: the size of each rectangle is at most 0,% of the global area - the third file contains large MBRs: the size of each rectangle is at most 0,% of the global area. The search rocedre sed a search file for each data file containing 00 rectangles, also randomly generated, with similar size roerties as the data rectangles. We sed the revios data files for retrieval of toological relations in R-, R + - and R*-trees. In the imlementation of R-trees we selected the adratic-slit algorithm and we set the minimm node caacity to m=0%; in the imlementation of R*-trees we set m=0% while in the imlementation of R + -trees the "minimm nmber of rectangle slits" was selected to be the cost fnction. These settings seem to be the most efficient ones for each method (Beckmann et al., 0; Sellis et al., ). Table illstrates the nmber of hits er search, that is, the nmber of retrieved MBRs for the three files. The nmber of hits er search is inversely roortional to the selectivity of the relation and it is related to the nmber of disk access. Note that the total nmber of MBRs in each colmn is larger than 0,000 becase a MBR may be retrieved for more than one relation. Toological Relation small MBRs medim MBRs large MBRs disjoint,,, meet.0.. overla. 0.. covered_by.0.0. inside eal covers.0.. contains Table Retrieved MBRs er relation for each data file The nmber of disk accesses er search is illstrated grahically in Figre. With the excetion of disjoint, the imrovement in the efficiency of retrieval sing R- trees comared to serial retrieval is immense; this is articlarly tre for small size MBRs where the difference is almost two orders of magnitde. The difference dros as the MBRs become larger. The increase in the size increases the density and, therefore, the ossibility that the reference object is not disjoint with other MBRs or intermediate nodes. The retrieval of disjoint is, as exected, worse than serial retrieval becase this relation reires the retrieval of all the nodes of the tree strctre as derived from the -ste strategy resented above. Clearly, a "real" system wold do a serial retrieval in sch a case instead of sing the tree strctre. Obviosly, toological relations can be retrieved sing the traditional window ery (i.e., not_disjoint relation) sed in satial data strctres. In that case, the distinction between the not_disjoint relation and the actal toological relation to be retrieved can be made dring the refinement ste. The cost of this aroach is roghly eivalent to the cost of meet (according to Table, meet involves the retrieval of almost all MBRs that are not disjoint with the reference MBR). Deending on the relation and the MBR size, or aroach can imrove the erformance by to 0%. Another imortant imrovement is the redction of the MBRs that are selected for the refinement ste. As illstrated in Table, the nmber of hits er search for some relations (e.g., inside, covers) is considerably lower than the hits er The nmber of disk accesses er search sing serial retrieval is eal to 00 for all relations (the size of the data file divided by the age caacity which is eal to 0 entries).

8 disk accesses er search... A R-tree R+-tree R*-tree m o cb i e cv cn toological relation disk accesses er search.... A A A A A A toological relation R-tree R+-tree R*-tree m o cb i e cv cn A A A A A A A A disk accesses er search toological relation R-tree R+-tree R*-tree A A m o cb i e cv cn Small data size Medim data size Large data size Fig. Performance comarison of R-tree variants search for meet (sally below 0%). Therefore the int of the refinement ste is a small sbset of the MBRs that wold be selected sing the window ery. According to Figre, the relations of m t can be groed into three categories with resect to the cost of retrieval. The first gro contains disjoint which is the most exensive relation and shold be rocessed by serial search. The second gro consists of meet, overla, inside and covered_by. The third gro consists of the three relations (eal, covers, contains) which are the least exensive to rocess. The cost difference between the second and the third gro increases as the size of the MBRs becomes larger. Note also that the rimary factor for the retrieval cost is the relation for the intermediate nodes (Table ). The relation inside is more exensive than covers, althogh it retrieves only one ott MBR configration (relation R - with resect to '). This is de to the large nmber of intermediate nodes that cold contain MBRs that satisfy the relation R -. On the other hand, the cost of inside is almost the same as the cost of covered_by, which can be inferred by Table. The comarison of the varios R-tree-based strctres follows, in general, the conclsions drawn in the literatre i.e., the variations R + -trees and R*-trees oterform the original R-trees. For small and medim MBRs, R + -trees erform slightly better than R*-trees. However this advantage is lost if the dlicate entries generate one extra level in the tree strctre, as haened for the large data size of or tests. In sch case the erformance of R + -trees is inferior for the most exensive relations bt remains cometent for the least exensive ones. The irreglar behavior of R + -trees is de to the lack of overla between nodes, which reslts in the ick exclsion of the majority of the intermediate nodes when one of the least exensive relations is retrieved. On the other hand, for the When the data density becomes high it is ossible that all of the entries in a fll node coincide on the same oint of the lane. In sch cases R + trees do not work (Greene, ). This haened for some data sets involving large MBRs dring or tests. exensive relations there is no significant gain and the erformance is worse comared to the other strctres becase of the greater nmber of nodes in the tree. In section we have stdied eries that involve the retrieval of objects that satisfy a toological relation with resect to a reference object. In the next section we extend or reslts for eries that involve comlex satial conditions in the form of disjnctions and conjnctions with resect to one and two reference objects.. COMPLEX QUERIES In some cases the refinement rovided by the relations of m t is not needed. In a cadastral alication, for instance, the difference between inside and covered_by may not be imortant. Consider the ery "find all land arcels in a given area". The land arcels of the reslt shold be inside or covered_by the area, that is, the interretation of in is inside covered_by. We can define toological relations of lower alitative resoltion sing disjnctions of the relations of m t. This is a common strategy in alitative satial knowledge reresentation; it has been sed for direction relations in (Paadias and Sellis, ) and for the above toological relations in (Randell et al., ). Qeries involving low resoltion relations can also be rocessed by the -ste method of section. The only difference is that the set of the MBRs to be retrieved by the first ste is the nion of the MBRs to be retrieved by each of the relations that belong to the disjnction. Frthermore, in some cases the retrieval times do not change. In the revios ery (relation in), the MBRs of the reslt are the same as those that wold be retrieved if the relation of interest were covered_by. This is becase the MBRs to be retrieved for inside constitte a sbset of the MBRs for covered_by (see Table ). Figre illstrates the sbset relations with resect to the ott MBRs.

9 disjoint contain covers overla eal covered_by inside meet Fig. Sbset relations according to the ott MBRs According to Figre the retrieval of the ery "find all objects that satisfy the relation meet or contain or eal or inside " involves the same retrieval time with the ery "find all objects that meet ". On the other hand, eries that involve conjnctions of toological relations with resect to one reference object (e.g., find all objects that are inside and covered_by ) have an emty reslt becase the relations of m t are airwise disjoint. Nevertheless we can erform semantic ery otimisation for eries of the form "find all objects that are inside and overla with ". We can determine that the reslt of this ery is emty if we know that and are disjoint. As Figre illstrates, if is inside and disjoint, it cannot be the case that overlas. Fig. A ery with two reference objects and emty reslt Table illstrates the relations between reference objects for which an emty reslt is retrned withot rnning the ery. When a ery involving two reference objects and is given, the toological relation between the reference objects is examined, and if it is one of the relations of the table, the ott is emty. For the above ery, in addition to disjoint, if and are related by meet, eal, inside or covered_by, the reslt is also emty. Each entry at row r i (, ) and colmn r j (, ) (where r i and r j are relations of m t ) is the comlement of the comosition relation r' i (,) and r j (, ) with resect to m t, (where r' i is the converse relation of r i ). For an extensive discssion abot comosition of toological relations see (Egenhofer, ). If the relation between the reference objects is not one of the relations of the array, then one of the two relations is retrieved sing R- trees and the alifying MBRs ' are filtered with resect to the other reference object. This rocess can be erformed in main memory by checking the relation that each ott MBR of the first ste satisfies with resect to the second reference object. Therefore the nmber of disk accesses deends only on the first relation. The selection of the first relation is based on the size of the reference MBR and on the cost gro that the relation belongs to. The relations covers, contains and eal are referable becase they reire the least disk accesses. If the sizes of the reference MBRs are considerably different, then the smallest reference MBR mst be selected becase the cost of retrieval is roortional to the data size (see Figre ). disjoint(, ) meet(, ) eal(, ) inside(, ) cvrdby(, ) contain(, ) covers(, ) overla(, ) disjoint(, ) meet(, ) eal(, ) inside(, ) cvrdby(, ) contain(, ) covers(, ) overla(, ) --- e ct cv m e i e i cb i ct d e i m e i e i cb e i cb m e i m e i d e i i cb d e i cb m e i e i cb ct cv d m i i cb cv i cb ct i ct cv o e i cb e i cb i cb ct cv e ct cv e ct cv m e i ct cv d m ct cv ct cv d e ct cv i ct cv i cb d m i ct ct cv ct cv m e i i cb cv i cb cv i cb cv Table Conjnctions of relations that yield emty reslts m e i e i cb ct cv o i cb ct i cb cv i cb e ct cv e ct cv i cb ct cv i cb i cb --- e ct cv e ct cv e i cb i ct e ct cv e i cb e i cb ---

