國立虎尾科技大學 |

Network Coding for Function Computation.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Network Coding for Function Computation./
Author:	Tripathy, Ardhendu Shekhar.
Description:	1 online resource (120 pages)
Notes:	Source: Dissertation Abstracts International, Volume: 79-11(E), Section: B.
Contained By:	Dissertation Abstracts International79-11B(E).
Subject:	Electrical engineering. -
Online resource:	click for full text (PQDT)
ISBN:	9780438077157

Network Coding for Function Computation.
Tripathy, Ardhendu Shekhar.

Network Coding for Function Computation. - 1 online resource (120 pages)

Source: Dissertation Abstracts International, Volume: 79-11(E), Section: B.

Thesis (Ph.D.)--Iowa State University, 2018.

Includes bibliographical references

Many applications such as parallel processing, distributed data analytics and sensor networks often need to compute functions of data that are observed in a distributed manner over a network. A network can be modeled as a directed graph, each vertex of which denotes a node that can carry out computations and communicate with its neighbors. The edges of the graph denote one-way noiseless communication links. A subset of nodes---called sources---observe independent messages, and a possibly different subset of nodes---called terminals---wish to compute a particular demand function of the messages. The information transmitted on the edges are specified by a set of functions, one for each edge; this set of functions is called a network code. We are interested in network codes that allow each terminal to compute the demanded function with zero-error.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2018

Mode of access: World Wide Web

ISBN: 9780438077157Subjects--Topical Terms:

596380
Electrical engineering.
Index Terms--Genre/Form:

554714
Electronic books.

Network Coding for Function Computation.
LDR:04374ntm a2200373Ki 4500 001 919173
005 20181116131021.5
006 m o u
007 cr mn||||a|a||
008 190606s2018 xx obm 000 0 eng d
020 $a 9780438077157
035 $a (MiAaPQ)AAI10790290
035 $a (MiAaPQ)iastate:17298
035 $a AAI10790290
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Tripathy, Ardhendu Shekhar. $3 1193683
245 1 0 $a Network Coding for Function Computation.
264 0 $c 2018
300 $a 1 online resource (120 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertation Abstracts International, Volume: 79-11(E), Section: B.
500 $a Adviser: Aditya Ramamoorthy.
502 $a Thesis (Ph.D.)--Iowa State University, 2018.
504 $a Includes bibliographical references
520 $a Many applications such as parallel processing, distributed data analytics and sensor networks often need to compute functions of data that are observed in a distributed manner over a network. A network can be modeled as a directed graph, each vertex of which denotes a node that can carry out computations and communicate with its neighbors. The edges of the graph denote one-way noiseless communication links. A subset of nodes---called sources---observe independent messages, and a possibly different subset of nodes---called terminals---wish to compute a particular demand function of the messages. The information transmitted on the edges are specified by a set of functions, one for each edge; this set of functions is called a network code. We are interested in network codes that allow each terminal to compute the demanded function with zero-error.
520 $a In the first part of this thesis, we assume that the message random variables are independent and uniformly distributed over a finite field. The demand function is set to be the finite field sum of all the messages observed in the network. A valid network code for this sum-network problem allows each terminal to compute the sum, and has an associated computation rate. We wish to find the best possible computation rate for a given sum-network; this value is called its computation capacity. Finding the computation capacity of a sum-network is known to be a difficult problem. Here we are able to evaluate it for certain systematically constructed sum-network problem instances. The construction procedure uses incidence structures, whose combinatorial properties allow us to be able to evaluate the computation capacity of the constructed sum-networks. An important aspect of the problem that we uncover is the strong dependence of the computation capacity on the finite field over which the sum is to computed. This is shown by a sum-network, whose computation capacity is 1 over a finite field and close to 0 over a different finite field. We also construct sum-networks whose computation capacity can take on arbitrarily many different values over different finite field alphabets.
520 $a In the second part of the thesis, we focus on a particular directed acyclic network that has four nodes and four edges. It is the simplest instance of a network that does not have a tree structure. Three of the nodes are sources that observe independent messages that are uniformly distributed over a finite discrete alphabet. The fourth node is a terminal which wants to compute a demand function of the three messages. The demand function is an arbitrary discrete-valued function. We focus on network codes that have different rates on each of the four edges, thus we have a rate tuple associated with every valid network code. The collection of rate tuples for all valid network codes form a rate region, and we describe a procedure to obtain an outer bound to this rate region. We illustrate our approach through different example demand functions. When the demand function is the finite field sum over GF(2), we give a network code whose rate tuple matches the outer bound.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2018
538 $a Mode of access: World Wide Web
650 4 $a Electrical engineering. $3 596380
650 4 $a Information science. $3 561178
650 4 $a Mathematics. $3 527692
655 7 $a Electronic books. $2 local $3 554714
690 $a 0544
690 $a 0723
690 $a 0405
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a Iowa State University. $b Electrical and Computer Engineering. $3 1179387
773 0 $t Dissertation Abstracts International $g 79-11B(E).
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10790290 $z click for full text (PQDT)