KR102107294B1 - Netwrok topology formation for mobile sensor networks based on network coding - Google Patents

Abstract

According to one embodiment, a utility function may be used to form an optimal network topology for a distributed network. More specifically, a method for forming a network topology for a distributed network topology, the method comprising: determining a utility function for each of a plurality of neighboring nodes that exist around specific nodes; Based on the utility function, it may be a method of forming a network topology including selecting a neighbor node to connect the link with the specific node.

Description

Network topology formation method for mobile sensor network based on network coding. {NETWROK TOPOLOGY FORMATION FOR MOBILE SENSOR NETWORKS BASED ON NETWORK CODING}

The following embodiments relate to a method for forming a network topology for a mobile sensor network based on network coding, and more specifically, a method for forming a network topology using a utility function that considers a distance reduction cost and a link formation cost. It is about.

Modern mobile devices such as smartphones and tablets are sophisticated computing and networking platforms with enhanced sensor capabilities. With the popularization of mobile devices, it is now possible to form an immediate large-scale ad hoc network through connections between mobile devices. Therefore, as mobile devices participate in the network by transmitting data collected by sensors and relaying data from other devices, it is important to configure an optimal network topology that can be optimized.

When constructing an optimal network topology, centralized optimization can calculate all possible topologies and require very high computational complexity. In addition, a bottleneck may be caused by the multicast flow.

Accordingly, a method of forming a network topology that reduces bottlenecks while reducing computational complexity may be needed.

According to an embodiment, a method for forming a network topology may be provided by determining a link with each other node from each node using a utility function that considers a distance shortening gain and a link formation cost.

According to an embodiment, a method of transmitting data using a network topology formed by generating link-independent data by network coding may be provided.

According to an embodiment, a method of forming a network topology for a distributed network topology, comprising: determining a utility function for each of a plurality of neighboring nodes that exist around specific nodes; Based on the utility function, it may be a method of forming a network topology including selecting a neighbor node to connect the link with the specific node.

The utility function may be a method of forming a network topology in consideration of a distance shortening gain determined according to the Euclidean distance from the specific node to the destination node.

The shortening gain may be a method of forming a network topology that increases as the Euclidean distance from the specific node to the destination node decreases or decreases as the Euclidean distance from the specific node to the destination node increases.

The utility function may be a method of forming a network topology considering a link formation cost required to form an outgoing link at the specific node.

The link formation cost may be a method of forming a network topology that decreases when a bidirectional link between the specific node and a neighbor node is connected, rather than when a one-way link between the specific node and a neighbor node is connected.

The network coding may be a network topology forming method that generates link-independent data by combining all packets transmitted to the specific node to generate one packet.

According to an embodiment, in the distributed network topology, receiving incoming data from a plurality of neighbor nodes linked by a specific node; Network coding the incoming data received from the plurality of neighbor nodes to generate one outgoing data; Transmitting the outgoing data to a neighbor node located closer to a destination node than the neighbor nodes that have transmitted the incoming data; And the distributed network topology determines utility functions for each of the plurality of neighbor nodes that receive the outgoing data by being present around the specific node and the specific node, and based on the determined utility function. It may be a data transmission method determined by selecting the neighbor node to transmit outgoing data.

The utility function may be a data transmission method that considers a distance shortening gain determined according to a Euclidean distance from the specific node to a destination node.

The utility function may be a data transmission method considering a link formation cost required to form an outgoing link at the specific node.

The network coding may be a data transmission method that generates link-independent data by combining all incoming data transmitted to the specific node to generate one outgoing data.

According to an embodiment, in a node performing a network topology forming method for a distributed network topology, the node includes a processor, and the processor is effective for each of a plurality of neighboring nodes that exist around specific nodes. It may be a node that determines a function and selects a neighbor node to connect a link with the specific node based on the utility function.

According to an embodiment, in a node that transmits data using a distributed network topology, the node includes a processor, and the processor is configured from a plurality of neighboring nodes linked by a specific node in the distributed network topology. A neighbor node that receives commuting data, network-codes incoming data received from the plurality of neighbor nodes to generate one outgoing data, and is located closer to a destination node than the neighbor nodes that have transmitted the incoming data. To the outgoing data, and the distributed network topology determines a utility function for each of the plurality of neighbor nodes that receive the outgoing data by being present around the specific node and the specific node, and the determined utility The outgoing data is transmitted based on a function It may be a node determined by selecting the neighbor node to be.

According to an embodiment, as a method of forming a network topology, a method of forming a network topology by determining a link with each other node from each node using a utility function that considers a distance shortening gain and a link formation cost is provided. can do.

According to an embodiment, as a method of transmitting data, a method of transmitting data using a network topology formed by generating link independent data by network coding may be provided.