10 Althogh the above methods rovide a good fondation for ractical alications, the nderlying theoretical assmtions do not always hold in ractical (GIS) environments. In the next section we discss how or work can be extended to catre real-world imrecision.. NON-CRISP MBRS The se of MBR filters for rocessing toological relations has a very delicate asect, as it relies heavily on the fact that the MBRs are cris reresentations of the objects. A cris MBR has to satisfy two constraints:. the object to be aroximated is flly contained within the rectangle and. each bondary of the MBR coincides with some art of the object s bondary. One may arge that a violation of either condition wold not constitte an MBR in its literal sense (the rectangle wold not be bonding, or the rectangle wold not be minimal); however, even with carefl imlementations of algorithms that create MBRs for satial objects, these constraints can be reglarly violated. This roblem is de to the difficlties of imlementing coordinate-based comter algorithms for geometry that reserve toology. One cannot always reliably derive from a coordinate reresentation toological roerties sch as the coincidence of two oints, or whether a oint close to a line is actally on the line etc. While the first constraint in the definition of cris MBRs can sally be flfilled, e.g., by making the MBRs slightly larger than reired, it is the second constraint (the minimality of rectangles) that introdces the roblem. Sometimes recision is traded for erformance and fast, bt slightly inaccrate, algorithms may be sed to calclate MBRs. Sch algorithms make sre that constraint is flfilled, bt occasionally violate constraint so that some MBRs may actally be larger than necessary. The same may haen de to nmerical inaccracies sch as ronding errors, articlarly, if floating oint arithmetic is sed to reresent the objects coordinates and MBRs are exressed by two integer airs. Therefore, often in imlementations of MBR-based satial access one has to assme some nmerical inaccracies. The -ste algorithm of section can be extended to deal with inaccracies involving slightly larger than cris MBRs. Or extension is based on the concet of concetal neighborhood (Freksa, Egenhofer and Al-Taha ). The concetal neighborhood of the one-dimensional relations of Figre forms a grah, in which each air of directly connected nodes corresonds to a air of relations that are concetal neighbors. Given an initial relation between two objects, if we continosly enlarge the rimary object, we follow the ath indicated by the arrows in Figre a. For instance, if the relation between the objects is R, then extending the rimary object, while keeing the reference object constant, gradally leads to relations R, R, R and R (see Figre ). Likewise, enlarging the reference object changes the relation between the two objects according to the directions of the edges in Figre b. A first-degree concetal neighbor of a relation R i is a relation R j that can be reached from R i via a directed edge in either neighborhood grah. For examle, relation has for first-degree concetal neighbors (relations and if we enlarge the rimary object, and relations and 0 if we enlarge the reference object). On the other hand, relation has one first-degree concetal neighbor, relation which is obtained by enlarging either object. A second-degree concetal neighbor of R i is a relation R j that has at least two first-degree concetal neighbors that are also first-degree neighbors of R i. For examle, the second-degree concetal neighbors of relation comrises the set of relations,,,, and ; they all have at least two common first-degree neighbors with. On the other hand, relation does not have any second-degree neighbors. These reslts can be easily extended to d sace where two relations are k-neighbors if they are k-neighbors in any dimension. R R R R R R R R R R R R R 0 (a) Enlargement of rimary object R R R R R R R 0 R R R R R R (b) Enlargement of reference object Fig. Concetal neighborhoods for relations in d sace We se the notion of concetal neighborhood to overcome the otential roblems that may arise from inaccracy. In addition to the configrations of Table, the candidate MBRs for the first ste may also satisfy the first- and second-degree concetal neighbor relations with resect to the cris MBRs that wold be retrieved. Table shold be sed instead of Table in sch cases. The dark grey rectangles of Table corresond to the cris MBR relations dislayed in Table. The light grey

11 eal contains inside covers rectangles are the additional ones that corresond to first and second neighborhood relations. The extra MBRs to be retrieved increase the retrieval cost, bt assre that all otential objects that satisfy the ery are retrieved even if there exists inaccracy that can reslt in -degree relation deformation. The largest increase in ott MBRs is observed for the relation eal, for which MBR relations have to be considered (in lie of MBR relation in the cris case). On the other hand, the ott MBRs for the relation overla remain constant.. CONCLUSION covered_by disjoint meet overla Table Retrieval sing -neighbors The reresentation and rocessing of toological relations is an imortant toic in a nmber of areas inclding Satial Qery Langages, Image and Mltimedia Databases and Geograhic Information Systems. Desite their imortance, toological relations have not been extensively sed in satial data strctres. In this aer we have focsed on the retrieval of toological relations in MBR-based data strctres. In articlar we have shown how the toological relations disjoint, meet, eal, overla, contains, inside, covers and covered_by (defined by the -intersection model) can be retrieved from R-trees and their variations. First we illstrated the ossible relations between MBRs and we described the corresonding toological relations. Then we stdied the toological information that MBRs convey abot the actal objects they enclose sing the concet of rojections. Finally we alied the reslts in R-tree-based data strctres and we conclded that they are sitable for toological relations, with R + - and R * - trees oterforming the original R-trees for most cases. We also investigated eries that involve comlex satial conditions in the form of disjnctions and conjnctions and we demonstrated how or method can be alied to non-cris MBRs. Althogh we have dealt with contigos regions, ractical alications do not necessarily deal with contigos objects. Geograhic entities, sch as contries with islands, consist of disconnected comonents. The revios reslts have been extended for this case. The only difference is that the nmber of MBRs to be retrieved for some relations increases since the relaxation of the contigity constraint alifies more MBRs as otential candidates. An imlementation of satial relations for non-contigos objects, as well as geograhic examles and detailed descritions can be fond (Paadias and Theodoridis, ) In order to model linear and oint data we need frther extensions becase the toological relations that can be defined, as well as the nmber of ossible rojection relations between MBRs, deend on the tye of objects. Egenhofer (), for instance, defined relations between lines based on the -intersection model, while Paadias and Sellis () have shown that the nmber of different rojections between a region reference object and a line rimary object is. The ideas in this aer can be extended to inclde linear and oint data, objects with holes etc. ACKNOWLEDGEMENTS We wold like to thank Cheryl Jacob for roof-reading this aer and roviding sefl sggestions. Dimitris Paadias was artially sorted by NSF - IRI. Part of this work was done while he was with the Deartment of Geoinformation, Technical University of Vienna. Yannis Theodoridis and Timos Sellis were artially sorted by the Deartment of Research and Technology of Greece (PENED ).