1 is a diagram illustrating a distributed network topology according to an embodiment.
2 illustrates a network topology in which data is transmitted without network coding applied, according to an embodiment.
3 illustrates a network topology in which data is transmitted while network coding is applied, according to an embodiment.
4 illustrates that a network forming game is broken down into a link forming game, according to one embodiment.
5 shows an algorithm for designing a distributed network topology based on a link formation game, according to an embodiment.
6 is a diagram illustrating the number of outgoing links per node when Nv = 50 according to an embodiment.
7 is a diagram showing the number of active links for a unit link formation cost in various network sizes according to an embodiment.
8 is a diagram illustrating a connection failure ratio with respect to a unit link formation cost in various network sizes according to an embodiment.
9 is a diagram showing network utility for the number of nodes and unit link formation cost included in a network topology according to an embodiment.
FIG. 10 is a diagram illustrating network utility for the number of nodes included in a network topology when using three different strategies according to an embodiment.
11 is a diagram illustrating a search space for the number of nodes included in a network topology when using three different strategies according to an embodiment.

Specific structural or functional descriptions of the embodiments are disclosed for illustrative purposes only, and may be implemented in various forms. Accordingly, the embodiments are not limited to a specific disclosure form, and the scope of the present specification includes modifications, equivalents, or substitutes included in the technical spirit.

The terms first or second may be used to describe various components, but these terms should be interpreted only for the purpose of distinguishing one component from other components. For example, the first component may be referred to as the second component, and similarly, the second component may also be referred to as the first component.

When an element is said to be "connected" to another element, it should be understood that other elements may be present, either directly connected to or connected to the other element.

Singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, terms such as “include” or “have” are intended to designate the existence of a described feature, number, step, action, component, part, or combination thereof, one or more other features or numbers, It should be understood that the existence or addition possibilities of steps, actions, components, parts or combinations thereof are not excluded in advance.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by a person skilled in the art. Terms such as those defined in a commonly used dictionary should be interpreted as having meanings consistent with meanings in the context of related technologies, and should not be interpreted as ideal or excessively formal meanings unless explicitly defined herein. Does not.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram illustrating a distributed network topology according to an embodiment.

Modern mobile devices, such as smartphones and tablets, can perform sophisticated computing with enhanced sensor capabilities (eg, user position sensing, atmospheric temperature and humidity measurements). In addition, recent mobile devices may be equipped with sensors (eg, medical sensors) for measurements such as ECG signals, body temperature, blood glucose, heart rate, and blood oxygen saturation. As mobile devices become popular, instantaneous large-scale ac-hoc networks can be formed by connecting mobile devices. The mobile device constituting the network can not only transmit data collected from sensors, but also relay data received from other mobile devices.

At this time, when transmitting data from the source node to the destination node, it is important to form an optimal network topology. Here, the mobile device may be a source node or a destination node. Alternatively, the mobile device may be a specific node that receives data from the source node, and the specific node may transmit the received data to a neighboring node or a destination node.

When designing the optimal network topology, a large network requires high computational complexity. Since the network topology configured by the mobile device may be composed of multiple nodes, the number of potential network topologies that can be formed by the nodes can increase exponentially. Here, the node may represent, for example, a mobile device such as a smart phone, a tablet, or a PDA. Therefore, optimizing the optimal network topology requires very high computational complexity, unless formulated as an optimization problem (eg, a convex optimization problem). In addition, when a mobile device moves, optimization of an optimal network topology considering dynamic factors requires higher computational complexity.

The source node included in the network topology may generate data using the measured data from the sensor, and transmit the generated data to a destination node or a specific node in the vicinity. At this time, since the network topology may include a plurality of source nodes and / or destination nodes, a multicast flow may be generated. However, due to a bottleneck problem, only one flow among multicast flows may be transmitted at a time. Therefore, a method for avoiding the bottleneck will be described in detail in FIGS. 2 and 3 below.

To form a network topology, a network forming game can be applied. At this time, each node constituting the network topology may determine a connection with a neighbor node in consideration of a distance shortening gain and a link formation cost of the utility function. For example, each node constituting the network topology may determine an action for an out-going link in consideration of a distance shortening gain and a link formation cost.

Here, the action may include whether the outgoing link at each node is active or inactive. For example, when the action is 1, the outgoing link may indicate active, and when the action is 0, the outgoing link may indicate inactive.

를 나타낸다. 1 is a network topology composed of 7 nodes according to an embodiment.

Indicates.

는 노드의 세트를 나타내며,

는 연결된 링크의 세트를 나타낸다. 예를 들면, 도 1에서

이고,

를 나타낼 수 있다. 또한, V4 노드에 대해서, V4 노드가 데이터를 수신하는 인커밍 링크는

로 나타낼 수 있고, V4 노드로부터 데이터가 전송되는 아웃고잉 링크는

로 나타낼 수 있다. 이때, 도 1에 나타난 모든 노드는 한번에 하나의 인커밍 링크만 가질 수 있다.

Denotes a set of nodes,

Indicates a set of linked links. For example, in FIG. 1

ego,

Can represent In addition, for the V4 node, the incoming link through which the V4 node receives data is

And an outgoing link through which data is transmitted from the V4 node

Can be represented as At this time, all nodes shown in FIG. 1 may have only one incoming link at a time.

일 수 있다. 또한, 아웃고잉 링크(outgoing link)는 특정 노드로부터 주변 노드로 데이터가 전송되는 링크를 나타내며, 예를 들면, V4 노드로부터 주변 노드로 데이터가 전송되는 아웃고잉 링크는

일 수 있다. 이때, 인커밍 링크를 통해 전송되는 데이터는 인커밍 데이터(incoming data)일 수 있으며, 아웃고잉 링크를 통해 전송되는 데이터는 아웃고잉 데이터(outgoing data)일 수 있다. Here, the incoming link (incoming link) represents a link from which data is transmitted from a neighboring node to a specific node. For example, an incoming link transmitted from a neighboring node to a V4 node is

Can be In addition, the outgoing link (outgoing link) represents a link through which data is transmitted from a specific node to a peripheral node. For example, an outgoing link through which data is transmitted from a V4 node to a peripheral node is

Can be In this case, data transmitted through the incoming link may be incoming data, and data transmitted through the outgoing link may be outgoing data.

에 포함된 Vi 노드는 소스 노드, 목적지 노드, 소스 노드와 목적지 노드 사이에 위치하여 데이터를 릴레이 하는 특정 노드를 모두 포함할 수 있다.

에 포함된 노드의 수는

로 나타날 수 있다. 또한,

와 같이, vi의 목적지 노드는 노드의 세트에 포함될 수 있으며, 목적지 노드의 세트

는

와 같이 나타날 수 있다.According to one embodiment,

The Vi node included in the source node, the destination node, may be located between the source node and the destination node may include all of the specific nodes to relay data.

The number of nodes included in

Can appear as In addition,

As such, the destination node of vi may be included in the set of nodes, and the set of destination nodes

The

May appear as

와 같이 나타날 수 있으며, 이때,

인 경우 Vi 노드와 Vj 노드 간의 링크는 데이터를 전송할 수 있는 활성화 상태를 나타내고,

인 경우 Vi 노드 와 Vj 노드 간의 링크는 데이터를 전송할 수 없는 비활성화 상태를 나타낸다. At this time, the link from the Vi node to the Vj node is

It may appear as

In the case of, the link between the Vi node and the Vj node represents an active state capable of transmitting data,

In this case, the link between the Vi node and the Vj node represents an inactive state in which data cannot be transmitted.

는 Vi 노드에서 Vj 노드로 연결된 것을 나타낸다. 즉, Vi 노드는 테일 노드(tail node)로서 연결된 링크

가 시작되는 노드를 나타낼 수 있으며, Vj 노드는 헤드 노드(head node)로서 연결된 링크

가 끝나는 노드를 나타낼 수 있다. 따라서,

와 달리,

는 테일 노드와 헤드 노드가 반대인 경우를 나타낸다. Here, linked link

Indicates a connection from the Vi node to the Vj node. That is, the Vi node is a link connected as a tail node.

May indicate a node from which the Vj node is connected as a head node.

May indicate a node ending. therefore,

Unlike,

Indicates a case where the tail node and the head node are opposite.

는 Vj 노드의 인커밍 링크를 나타내거나 또는 Vi 노드의 아웃고잉 링크를 나타낼 수 있다. 또한, 연결된 링크의 세트

는 활성화된 링크만 포함할 수 있으며,

는 네트워크 토폴로지

의 활성화된 링크의 개수를 나타낼 수 있다.

May indicate the incoming link of the Vj node or the outgoing link of the Vi node. Also, a set of linked links

Can only contain active links,

Network topology

It may indicate the number of active links.

는 Vi 노드와 Vj 노드간의 유클리디안 거리를 나타낼 수 있다. 예를 들면, Vi 노드와 Vj 노드가 동일한 경우

를 나타내며, Vi 노드와 Vj 노드가 도달할 수 없는 경우

를 나타낼 수 있다. 이때, Vi 노드의 이웃 노드의 세트

는

로서 나타날 수 있으며,

는 연결 경계(connection boundary)를 나타낼 수 있다. 따라서,

인 경우, Vi 노드와 Vj 노드 간의 연결은 형성될 수 없으므로

=0 이며 또한

=0 이다. 그리고, 병목 현상을 피하기 위해,

는 활성화된 인커밍 링크를 가진 Vj 노드의 이웃 노드의 세트를 나타내며,

와 같이 노드는 한번에 하나의 인커밍 링크를 가질 수 있다. In addition,

Can represent the Euclidean distance between the Vi node and the Vj node. For example, if the Vi node and the Vj node are the same

And Vi node and Vj node cannot be reached

Can represent At this time, the set of neighbor nodes of the Vi node

The

May appear as,

May indicate a connection boundary. therefore,

In the case of, since the connection between the Vi node and the Vj node cannot be formed,

= 0 and also

= 0. And, to avoid bottlenecks,

Denotes a set of neighbor nodes of the Vj node with an active incoming link,

As such, a node can have one incoming link at a time.