12 Max Egenhofer was artially sorted by the National Science Fondation nder grant nmbers IRI-00 and SBR-0 (for the National Center for Geograhic Information and Analysis), Intergrah Cororation, Environmental Systems Research Institte, the Scientific Division of the North Atlantic Treaty Organisation, and the Maine Mathematics and Science Alliance. REFERENCES Allen, J.F. () Maintaining Knowledge abot Temoral Intervals. Commnications of ACM, Vol (),. -. Beckmann, N., Kriegel, H.P. Schneider, R., Seeger, B. (0) The R*-tree: an Efficient and Robst Access Method for Points and Rectangles. In the Proceedings of ACM- SIGMOD Conference. Brinkhoff, T., Kriegel, H.P, Schneider, R. () Comarison of Aroximations of Comlex Objects sed for Aroximation-based Qery Processing in Satial Database Systems. In the Proceedings of th International Conference on Data Engineering. Brinkhoff, T., Kriegel, H.P., Schneider, R., Seeger, B. () Mlti-Ste Processing of Satial Joins. In the Proceedings of ACM-SIGMOD Conference. Clementini, E., Sharma, J., Egenhofer, M. () Modeling Toological Satial Relations: Strategies for Qery Processing. To aear in the International Jornal of Comter and Grahics. Egenhofer, M. () Reasoning abot Binary Toological Relations. In the Proceedings of the Second Symosim on the Design and Imlementation of Large Satial Databases. Sringer Verlag LNCS. Egenhofer, M. () Definition of Line-Line Relations for Geograhic Databases. Data Engineering, Vol (),. 0-. Egenhofer, M. () Satial SQL: A Qery and Presentation Langage. IEEE Transactions on Data and Knowledge Engineering, (), -. Egenhofer, M., Al-Taha, K. () Reasoning abot Gradal Changes of Toological Relationshis. In the Proceedings of the International Conference GIS - From Sace to Territory: Theories and Methods of Satio-Temoral Reasoning in Geograhic Sace. Sringer Verlag LNCS. Frank, A. U. () Qalitative Satial Reasoning: Cardinal Directions as an Examle. To aear in the International Jornal of Geograhic Information Systems. Freksa, C. () Temoral Reasoning Based on Semi- Intervals. Artificial Intelligence Vol,. -. Glasgow, J.I., Paadias, D. () Comtational Imagery. Cognitive Science, Vol,. -. Greene, D. () An Imlementation and Performance Analysis of Satial Data Access Methods. In the Proceedings of the th International Conference on Data Engineering. Grigni M., Paadias, D., Paadimitrio, C. () Toological Inference. Sbmitted. Gttman, A. () R-trees: a Dynamic Index Strctre for Satial Searching. In the Proceedings of ACM-SIGMOD Conference. Hadzilacos, T., Tryfona, N. () A Model for Exressing Toological Integrity Constraints in Geograhic Databases. In the Proceedings of the International Conference GIS - From Sace to Territory: Theories and Methods of Satio-Temoral Reasoning in Geograhic Sace. Sringer Verlag LNCS. Keighan, E. () Managing Satial Data within the Framework of the Relational Model. Technical Reort, Oracle Cororation, Canada. Mark, D., Egenhofer, M. () Calibrating the Meaning of Satial Predicates from Natral Langage: Line Region Relations. In the Proceedings of the th International Symosim on Satial Data Handling. Taylor Francis. Mark, D., Xia, F. () Determining Satial Relations between Lines and Regions in Arc/Info sing the - Intersection Model. In ESRI User Conference. MGE () MGE Analyst Reference Manal. Intergrah Cororation. Paadias, D., Sellis, T. () The Qalitative Reresentation of Satial Knowledge in two-dimensional Sace. VLDB Jornal, Secial Isse on Satial Databases, Vol,. -. Paadias, D., Theodoridis, Y. () Satial Relations, Minimm Bonding Rectangles and Satial Data Strctres. Technical Reort, KDBSLAB-TR--0, National Technical University of Athens, Greece. Paadias, D., Theodoridis, Y., Sellis, T. () The Retrieval of Direction Relations Using R trees. In the Proceedings of the th Conference on Database and Exert Systems Alications. Sringer Verlag LNCS. Paadias, D., Sellis, T. () A Pictorial Qery-By-Examle Langage. To aear in the Jornal of Visal Langages and Comting, Secial Isse on Visal Qery Systems, March. Randell, D. A., Ci, Z., Cohn., A., () A Satial Logic Based on Regions and Connection. In the Proceedings of the rd International Conference on Princiles of Knowledge Reresentation and Reasoning. Morgan Kafmann. Rossoolos, N., Kelley, F., Vincent, F. () Nearest Neighbor Qeries. In the Proceedings of ACM-SIGMOD Conference. Sellis, T., Rossoolos, N., Falotsos, C. () The R + -tree: A Dynamic Index for Mlti-Dimensional Objects. In the Proceedings of the th VLDB conference. Sistla, P., Y, C., Haddad, R. () Reasoning abot Satial Relationshis in Pictre Retrieval Systems. In the Proceedings of the 0th VLDB Conference.

L L.w-l wl l+!!j. Topological Relations in the World of Minimum Bounding Rectangles: A Study with R-trees. Yannis Theodoridis

L L.w-l wl l+!!j. Topological Relations in the World of Minimum Bounding Rectangles: A Study with R-trees. Yannis Theodoridis Topological Relations in the World of Minimum Bounding Rectangles: A Study with R-trees Dimitris Papadias Department of Computer Science and Engineering University of California, San Diego CA 92093-04,

More information

A Simulation-based Spatial Decision Support System for a New Airborne Weather Data Acquisition System

A Simulation-based Spatial Decision Support System for a New Airborne Weather Data Acquisition System A Simlation-based Satial Decision Sort System for a New Airborne Weather Data Acqisition System Erol Ozan Deartment of Engineering Management Old Dominion University Norfol, VA 23529 Pal Kaffmann Deartment

More information

Chapter 2 Introduction to the Stiffness (Displacement) Method. The Stiffness (Displacement) Method

Chapter 2 Introduction to the Stiffness (Displacement) Method. The Stiffness (Displacement) Method CIVL 7/87 Chater - The Stiffness Method / Chater Introdction to the Stiffness (Dislacement) Method Learning Objectives To define the stiffness matrix To derive the stiffness matrix for a sring element

More information

A KINEMATIC WAVE MODEL FOR EMERGENCY EVACUATION PLANNING

A KINEMATIC WAVE MODEL FOR EMERGENCY EVACUATION PLANNING A KINEMATIC WAVE MODEL FOR EMERGENCY EVACUATION PLANNING KAI-FU QIU, LIANG CHEN, ZHE WANG WEN-LONG JIN Deartment of Atomation Center for Intelligent Transortation Systems University of Science and Technology

More information

Approximating min-max k-clustering

Approximating min-max k-clustering Aroximating min-max k-clustering Asaf Levin July 24, 2007 Abstract We consider the roblems of set artitioning into k clusters with minimum total cost and minimum of the maximum cost of a cluster. The cost

More information

Constrained tri-connected planar straight line graphs

Constrained tri-connected planar straight line graphs Constrained tri-connected lanar straight line grahs Marwan Al-Jbeh Gill Bareet Mashhood Ishae Diane L. Soaine Csaba D. Tóth Andrew Winslow Abstract It is known that for any set V of n 4 oints in the lane,

More information

The analysis of scoring matrix in voting system with multiple criteria and multiple places

The analysis of scoring matrix in voting system with multiple criteria and multiple places 0 rd International Conference on Advanced Management Science IPEDR vol.9 (0) (0) IACSIT Press, Singaore The analysis of scoring matrix in voting system with mltile criteria and mltile laces Fh-Hwa Franlin