2 illustrates a network topology in which data is transmitted without network coding applied, according to an embodiment. On the other hand, Figure 3 shows a network topology in which data is transmitted while network coding is applied, according to an embodiment.

Here, FIGS. 2 and 3 are both based on FIG. 1. That is, referring to FIG. 1, neighbor nodes of the V4 node in FIGS. 2 and 3 may be V2, V3, V5, V6, and V7 nodes.

가 Vi 노드에서 수집된 데이터인 경우, 데이터는 특정 노드에서 타겟 노드로 전송될 수 있다. 여기서, 특정 노드는 소스 노드일 수 있으며, 또한 소스 노드와 목적지 노드 사이에 위치하여 데이터를 전송하는 노드 중에서 한 노드일 수 있다. 그리고 타겟 노드는 목적지 노드일 수 있으며, 또한 소스 노드와 목적지 노드 사이에 위치하며 데이터를 수신하는 이웃 노드 중의 한 노드일 수 있다. In Figure 2

When is a data collected from a Vi node, data may be transmitted from a specific node to a target node. Here, the specific node may be a source node, or may be a node located between a source node and a destination node to transmit data. In addition, the target node may be a destination node, or may be a node located between a source node and a destination node and among neighboring nodes receiving data.

는 연결된 링크

에서 전송되는 링크 종속적인 데이터(link dependent data)를 나타낼 수 있다. 네트워크 토폴로지

의 네트워크 상태는 상호 종속성 조건에서

로 표현될 수 있고,

는 아래의 수학식 1과 같이 정의될 수 있다.

Linked link

It may indicate link dependent data transmitted in (link dependent data). Network topology

The network state of

Can be expressed as

Can be defined as in Equation 1 below.

일 경우,

를 만족할 수 있지만, 그렇지 않을 경우

일 수 있다. In Figure 2,

If it is,

Can be satisfied, but if not

Can be

에 네트워크 코딩이 적용될 경우, 네트워크의 상태는

로서 표현될 수 있고,

는 아래의 수학식 2와 같이 정의될 수 있다. However, the network topology

If network coding is applied to the

Can be expressed as

Can be defined as Equation 2 below.

는 Vi 노드에서

노드로 전송되는 네트워크 코딩된 패킷(또는 데이터)을 나타낼 수 있다. 이때,

는 헤더로서의 글로벌 코딩 계수

의 벡터이고 페이로드로서의

일 수 있다.

는 아래의 수학식 3과 같이 정의될 수 있다. 여기서 연산은 갈로이스 필드(GF)에서의 합 및 곱을 나타낼 수 있다. here,

In the Vi node

It may represent a network coded packet (or data) transmitted to a node. At this time,

Is a global coding factor as a header

Is a vector of and as a payload

Can be

Can be defined as in Equation 3 below. Here, the operation may represent the sum and product in the Galois field (GF).

는

노드로 전송되는 모든 패킷을 조합하여 하나의 단일 패킷

를 생성할 수 있다. 그러므로 네트워크 코딩은 수학식 1에서 링크 종속성을 제거할 수 있으며, 링크 종속적인 데이터

를 링크 독립적인 데이터(link independent data)

로 변환할 수 있다. 이와 같이 네트워크 코딩 함수

가 적용된 일 실시예는 도 3과 같을 수 있다.Therefore, the network coding function

The

Combines all packets sent to a node, resulting in a single packet

Can generate Therefore, network coding can remove link dependency in Equation 1 and link-dependent data

To link independent data

Can be converted to Network coding functions like this

3 may be as shown in FIG. 3.

4 illustrates that a network forming game is broken down into a link forming game, according to one embodiment.

에 포함된 노드 간의 연결은 엣지-디스조인트 서브그래프

에 의한 2 노드 간의 독립적인 게임으로 분해될 수 있다. 엣지-디스조인트 서브그래프

에 의한 2 노드 간의 연결은, 링크 형성 게임

에 의한 2 노드 간의 연결로 분해될 수 있다.Network formation game

Connections between nodes included in the edge-disjoint subgraph

It can be decomposed into an independent game between 2 nodes. Edge-disjoint subgraph

The connection between 2 nodes by, link formation game

It can be decomposed into a connection between 2 nodes.

에 의한 2 노드 간의 연결은 목적지 노드의 인덱스 세트 D에 의해 여러 목적지 노드를 향하는 링크를 포함할 수 있다. 예를 들면, 도 4에서 확인가능하듯이 V4 노드와 V5 노드 간의 연결은 목적지 노드로서 V6 노드와 V7 노드를 향하는 링크를 포함할 수 있다.Here, the edge-disjoint subgraph

The connection between the two nodes by may include links to various destination nodes by index set D of the destination node. For example, as can be seen in FIG. 4, the connection between the V4 node and the V5 node may include a link to the V6 node and the V7 node as the destination node.

에 의한 2 노드 간의 연결은 하나의 목적지 노드를 향하는 링크를 포함할 수 있다. 예를 들면, 도 4에서 확인가능하듯이 V4 노드와 V5 노드 간의 연결은 목적지 노드로서 V7 노드를 향하는 링크를 포함할 수 있다. On the other hand, link formation game

The connection between the two nodes by may include a link to one destination node. For example, as can be seen in FIG. 4, the connection between the V4 node and the V5 node may include a link to the V7 node as a destination node.

><Network formation game

>

, 목적지 노드의 인덱스의 세트

일 때, 네트워크 형성 게임은 아래의 수학식 4와 같이 표현될 수 있다. 여기서

로서, Vi 노드의 목적지 노드를 위한 인덱스 벡터를 나타낼 수 있다. Set of nodes

, The set of indexes of the destination node

In this case, the network formation game may be expressed as Equation 4 below. here

As, can represent an index vector for the destination node of the Vi node.

는 플레이어(즉, 노드)의 유한한 세트를 나타내며,

는 Vi 노드의 액션의 유한한 세트를 나타내며,

는 Vi 노드의 효용 함수를 나타낸다. here,

Denotes a finite set of players (i.e. nodes),

Denotes a finite set of actions of the Vi node,

Denotes the utility function of the Vi node.

에 포함되며, 이웃 노드 Vj와 링크 형성하는 결정을 하는 네트워크 형성 게임의 플레이어로서 고려될 수 있다. 그리고, Vi 노드의 액션은

와 같이 표현될 수 있고, 이는 플레이어의 이웃 노드를 향한 아웃고잉 링크의 활성화 또는 비활성화를 나타낼 수 있다.At this time, the Vi node is a set of nodes

And may be considered as a player in a network forming game that makes a decision to link with a neighbor node Vj. And the action of Vi node

It may be expressed as, which may indicate activation or deactivation of the outgoing link toward the player's neighbor node.

는 Vi 노드를 제외한 플레이어(즉, 노드)의 액션의 세트를 나타낼 수 있다. The utility function of the Vi node can be expressed as Equation 5 below. At this time,

May represent a set of actions of the player (ie, node) except for the Vi node.

에 대한 거리단축이득(reward)는 아래의 수학식 6과 같이 표현될 수 있다. 이때,

는 액션

에 의해 결정될 수 있다. Player Vi's action in equation (5)

The distance reduction gain for can be expressed as Equation 6 below. At this time,

The action

It can be determined by.

에 의해서 목적지 노드를 향하는 거리 감소(distance reduction)를 나타낼 수 있다. 여기서,

는 Vi 노드에서 Vj 노드로 향하는 거리의 벡터를 나타낼 수 있으며,

는 거리의 벡터

에 반비례(inversely proportional)할 수 있다.In Equation 6, the distance reduction reward is the action

By this, it is possible to indicate a distance reduction toward the destination node. here,

Can represent a vector of distances from the Vi node to the Vj node,

Vector of the street

Can be inversely proportional to.

For example, when transmitting data from the node Vi to the destination node Vd, data may be transmitted from a neighbor node Vk and Vj to a Vj node that is closer to the destination node Vd. That is, data may be transmitted from the Vi node to the Vj node, and data may be transmitted from the Vj node to the Vd node.

에서, 링크형성비용(cost)은 아래의 수학식 7과 같이 표현될 수 있다. 여기서,

는 링크 형성하기 위한 단위 링크형성비용(unit cost)를 나타낼 수 있다. Selected Action in Equation 5 above

In, the link formation cost can be expressed as Equation 7 below. here,

Can denote a unit cost of link formation.

는 노드 Vi가 형성하는 아웃고잉 링크를 위해 필요한 총 지출(total payment)을 나타낼 수 있다. 여기서, 총 지출은 아웃고잉 링크를 통해 데이터가 전송될 때 필요한 링크형성비용을 나타낼 수 있다. Equation (7)

Can represent the total payment required for the outgoing link formed by node Vi. Here, the total expenditure may represent a link formation cost required when data is transmitted through the outgoing link.

인 경우

는

일 수 있다. 다른 예를 들면, Vi 노드와 Vj 노드 간의 링크를 모두 형성한 경우,

일 수 있다. 또 다른 예를 들면,

인 경우,

일 수 있다. For example,

If

The

Can be For another example, when all the links between the Vi node and the Vj node are formed,

Can be Another example:

If it is,

Can be

According to an embodiment, the network forming game may be decomposed into a plurality of link forming games. Therefore, in the case of a link formation game based on network coding, computational complexity can be reduced.

<Network coding>

는 네트워크 토폴로지

를 구성하는 엣지-디스조인트 서브그래프(edge-disjoint subgraphs)일 수 있다. 이때, 엣지-디스조인트 서브그래프

는

일 때,

를 만족하며,

및

를 만족할 수 있다.

Network topology

It may be edge-disjoint subgraphs constituting (edge). At this time, the edge-disjoint subgraph

The

when,

Satisfying,

And

Can be satisfied.

를 나타낼 수 있다. 목적 노드 인덱스 세트 D를 갖는 엣지-디스조인트 서브그래프

을 위한 네트워크 형성 게임은 아래의 수학식 8과 같이 표현될 수 있다. In other words, the network topology as the union of the edge-disjoint subgraph.