More information

MODELING THE RELIABILITY OF C4ISR SYSTEMS HARDWARE/SOFTWARE COMPONENTS USING AN IMPROVED MARKOV MODEL

MODELING THE RELIABILITY OF C4ISR SYSTEMS HARDWARE/SOFTWARE COMPONENTS USING AN IMPROVED MARKOV MODEL Technical Sciences and Alied Mathematics MODELING THE RELIABILITY OF CISR SYSTEMS HARDWARE/SOFTWARE COMPONENTS USING AN IMPROVED MARKOV MODEL Cezar VASILESCU Regional Deartment of Defense Resources Management

More information

3.1 Experimental Design

3.1 Experimental Design 3 Relay Feedback Åström and Hägglnd [1] sggest the relay feedback test to generate sstained oscillation as an alternative to the conventional continos cycling techniqe. It is very effective in determining

More information

Gravity Waves in Shear and Implications for Organized Convection

Gravity Waves in Shear and Implications for Organized Convection SEPTEMBER 2009 S T E C H M A N N A N D M A J D A 2579 Gravity Waves in Shear and Imlications for Organized Convection SAMUEL N. STECHMANN Deartment of Mathematics, and Deartment of Atmosheric and Oceanic

More information

Rough Sets in Approximate Spatial Reasoning

Rough Sets in Approximate Spatial Reasoning ough Sets in Aroximate Satial easoning Thomas Bittner and John G. Stell Centre de recherche en geomatique, Laval University, Quebec, Canada. Thomas.Bittner@scg.ulaval.ca Deartment of Comuter Science, Keele

More information

Forced repetitions over alphabet lists

Forced repetitions over alphabet lists Forced reetitions over alhabet lists Neerja Mhaskar 1 and Michael Soltys 2 1 McMaster University Det. of Comting & Software 1280 Main Street West Hamilton, Ontario L8S 4K1, CANADA ohlin@mcmaster.ca 2 California

More information

An Introduction To Range Searching

An Introduction To Range Searching An Introduction To Range Searching Jan Vahrenhold eartment of Comuter Science Westfälische Wilhelms-Universität Münster, Germany. Overview 1. Introduction: Problem Statement, Lower Bounds 2. Range Searching

More information

A priori error analysis for transient problems using Enhanced Velocity approach in the discrete-time setting.

A priori error analysis for transient problems using Enhanced Velocity approach in the discrete-time setting. arxiv:181.04809v1 [math.na] 1 Dec 018 A riori error analysis for transient roblems sing nhanced Velocity aroach in the discrete-time setting. Yerlan Amanbek 1, and Mary F. Wheeler 1 1 Center for Sbsrface

More information

A manufacturing network design model based on processor and worker capabilities

A manufacturing network design model based on processor and worker capabilities A manfactring network design model based on rocessor and worker caabilities Marc Paqet, Alain Martel and Benoit Montreil Aril 2005 Working Paer DT-2005-AM-1 (Revised March 2006) Network Organization Technology

More information

Program Burn Algorithms Based on Detonation Shock Dynamics

Program Burn Algorithms Based on Detonation Shock Dynamics Program Brn Algorithms Based on Detonation Shock Dynamics Discrete Aroimations of Detonation Flows with Discontinos Front Models J. B. Bdzil, D. S. Stewart and T. L. Jackson DX-division, Los Alamos National

More information

Designing Parametric Cubic Curves. Prof. George Wolberg Dept. of Computer Science City College of New York

Designing Parametric Cubic Curves. Prof. George Wolberg Dept. of Computer Science City College of New York Designing Parametric Cbic Crves Prof. George Wolberg Det. of Comter Science City College of New York Objectives Introdce the tyes of crves - Interolating -Hermite - Bezier - B-Sline Analyze their erformance

More information

Existence of HCMU metrics on higher genus Riemann surfaces

Existence of HCMU metrics on higher genus Riemann surfaces Existence of HCMU metrics on higher gens Riemann srfaces October 4, 0 bstract We will generalize the reslt in [CCW05] and roe the existence of HCMU metrics on higher gens K-srfaces, i.e. Riemann srfaces

More information

GIVEN an input sequence x 0,..., x n 1 and the

GIVEN an input sequence x 0,..., x n 1 and the 1 Running Max/Min Filters using 1 + o(1) Comarisons er Samle Hao Yuan, Member, IEEE, and Mikhail J. Atallah, Fellow, IEEE Abstract A running max (or min) filter asks for the maximum or (minimum) elements

More information

Open Access Stability Analysis of Internet Congestion Control Model with Compound TCP Under RED

Open Access Stability Analysis of Internet Congestion Control Model with Compound TCP Under RED Send Orders for Rerints to rerints@benthamscienceae 986 The Oen Atomation and Control Systems Jornal 205 7 986992 Oen Access Stability Analysis of Internet Congestion Control Model with Comond TCP Under

More information

FRONT TRACKING FOR A MODEL OF IMMISCIBLE GAS FLOW WITH LARGE DATA

FRONT TRACKING FOR A MODEL OF IMMISCIBLE GAS FLOW WITH LARGE DATA FONT TACKING FO A MODEL OF IMMISCIBLE GAS FLOW WITH LAGE DATA HELGE HOLDEN, NILS HENIK ISEBO, AND HILDE SANDE Abstract. In this aer we stdy front tracking for a model of one dimensional, immiscible flow

More information

An Analysis of Reliable Classifiers through ROC Isometrics

An Analysis of Reliable Classifiers through ROC Isometrics An Analysis of Reliable Classifiers through ROC Isometrics Stijn Vanderlooy s.vanderlooy@cs.unimaas.nl Ida G. Srinkhuizen-Kuyer kuyer@cs.unimaas.nl Evgueni N. Smirnov smirnov@cs.unimaas.nl MICC-IKAT, Universiteit

More information

Module 4. Analysis of Statically Indeterminate Structures by the Direct Stiffness Method. Version 2 CE IIT, Kharagpur

Module 4. Analysis of Statically Indeterminate Structures by the Direct Stiffness Method. Version 2 CE IIT, Kharagpur Modle Analysis of Statically Indeterminate Strctres by the Direct Stiffness Method Version CE IIT, Kharagr Lesson The Direct Stiffness Method: Trss Analysis (Contined) Version CE IIT, Kharagr Instrctional

More information

Centrum voor Wiskunde en Informatica

Centrum voor Wiskunde en Informatica Centrm voor Wisknde en Informatica Probability, Networks and Algorithms Probability, Networks and Algorithms Some remarks on roerties of generalized Pickands constants K.G. Debicki, REPOR PNA-R4 FEBRUARY,

More information

For q 0; 1; : : : ; `? 1, we have m 0; 1; : : : ; q? 1. The set fh j(x) : j 0; 1; ; : : : ; `? 1g forms a basis for the tness functions dened on the i

For q 0; 1; : : : ; `? 1, we have m 0; 1; : : : ; q? 1. The set fh j(x) : j 0; 1; ; : : : ; `? 1g forms a basis for the tness functions dened on the i Comuting with Haar Functions Sami Khuri Deartment of Mathematics and Comuter Science San Jose State University One Washington Square San Jose, CA 9519-0103, USA khuri@juiter.sjsu.edu Fax: (40)94-500 Keywords:

More information

Finite-State Verification or Model Checking. Finite State Verification (FSV) or Model Checking

Finite-State Verification or Model Checking. Finite State Verification (FSV) or Model Checking Finite-State Verification or Model Checking Finite State Verification (FSV) or Model Checking Holds the romise of roviding a cost effective way of verifying imortant roerties about a system Not all faults

More information

LONG GAPS BETWEEN DEFICIENT NUMBERS

LONG GAPS BETWEEN DEFICIENT NUMBERS LONG GAPS BETWEEN DEFICIENT NUMBERS PAUL POLLACK Abstract. Let n be a natral nmber. If the sm of the roer divisors of n is less than n, then n is said to be deficient. Let G(x) be the largest ga between

More information

arxiv: v2 [math.na] 7 Sep 2016

arxiv: v2 [math.na] 7 Sep 2016 A MODIFIED IMPLEMENTATION OF MINRES TO MONITOR RESIDUAL SUBVECTOR NORMS FOR BLOCK SYSTEMS ROLAND HERZOG AND KIRK M. SOODHALTER Abstract. Saddle-oint systems, i.e., strctred linear systems with symmetric

More information

Krauskopf, B., Lee, CM., & Osinga, HM. (2008). Codimension-one tangency bifurcations of global Poincaré maps of four-dimensional vector fields.