Can represent Edge-disjoint subgraph with target node index set D

The network formation game for can be expressed as Equation 8 below.

Here, actions determined by players performing a network formation game may be expressed as a union of links representing active and inactive in a network topology. Therefore, the product operation representing the simultaneous consideration of network forming games for a plurality of edge-disjoint subgraphs is expressed as Equation 9 below, which forms a network for the edge-disjoint subgraph. The result of the game can be considered as a union of network states.

은 엣지-디스조인트 서브그래프를 위한 네트워크 형성게임

로 분해될 수 있다. According to an embodiment, as shown in Equation 10 below, a network formation game for a network topology based on network coding

Is a network forming game for edge-disjoint subgraphs

Can be decomposed into

If network coding is not applied, it may be as shown in Equation 11 below.

However, when network coding is applied, Equation 11 may be expressed as Equation 12. Therefore, the inter-link dependency of Equation 11 can be eliminated in Equation 12.

이 하나의 목적지 노드 Vd를 갖는 링크 형성 게임

로 분해될 수 있음을 나타낼 수 있다. 따라서, 네트워크 코딩에 기초하여 멀티캐스트 플로우(multicast flow)를 갖는 네트워크 형성 게임은 유니캐스트 플로우를 갖는 독립적인 링크 형성 게임으로 분해될 수 있다. According to another embodiment, Equation 11 is a network formation game having a plurality of destination nodes.

Link formation game with this one destination node Vd

It can indicate that can be decomposed. Accordingly, a network forming game with a multicast flow based on network coding can be decomposed into an independent link forming game with a unicast flow.

가 노드 Vi의 가상의 서브 노드(virtual subnode)이며, Vd는 목적지 노드를 나타낼 수 있다. 목적지 노드 Vd를 갖는

의 가상의 서브 링크는

로 표시될 수 있다. 목적지 노드 Vd를 위한 가상의 서브그래프는

로 표현될 수 있다. 이때, 가상의 서브그래프

는

일 때,

를 만족하며,

및

를 만족할 수 있다. 여기서, 가상의 서브그래프는 엣지-디스조인트 서브그래프(Edge-disjoint subgraphs)일 수 있다.

Is a virtual subnode of node Vi, and Vd can represent a destination node. With destination node Vd

The virtual sublink of

It may be indicated by. The hypothetical subgraph for the destination node Vd

Can be expressed as At this time, a virtual subgraph

The

when,

Satisfying,

And

Can be satisfied. Here, the virtual subgraphs may be edge-disjoint subgraphs.

는 링크 형성 게임

으로 분해될 수 있다.At this time, the network forming game for the edge-disjoint subgraph based on network coding as shown in Equation 13 below.

Link formation game

Can be decomposed into

Accordingly, Equation 14 below may be derived based on Equation 13 and Equation 12.

는 유니캐스트 플로우를 갖는 링크 형성 게임

를 갖는 독립적인 링크로 분해될 수 있다. Hence, a network-forming game for edge-disjoint subgraphs with multicast flows.

Is a link formation game with unicast flow

It can be disassembled into independent links with.

<Design of link formation game and distributed network topology>

로서 두 플레이어(노드)들의 집합을 나타내며,

로서 플레이어 Vi의 가능한 액션 집합을 나타낼 수 있다. 예를 들면, a_i=1인 경우 플레이어 Vi는 활성화 링크를 가지면, a_i=0인 경우 플레이어 Vi는 비활성화 링크를 가질 수 있다. The link formation game may consist of 2 players (nodes) having a unicast flow. At this time, the link formation game may be expressed as Equation 15 below. here,

Denotes a set of two players (nodes),

It can represent the set of possible actions of player Vi. For example, if a _i = 1, the player Vi may have an active link, and if a _i = 0, the player Vi may have an inactive link.

는 거리를 나타내는

에 반비례(inversely proportional)할 수 있다.In this case, the utility function may be expressed as Equation 16 below. here,

Indicating the distance

Can be inversely proportional to.

In the link formation game, the link formation cost of the utility function may be expressed as Equation 17 below.

및

는 NE를 만족하는 Vi와 Vj의 액션을 각각 나타낸다. 예를 들면, 링크 형성 게임에 대한 a_i가 0 또는 1을 갖는 경우,

은 a_i중에서 NE을 만족하는 액션으로서 '0' 또는 '1' 또는 '0, 1' 둘 다일 수 있다. The Nash equilibrium (NE) for the link formation game can be expressed as Equation 18 below. here,

And

Denotes actions of Vi and Vj satisfying NE, respectively. For example, if a _i for a link formation game has 0 or 1,

Is an action that satisfies NE among a _i , and may be both '0' or '1' or '0, 1'.

로 표현되는 복수개의 NE를 가질 수 있다. 복수개의 NE중에서 선택된 NE는 Vi 노드 및 Vj 노드의 아웃고잉 링크를 활성화(active) 및 비활성화(inactive)하는 최적화 결정을 가능하도록 할 수 있다. If there are multiple NEs, the link formation game

It may have a plurality of NE represented by. The NE selected from the plurality of NEs may enable optimization decisions to activate (active) and inactive (active) the outgoing links of the Vi node and the Vj node.

5 shows an algorithm for designing a distributed network topology based on a link formation game, according to an embodiment.

The algorithm can proceed in the following order.

일 때, 노드의 세트

, 이웃 노드의 세트

, 목적지 노드의 인덱스 세트

, 효용 함수

일 수 있다.

If is, set of nodes

, Set of neighbor nodes

, The index set of the destination node

, Utility function

Can be

)할 수 있다. 이후 노드의 세트

를 유니캐스트 플로우를 가지는 2개의 플레이어(노드)로 분해하고, 현재의 링크의 액션

을 임시 저장할 수 있다. Reset all connected link states (

)can do. Set of nodes after

Decomposes into 2 players (nodes) with unicast flow, and the action of the current link

Can be temporarily stored.

,에 대해 링크 형성 게임을 만족하는

및

를 식별할 수 있다. 여기서,

는 아래의 수학식 19를 만족할 수 있다. 이때,

는 Vi 노드에서 Vj 노드로의 연결된 링크(

)일 수 있으며,

는 Vj 노드에서 Vi 노드로의 연결된 링크(

)일 수 있다. Each disassembled player, for example

About to satisfy the link formation game,

And

Can be identified. here,

Can satisfy Equation 19 below. At this time,

Is the connected link from the Vi node to the Vj node (

),

Is the linked link from the Vj node to the Vi node (

).

및

는 NE을 만족하는 것으로서, 알고리즘에 의해 적어도 하나의

및

는 식별될 수 있다. 또한, 복수개의

및

가 식별된 경우, 최적화

및

는 선택될 수 있다.

및

에 의해

의 링크는 결정될 수 있다.Identified

And

Is satisfied with NE, at least one by algorithm

And

Can be identified. Also, multiple

And

If is identified, optimization

And

Can be selected.

And

By

The link of can be determined.

를 임시 저장된

와 비교할 수 있다.

는 제1 링크상태를 나타낼 수 있고,

는 제2 링크상태를 나타낼 수 있다. 이때, 제1 링크상태는 현재의 링크상태일 수 있고, 제2 링크상태는 검출된 링크상태일 수 있다. According to one embodiment, Equation 19 is satisfied

Temporarily stored

Can be compared with

Can indicate the first link state,

May indicate the second link state. At this time, the first link state may be the current link state, and the second link state may be the detected link state.

와 제1 링크상태

가 동일하지 않은 경우 임시 저장된

를 대신하여

로 변경하여 저장할 수 있다. 이때, 제2 링크상태

에 기초하여, 다시 제3의 링크상태를 검출할 수 있다. 마찬가지로, 제2 링크상태와 제3의 링크상태를 비교하는 과정을 반복할 수 있다. At this time, as a result of comparing the temporarily stored first link state with the detected second link state, the second link state

And first link status

Temporarily stored if is not the same

On behalf of

You can save it by changing to. At this time, the second link state

Based on this, the third link state can be detected again. Similarly, the process of comparing the second link state and the third link state can be repeated.

와 제1 링크상태

가 동일한 경우, 검출된 제2 링크상태에 기초하여 네트워크 토폴로지를 구성하는 노드 간의 링크를 결정할 수 있다. Alternatively, as a result of comparing the temporarily stored first link state with the detected second link state, the second link state

And first link status

If is the same, the link between nodes constituting the network topology may be determined based on the detected second link state.

및

를 식별하여, 노드 간의 링크를 결정될 수 있다. 이로 인해, 복수의 노드로 구성되는 네트워크 토폴로지는 형성될 수 있다.In addition, each of the plurality of nodes constituting the network topology satisfies the link formation game.

And

By identifying the link between the nodes can be determined. Due to this, a network topology composed of a plurality of nodes can be formed.

6 is a diagram illustrating the number of outgoing links per node when Nv = 50 according to an embodiment.

로 설정될 수 있으며, Vi 노드의 이웃 노드는

로서 결정될 수 있다. 여기서, 각각의 노드는 도 5의 알고리즘에 기초하여 이웃 노드를 향한 아웃고잉 링크를 결정할 수 있다. Assuming that there is a node of Nv in the cell of radius R, the Vi node can send data to the destination node. At this time, the connection boundary (connection boundary)

Can be set to, the neighbor node of the Vi node

It can be determined as. Here, each node may determine an outgoing link to a neighbor node based on the algorithm of FIG. 5.

는 아래의 수학식 20와 같이 표현될 수 있으며, 수학식 20에 표현된 대로 노드의 위치가 고려될 수 있다.At this time, the equation (16)

May be expressed as in Equation 20 below, and the position of the node may be considered as expressed in Equation 20.