Krauskopf, B., Lee, CM., & Osinga, HM. (2008). Codimension-one tangency bifurcations of global Poincaré maps of four-dimensional vector fields. Kraskopf, B, Lee,, & Osinga, H (28) odimension-one tangency bifrcations of global Poincaré maps of for-dimensional vector fields Early version, also known as pre-print Link to pblication record in Explore

More information

Pressure coefficient evaluation on the surface of the SONDA III model tested in the TTP Pilot Transonic Wind Tunnel

Pressure coefficient evaluation on the surface of the SONDA III model tested in the TTP Pilot Transonic Wind Tunnel Jornal of Physics: Conference Series OPEN ACCESS Pressre coefficient evalation on the srface of the SONDA III model tested in the TTP Pilot Transonic Wind Tnnel To cite this article: M L C C Reis et al

More information

CMSC 425: Lecture 4 Geometry and Geometric Programming

CMSC 425: Lecture 4 Geometry and Geometric Programming CMSC 425: Lecture 4 Geometry and Geometric Programming Geometry for Game Programming and Grahics: For the next few lectures, we will discuss some of the basic elements of geometry. There are many areas

More information

APPLICATION OF ENERGY ABSORBING CONNECTERS TO STEEL CONTINUOUS GIRDER BRIDGES

APPLICATION OF ENERGY ABSORBING CONNECTERS TO STEEL CONTINUOUS GIRDER BRIDGES 13 th World Conference on Earthqake Engineering Vancover, B.C., Canada Agst 1-6, 24 Paer No. 3367 APPLICATION OF ENERGY ABSORBING CONNECTERS TO STEEL CONTINUOUS GIRDER BRIDGES Hiroshi ZUI 1 and Yoshio

More information

Network Farthest-Point Diagrams and their Application to Feed-Link Network Extension

Network Farthest-Point Diagrams and their Application to Feed-Link Network Extension Network Farthest-Point Diagrams and their Alication to Feed-Link Network Extension Prosenjit Bose Kai Dannies Jean-Lo De Carfel Christoh Doell Carsten Grimm Anil Maheshwari Stefan Schirra Michiel Smid

More information

BIFURCATIONS OF TRAVELING WAVE SOLUTIONS FOR A GENERALIZED CAMASSA-HOLM EQUATION

BIFURCATIONS OF TRAVELING WAVE SOLUTIONS FOR A GENERALIZED CAMASSA-HOLM EQUATION Jornal of Alied Analsis and Comtation Volme 8, Nmber 6, December 8, 85 86 Website:htt://jaac.ijornal.cn DOI:.98/8.85 BIFURCATIONS OF TRAVELING WAVE SOLUTIONS FOR A GENERALIZED CAMASSA-HOLM EQUATION Minzhi

More information

4.2 First-Order Logic

4.2 First-Order Logic 64 First-Order Logic and Type Theory The problem can be seen in the two qestionable rles In the existential introdction, the term a has not yet been introdced into the derivation and its se can therefore

More information

Partial stabilization of a class of hydraulic mechanical systems with Casimir functions

Partial stabilization of a class of hydraulic mechanical systems with Casimir functions Proceedings of the 7th World Congress he International Federation of Atomatic Control Seol, Korea, Jly 6-, 28 Partial stabilization of a class of hydralic mechanical systems with Casimir fnctions Sator

More information

Study on the impulsive pressure of tank oscillating by force towards multiple degrees of freedom

Study on the impulsive pressure of tank oscillating by force towards multiple degrees of freedom EPJ Web of Conferences 80, 0034 (08) EFM 07 Stdy on the implsive pressre of tank oscillating by force towards mltiple degrees of freedom Shigeyki Hibi,* The ational Defense Academy, Department of Mechanical

More information

Spatial Coupling of a Lattice Boltzmann fluid model with a Finite Difference Navier-Stokes solver

Spatial Coupling of a Lattice Boltzmann fluid model with a Finite Difference Navier-Stokes solver Satial Coling of a Lattice Boltzmann flid model with a Finite Difference Navier-Stoes solver Jonas Latt Bastien Choard Pal Albqerqe arxiv:hysics/0511243v1 [hysics.com-h] 29 Nov 2005 Agst 25, 2018 Abstract

More information

A Unified 2D Representation of Fuzzy Reasoning, CBR, and Experience Based Reasoning

A Unified 2D Representation of Fuzzy Reasoning, CBR, and Experience Based Reasoning University of Wollongong Research Online Faculty of Commerce - aers (Archive) Faculty of Business 26 A Unified 2D Reresentation of Fuzzy Reasoning, CBR, and Exerience Based Reasoning Zhaohao Sun University

More information

COMPUTING SCIENCE. University of Newcastle upon Tyne. Behaviour-Preserving Transition Insertions in Unfolding Prefixes. V.

COMPUTING SCIENCE. University of Newcastle upon Tyne. Behaviour-Preserving Transition Insertions in Unfolding Prefixes. V. UNIVERSITY OF NEWCASTLE University of Newcastle on Tyne COMPUTING SCIENCE Behavior-Preserving Transition Insertions in Unfolding Prefixes V. Khomenko TECHNICAL REPORT SERIES No. CS-TR-952 March, 2006 TECHNICAL

More information

Control Systems of a Non-stationary Plant Based on MPC and PID Type Fuzzy Logic Controller

Control Systems of a Non-stationary Plant Based on MPC and PID Type Fuzzy Logic Controller Proceedings o the International MltiConerence o Engineers and Comter Scientists 2016 Vol I,, March 16-18, 2016, Hong Kong Control Systems o a Non-stationary Plant Based on MPC and PID Tye Fzzy Logic Controller

More information

Computations of the Shock Waves at Hypersonic Velocities Taken into Account the Chemical Reactions that Appear in the Air at High Temperatures

Computations of the Shock Waves at Hypersonic Velocities Taken into Account the Chemical Reactions that Appear in the Air at High Temperatures Comtations of the Shock Waves at Hyersonic Velocities aken into Accont the Chemical Reactions that Aear in the Air at High emeratres Mihai Leonida NICLESC*,, Maris Gabriel COJOCAR Mihai Victor PRICOP,

More information

Principles of Computed Tomography (CT)

Principles of Computed Tomography (CT) Page 298 Princiles of Comuted Tomograhy (CT) The theoretical foundation of CT dates back to Johann Radon, a mathematician from Vienna who derived a method in 1907 for rojecting a 2-D object along arallel

More information

Implementation and Validation of Finite Volume C++ Codes for Plane Stress Analysis

Implementation and Validation of Finite Volume C++ Codes for Plane Stress Analysis CST0 191 October, 011, Krabi Imlementation and Validation of Finite Volume C++ Codes for Plane Stress Analysis Chakrit Suvanjumrat and Ekachai Chaichanasiri* Deartment of Mechanical Engineering, Faculty

More information

1. Tractable and Intractable Computational Problems So far in the course we have seen many problems that have polynomial-time solutions; that is, on

1. Tractable and Intractable Computational Problems So far in the course we have seen many problems that have polynomial-time solutions; that is, on . Tractable and Intractable Comptational Problems So far in the corse we have seen many problems that have polynomial-time soltions; that is, on a problem instance of size n, the rnning time T (n) = O(n

More information

Numerical Methods for Particle Tracing in Vector Fields

Numerical Methods for Particle Tracing in Vector Fields On-Line Visualization Notes Numerical Methods for Particle Tracing in Vector Fields Kenneth I. Joy Visualization and Grahics Research Laboratory Deartment of Comuter Science University of California, Davis

More information

An Ant Colony Optimization Approach to the Probabilistic Traveling Salesman Problem

An Ant Colony Optimization Approach to the Probabilistic Traveling Salesman Problem An Ant Colony Otimization Aroach to the Probabilistic Traveling Salesman Problem Leonora Bianchi 1, Luca Maria Gambardella 1, and Marco Dorigo 2 1 IDSIA, Strada Cantonale Galleria 2, CH-6928 Manno, Switzerland

More information

Combining Logistic Regression with Kriging for Mapping the Risk of Occurrence of Unexploded Ordnance (UXO)

Combining Logistic Regression with Kriging for Mapping the Risk of Occurrence of Unexploded Ordnance (UXO) Combining Logistic Regression with Kriging for Maing the Risk of Occurrence of Unexloded Ordnance (UXO) H. Saito (), P. Goovaerts (), S. A. McKenna (2) Environmental and Water Resources Engineering, Deartment

More information

RANDOM WALKS AND PERCOLATION: AN ANALYSIS OF CURRENT RESEARCH ON MODELING NATURAL PROCESSES

RANDOM WALKS AND PERCOLATION: AN ANALYSIS OF CURRENT RESEARCH ON MODELING NATURAL PROCESSES RANDOM WALKS AND PERCOLATION: AN ANALYSIS OF CURRENT RESEARCH ON MODELING NATURAL PROCESSES AARON ZWIEBACH Abstract. In this aer we will analyze research that has been recently done in the field of discrete

More information

Estimation of Lateral Displacements for Offshore Monopiles in Clays based on CPT Results

Estimation of Lateral Displacements for Offshore Monopiles in Clays based on CPT Results Estimation of Lateral Dislacements for Offshore Monoiles in Clas based on CPT Reslts *Garam Kim 1), Jinoh Kim 2), Incheol Kim 3), Dongho Kim 4), Bngkwon Bn 5), Yanghoon Roh 6), Ohchang Kwon 7) and Jnhwan

More information

Nonlinear parametric optimization using cylindrical algebraic decomposition

Nonlinear parametric optimization using cylindrical algebraic decomposition Proceedings of the 44th IEEE Conference on Decision and Control, and the Eropean Control Conference 2005 Seville, Spain, December 12-15, 2005 TC08.5 Nonlinear parametric optimization sing cylindrical algebraic

More information

Linear diophantine equations for discrete tomography

Linear diophantine equations for discrete tomography Journal of X-Ray Science and Technology 10 001 59 66 59 IOS Press Linear diohantine euations for discrete tomograhy Yangbo Ye a,gewang b and Jiehua Zhu a a Deartment of Mathematics, The University of Iowa,

More information

Brownian Motion and Random Prime Factorization

Brownian Motion and Random Prime Factorization Brownian Motion and Random Prime Factorization Kendrick Tang June 4, 202 Contents Introduction 2 2 Brownian Motion 2 2. Develoing Brownian Motion.................... 2 2.. Measure Saces and Borel Sigma-Algebras.........

More information

Evaluation of de ections of the vertical on the sphere and the plane: a comparison of FFT techniques

Evaluation of de ections of the vertical on the sphere and the plane: a comparison of FFT techniques Jornal of Geodesy (1997) 71: 461±468 Evalation of de ections of the vertical on the shere and the lane: a comarison of FFT techniqes Q. W. Li 1, Y. Cai Li, Michael G. Sideris 1 Institte of Geodesy and

More information

Lecture Notes On THEORY OF COMPUTATION MODULE - 2 UNIT - 2

Lecture Notes On THEORY OF COMPUTATION MODULE - 2 UNIT - 2 BIJU PATNAIK UNIVERSITY OF TECHNOLOGY, ODISHA Lectre Notes On THEORY OF COMPUTATION MODULE - 2 UNIT - 2 Prepared by, Dr. Sbhend Kmar Rath, BPUT, Odisha. Tring Machine- Miscellany UNIT 2 TURING MACHINE

More information

Lilian Markenzon 1, Nair Maria Maia de Abreu 2* and Luciana Lee 3

Lilian Markenzon 1, Nair Maria Maia de Abreu 2* and Luciana Lee 3 Pesquisa Oeracional (2013) 33(1): 123-132 2013 Brazilian Oerations Research Society Printed version ISSN 0101-7438 / Online version ISSN 1678-5142 www.scielo.br/oe SOME RESULTS ABOUT THE CONNECTIVITY OF

More information

This is a repository copy of Parallel-in-Space-and-Time Simulation of the Three-Dimensional, Unsteady Navier-Stokes Equations for Incompressible Flow.

This is a repository copy of Parallel-in-Space-and-Time Simulation of the Three-Dimensional, Unsteady Navier-Stokes Equations for Incompressible Flow. This is a reository coy of Parallel-in-Sace-and-Time Simlation of the Three-Dimensional, Unsteady Naier-Stokes Eqations for Incomressible Flow. White Rose Research Online URL for this aer: htt://erints.whiterose.ac.k/90737/

More information

Strong Matching of Points with Geometric Shapes

Strong Matching of Points with Geometric Shapes Strong Matching of Points with Geometric Shaes Ahmad Biniaz Anil Maheshwari Michiel Smid School of Comuter Science, Carleton University, Ottawa, Canada December 9, 05 In memory of Ferran Hurtado. Abstract

More information

Konyalioglu, Serpil. Konyalioglu, A.Cihan. Ipek, A.Sabri. Isik, Ahmet

Konyalioglu, Serpil. Konyalioglu, A.Cihan. Ipek, A.Sabri. Isik, Ahmet The Role of Visalization Approach on Stdent s Conceptal Learning Konyaliogl, Serpil Department of Secondary Science and Mathematics Edcation, K.K. Edcation Faclty, Atatürk University, 25240- Erzrm-Trkey;

More information

Estimation of the large covariance matrix with two-step monotone missing data

Estimation of the large covariance matrix with two-step monotone missing data Estimation of the large covariance matrix with two-ste monotone missing data Masashi Hyodo, Nobumichi Shutoh 2, Takashi Seo, and Tatjana Pavlenko 3 Deartment of Mathematical Information Science, Tokyo

More information

Sources of Non Stationarity in the Semivariogram

Sources of Non Stationarity in the Semivariogram Sorces of Non Stationarity in the Semivariogram Migel A. Cba and Oy Leangthong Traditional ncertainty characterization techniqes sch as Simple Kriging or Seqential Gassian Simlation rely on stationary

More information

Paper C Exact Volume Balance Versus Exact Mass Balance in Compositional Reservoir Simulation

Paper C Exact Volume Balance Versus Exact Mass Balance in Compositional Reservoir Simulation Paer C Exact Volume Balance Versus Exact Mass Balance in Comositional Reservoir Simulation Submitted to Comutational Geosciences, December 2005. Exact Volume Balance Versus Exact Mass Balance in Comositional

More information

Elements. Using ABSTRACT. 1. Introduction. Anish Bangalore, India. tioned devices. often consist packed. This offering both.

Elements. Using ABSTRACT. 1. Introduction. Anish Bangalore, India. tioned devices. often consist packed. This offering both. Jornal of Alied Mathematics and Physics,, 201, 1, 20-25 Pblished Online November 201 (htt://www.scir.org/jornal/jam) htt://dx.doi.org/10.426/jam..201.16005 A Monolithic, FEM-Based Aroach for the Coled

More information

On the Toppling of a Sand Pile

On the Toppling of a Sand Pile Discrete Mathematics and Theoretical Comuter Science Proceedings AA (DM-CCG), 2001, 275 286 On the Toling of a Sand Pile Jean-Christohe Novelli 1 and Dominique Rossin 2 1 CNRS, LIFL, Bâtiment M3, Université

More information

ABSTRACT MODEL REPAIR

ABSTRACT MODEL REPAIR Logical Methods in Comuter Science Vol. 11(3:11)2015,. 1 43 www.lmcs-online.org Submitted Jul. 2, 2014 Published Se. 17, 2015 ABSTRACT MODEL REPAIR GEORGE CHATZIELEFTHERIOU a, BORZOO BONAKDARPOUR b, PANAGIOTIS

More information

Chapter 1 Fundamentals

Chapter 1 Fundamentals Chater Fundamentals. Overview of Thermodynamics Industrial Revolution brought in large scale automation of many tedious tasks which were earlier being erformed through manual or animal labour. Inventors

More information

Effects of Soil Spatial Variability on Bearing Capacity of Shallow Foundations

Effects of Soil Spatial Variability on Bearing Capacity of Shallow Foundations Geotechnical Safety and Risk V T. Schweckendiek et al. (Eds.) 2015 The athors and IOS Press. This article is pblished online with Open Access by IOS Press and distribted nder the terms of the Creative

More information

Implementation of a Column Generation Heuristic for Vehicle Scheduling in a Medium-Sized Bus Company

Implementation of a Column Generation Heuristic for Vehicle Scheduling in a Medium-Sized Bus Company 7e Conférence Internationale de MOdélisation et SIMulation - MOSIM 08 - du 31 mars au 2 avril 2008 Paris- France «Modélisation, Otimisation MOSIM 08 et Simulation du 31 mars des Systèmes au 2 avril : 2008

More information

Temporal Social Network: Group Query Processing

Temporal Social Network: Group Query Processing Temporal Social Network: Grop Qery Processing Xiaoying Chen 1 Chong Zhang 2 Yanli H 3 Bin Ge 4 Weidong Xiao 5 Science and Technology on Information Systems Engineering Laboratory National University of

More information

The Graph Accessibility Problem and the Universality of the Collision CRCW Conflict Resolution Rule

The Graph Accessibility Problem and the Universality of the Collision CRCW Conflict Resolution Rule The Grah Accessibility Problem and the Universality of the Collision CRCW Conflict Resolution Rule STEFAN D. BRUDA Deartment of Comuter Science Bisho s University Lennoxville, Quebec J1M 1Z7 CANADA bruda@cs.ubishos.ca

More information

Actual exergy intake to perform the same task

Actual exergy intake to perform the same task CHAPER : PRINCIPLES OF ENERGY CONSERVAION INRODUCION Energy conservation rinciles are based on thermodynamics If we look into the simle and most direct statement of the first law of thermodynamics, we

More information

Convex Hull-Based Metric Refinements for Topological Spatial Relations

Convex Hull-Based Metric Refinements for Topological Spatial Relations ABSTRACT Convex Hull-Based Metric Refinements for Topological Spatial Relations Fuyu (Frank) Xu School of Computing and Information Science University of Maine Orono, ME 04469-5711, USA fuyu.xu@maine.edu

More information

Modeling and Optimization of propylene polymerization with branching

Modeling and Optimization of propylene polymerization with branching Jiří Jaromír Klemeš, Petar Sabev Varbanov and Peng Yen Liew (Editors) Proceedings of the 4 th Eroean Symosim on Comter Aided Process Engineering ESCAPE 4 Jne 5-8, 04, Bdaest, Hngary. 04 Elsevier B.V. All

More information

Computer arithmetic. Intensive Computation. Annalisa Massini 2017/2018

Computer arithmetic. Intensive Computation. Annalisa Massini 2017/2018 Comuter arithmetic Intensive Comutation Annalisa Massini 7/8 Intensive Comutation - 7/8 References Comuter Architecture - A Quantitative Aroach Hennessy Patterson Aendix J Intensive Comutation - 7/8 3

More information

STATIC, STAGNATION, AND DYNAMIC PRESSURES

STATIC, STAGNATION, AND DYNAMIC PRESSURES STATIC, STAGNATION, AND DYNAMIC PRESSURES Bernolli eqation is g constant In this eqation is called static ressre, becase it is the ressre that wold be measred by an instrment that is static with resect

More information

Solved Problems. (a) (b) (c) Figure P4.1 Simple Classification Problems First we draw a line between each set of dark and light data points.

Solved Problems. (a) (b) (c) Figure P4.1 Simple Classification Problems First we draw a line between each set of dark and light data points. Solved Problems Solved Problems P Solve the three simle classification roblems shown in Figure P by drawing a decision boundary Find weight and bias values that result in single-neuron ercetrons with the

More information

Lower Confidence Bound for Process-Yield Index S pk with Autocorrelated Process Data

Lower Confidence Bound for Process-Yield Index S pk with Autocorrelated Process Data Quality Technology & Quantitative Management Vol. 1, No.,. 51-65, 15 QTQM IAQM 15 Lower onfidence Bound for Process-Yield Index with Autocorrelated Process Data Fu-Kwun Wang * and Yeneneh Tamirat Deartment

More information

A SIMPLE PLASTICITY MODEL FOR PREDICTING TRANSVERSE COMPOSITE RESPONSE AND FAILURE

A SIMPLE PLASTICITY MODEL FOR PREDICTING TRANSVERSE COMPOSITE RESPONSE AND FAILURE THE 19 TH INTERNATIONAL CONFERENCE ON COMPOSITE MATERIALS A SIMPLE PLASTICITY MODEL FOR PREDICTING TRANSVERSE COMPOSITE RESPONSE AND FAILURE K.W. Gan*, M.R. Wisnom, S.R. Hallett, G. Allegri Advanced Comosites

More information

Correspondence Between Fractal-Wavelet. Transforms and Iterated Function Systems. With Grey Level Maps. F. Mendivil and E.R.

Correspondence Between Fractal-Wavelet. Transforms and Iterated Function Systems. With Grey Level Maps. F. Mendivil and E.R. 1 Corresondence Between Fractal-Wavelet Transforms and Iterated Function Systems With Grey Level Mas F. Mendivil and E.R. Vrscay Deartment of Alied Mathematics Faculty of Mathematics University of Waterloo

More information

Lecture Notes: Finite Element Analysis, J.E. Akin, Rice University

Lecture Notes: Finite Element Analysis, J.E. Akin, Rice University 9. TRUSS ANALYSIS... 1 9.1 PLANAR TRUSS... 1 9. SPACE TRUSS... 11 9.3 SUMMARY... 1 9.4 EXERCISES... 15 9. Trss analysis 9.1 Planar trss: The differential eqation for the eqilibrim of an elastic bar (above)

More information

OPTIMAL INTEREST RATE RULES UNDER ONE-SIDED OUTPUT AND INFLATION TARGETS. Peter J. Stemp

OPTIMAL INTEREST RATE RULES UNDER ONE-SIDED OUTPUT AND INFLATION TARGETS. Peter J. Stemp OPTIMAL INTEREST RATE RULES UNDER ONE-SIDED OUTPUT AND INFLATION TARGETS b Peter J. Stem Deartment of Economics Monash Uniersit PO Box 197, Calfield East Victoria, 3145 Astralia email: eter.stem@bseco.monash.ed.a

More information

q-ary Symmetric Channel for Large q

q-ary Symmetric Channel for Large q List-Message Passing Achieves Caacity on the q-ary Symmetric Channel for Large q Fan Zhang and Henry D Pfister Deartment of Electrical and Comuter Engineering, Texas A&M University {fanzhang,hfister}@tamuedu

More information

Shadow Computing: An Energy-Aware Fault Tolerant Computing Model

Shadow Computing: An Energy-Aware Fault Tolerant Computing Model Shadow Comuting: An Energy-Aware Fault Tolerant Comuting Model Bryan Mills, Taieb Znati, Rami Melhem Deartment of Comuter Science University of Pittsburgh (bmills, znati, melhem)@cs.itt.edu Index Terms

More information

ABSTRACT MODEL REPAIR

ABSTRACT MODEL REPAIR ABSTRACT MODEL REPAIR GEORGE CHATZIELEFTHERIOU a, BORZOO BONAKDARPOUR b, PANAGIOTIS KATSAROS c, AND SCOTT A. SMOLKA d a Deartment of Informatics, Aristotle University of Thessaloniki, 54124 Thessaloniki,

More information

CONSIDER an array of N sensors (residing in threedimensional. Investigating Hyperhelical Array Manifold Curves Using the Complex Cartan Matrix

CONSIDER an array of N sensors (residing in threedimensional. Investigating Hyperhelical Array Manifold Curves Using the Complex Cartan Matrix IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING (SPECIAL ISSUE), VOL.?, NO.?,???? 1 Investigating yperhelical Array Manifold Crves Using the Complex Cartan Matrix Athanassios Manikas, Senior Member,

More information

Online Appendix to Accompany AComparisonof Traditional and Open-Access Appointment Scheduling Policies

Online Appendix to Accompany AComparisonof Traditional and Open-Access Appointment Scheduling Policies Online Aendix to Accomany AComarisonof Traditional and Oen-Access Aointment Scheduling Policies Lawrence W. Robinson Johnson Graduate School of Management Cornell University Ithaca, NY 14853-6201 lwr2@cornell.edu

More information

GOOD MODELS FOR CUBIC SURFACES. 1. Introduction

GOOD MODELS FOR CUBIC SURFACES. 1. Introduction GOOD MODELS FOR CUBIC SURFACES ANDREAS-STEPHAN ELSENHANS Abstract. This article describes an algorithm for finding a model of a hyersurface with small coefficients. It is shown that the aroach works in

More information

Supplementary Materials for Robust Estimation of the False Discovery Rate

Supplementary Materials for Robust Estimation of the False Discovery Rate Sulementary Materials for Robust Estimation of the False Discovery Rate Stan Pounds and Cheng Cheng This sulemental contains roofs regarding theoretical roerties of the roosed method (Section S1), rovides

More information

System Reliability Estimation and Confidence Regions from Subsystem and Full System Tests

System Reliability Estimation and Confidence Regions from Subsystem and Full System Tests 009 American Control Conference Hyatt Regency Riverfront, St. Louis, MO, USA June 0-, 009 FrB4. System Reliability Estimation and Confidence Regions from Subsystem and Full System Tests James C. Sall Abstract

More information

An Adaptive Three-bus Power System Equivalent for Estimating Voltage Stability Margin from Synchronized Phasor Measurements

An Adaptive Three-bus Power System Equivalent for Estimating Voltage Stability Margin from Synchronized Phasor Measurements An Adative Three-bus Power System Equivalent for Estimating oltage Stability argin from Synchronized Phasor easurements Fengkai Hu, Kai Sun University of Tennessee Knoxville, TN, USA fengkaihu@utk.edu

More information

Designing of Virtual Experiments for the Physics Class

Designing of Virtual Experiments for the Physics Class Designing of Virtal Experiments for the Physics Class Marin Oprea, Cristina Miron Faclty of Physics, University of Bcharest, Bcharest-Magrele, Romania E-mail: opreamarin2007@yahoo.com Abstract Physics

More information

Sensitivity Analysis in Bayesian Networks: From Single to Multiple Parameters

Sensitivity Analysis in Bayesian Networks: From Single to Multiple Parameters Sensitivity Analysis in Bayesian Networks: From Single to Mltiple Parameters Hei Chan and Adnan Darwiche Compter Science Department University of California, Los Angeles Los Angeles, CA 90095 {hei,darwiche}@cs.cla.ed

More information

UNCERTAINTY FOCUSED STRENGTH ANALYSIS MODEL

UNCERTAINTY FOCUSED STRENGTH ANALYSIS MODEL 8th International DAAAM Baltic Conference "INDUSTRIAL ENGINEERING - 19-1 April 01, Tallinn, Estonia UNCERTAINTY FOCUSED STRENGTH ANALYSIS MODEL Põdra, P. & Laaneots, R. Abstract: Strength analysis is a

More information

A Bound on the Error of Cross Validation Using the Approximation and Estimation Rates, with Consequences for the Training-Test Split

A Bound on the Error of Cross Validation Using the Approximation and Estimation Rates, with Consequences for the Training-Test Split A Bound on the Error of Cross Validation Using the Aroximation and Estimation Rates, with Consequences for the Training-Test Slit Michael Kearns AT&T Bell Laboratories Murray Hill, NJ 7974 mkearns@research.att.com

More information

Network Configuration Control Via Connectivity Graph Processes

Network Configuration Control Via Connectivity Graph Processes Network Configuration Control Via Connectivity Grah Processes Abubakr Muhammad Deartment of Electrical and Systems Engineering University of Pennsylvania Philadelhia, PA 90 abubakr@seas.uenn.edu Magnus

More information

Evaluating Circuit Reliability Under Probabilistic Gate-Level Fault Models

Evaluating Circuit Reliability Under Probabilistic Gate-Level Fault Models Evaluating Circuit Reliability Under Probabilistic Gate-Level Fault Models Ketan N. Patel, Igor L. Markov and John P. Hayes University of Michigan, Ann Arbor 48109-2122 {knatel,imarkov,jhayes}@eecs.umich.edu

More information

FOUNTAIN codes [3], [4] provide an efficient solution

FOUNTAIN codes [3], [4] provide an efficient solution Inactivation Decoding of LT and Raptor Codes: Analysis and Code Design Francisco Lázaro, Stdent Member, IEEE, Gianligi Liva, Senior Member, IEEE, Gerhard Bach, Fellow, IEEE arxiv:176.5814v1 [cs.it 19 Jn

More information

Scaling Multiple Point Statistics for Non-Stationary Geostatistical Modeling

Scaling Multiple Point Statistics for Non-Stationary Geostatistical Modeling Scaling Multile Point Statistics or Non-Stationary Geostatistical Modeling Julián M. Ortiz, Steven Lyster and Clayton V. Deutsch Centre or Comutational Geostatistics Deartment o Civil & Environmental Engineering

More information

Uncorrelated Multilinear Principal Component Analysis for Unsupervised Multilinear Subspace Learning

Uncorrelated Multilinear Principal Component Analysis for Unsupervised Multilinear Subspace Learning TNN-2009-P-1186.R2 1 Uncorrelated Multilinear Princial Comonent Analysis for Unsuervised Multilinear Subsace Learning Haiing Lu, K. N. Plataniotis and A. N. Venetsanooulos The Edward S. Rogers Sr. Deartment

More information

On Wald-Type Optimal Stopping for Brownian Motion

On Wald-Type Optimal Stopping for Brownian Motion J Al Probab Vol 34, No 1, 1997, (66-73) Prerint Ser No 1, 1994, Math Inst Aarhus On Wald-Tye Otimal Stoing for Brownian Motion S RAVRSN and PSKIR The solution is resented to all otimal stoing roblems of

More information