는

에 반비례하는 관계로서,

일 때

일 수 있다. 따라서, Vi 노드가 목적지 노드에 가까이 있는 Vj 노드로 아웃고잉 링크를 연결할 경우, 양의 거리단축이득(positive rewards)가 Vi 노드에 주어질 수 있다.

The

Inversely proportional to,

when

Can be Therefore, when the Vi node connects the outgoing link to the Vj node close to the destination node, positive rewards may be given to the Vi node.

According to an embodiment, when forming a network topology in which data is transmitted from a source node to a destination node, an outgoing link of a node closer to the destination node may be more important than an outgoing link of a node remote from the destination node. . This is because a node located farther from the destination node has a higher probability of having path diversity.

및

를 고려할 때, 2개의 노드 페어는 ⅰ) 다음 수학식 21과 같이 각각의 노드 페어에서 노드 간의 거리가 동일한 경우이고 ⅱ) 다음 수학식 22과 같이 다른 위치일 수 있다. For example, two node pairs

And

Considering, the two node pairs may be ⅰ) the distance between nodes in each node pair is the same as in the following Equation 21, and ii) may be different positions as in the following Equation 22.

In this case, the difference according to Equation 20 between nodes is as shown in Equation 23 below.

보다

가 목적지 노드에 보다 가까우며 보다 많은 거리단축이득(reward)를 가질 수 있음을 알 수 있다. 이때, 효용 함수는 아래의 수학식 24와 같이 표현될 수 있다.Therefore, of the two node pairs

see

It can be seen that is closer to the destination node and can have more distance reduction. At this time, the utility function may be expressed as Equation 24 below.

Therefore, the network utility for evaluating the performance of the formed network topology can be expressed as Equation 25 below.

This may include the total shortening gain from all destination nodes and the total link formation cost required to form the links constituting the network topology. Therefore, the network utility can be used to measure the distance reduction gain obtained from the node, which reduces the link formation cost of link formation.

)인 경우를 전제로 할 수 있다. 6 is a connection boundary R = 10, and is a graph showing the average value of 1000 independent experiments, with two destination nodes (

).

The graphs 610 and NEAR are close to the destination node, the graphs 620 and MID are located in the middle of the destination node, and the graphs 630 and FAR may indicate the case where the destination node is far away.

는 소스 노드와 목적지 노드 간의 거리를 고려하기 때문이다. As can be seen from the graphs 610 to 630, it can be seen that the closer to the destination node, the more outgoing links it has. This is in the utility function

This is because the distance between the source node and the destination node is considered.

7 is a diagram showing the number of active links for a unit link formation cost in various network sizes according to an embodiment.

When the graph 710 is the network size Nv = 10, when the graph 720 is the network size Nv = 20, when the graph 730 is the network size Nv = 30, the graph 740 is the network size Nv = 40 In this case, the graph 750 may indicate a case where the network size Nv = 50.

7, it can be seen that as the number of network sizes increases, the number of activation links included in the network topology also increases. In addition, it can be seen that when the number of smaller active links is included even in the same network size, the unit link formation cost increases.

8 is a diagram illustrating a connection failure ratio with respect to a unit link formation cost in various network sizes according to an embodiment.

The number of active links included in the network topology can have a significant impact on a successful connection from the source node to the destination node. At this time, in order to evaluate the effect of the number of active links on a successful connection, a connection failure ratio may be measured.

When the unit link formation cost is 0.8 to 1, the connection failure rate of the network topology is very steeply increased, so it can be confirmed that there is a high possibility that data transmission is not successfully performed.

9 is a diagram showing network utility for the number of nodes and unit link formation cost included in a network topology according to an embodiment.

가 증가하거나 또는 단위 링크형성비용

이 감소하는 경우, 네트워크 유틸리티는 증가할 수 있다. 이는 보다 많은 거리단축이득이 획득되거나 또는 링크 형성하는 링크형성비용이 감소될수록 네트워크 유틸리티가 증가하기 때문이다. 예를 들면, 단위 링크형성비용

=1인 경우, 어떠한 링크도 형성되지 않아 네트워크 유틸리티는 0일 수 있다. Number of nodes in the network topology

Increases or the cost of unit link formation

If this decreases, network utilities can increase. This is because network utility increases as more distance reduction gain is obtained or link formation cost of link formation is reduced. For example, unit link formation cost

If = 1, no link is formed and the network utility may be 0.

FIG. 10 is a diagram illustrating network utility for the number of nodes included in a network topology when using three different strategies according to an embodiment. 11 is a diagram illustrating a search space for the number of nodes included in a network topology when using three different strategies according to an embodiment.

The performance of the algorithm of FIG. 5 can be evaluated in comparison with two other strategies in terms of network utility and computational complexity. The two strategies are:

1. Non-NC Centralized Strategy

Since network coding is not applied, it can be expressed as Equation 26 below as a strategy that satisfies inter-link dependency.

2. NC Centralized Strategy

As a strategy in which network coding is applied and dependencies between links are ignored, it can be expressed as Equation 27 below. Here, the NC-Centralized strategy represents a method of selecting an action in which the entire utility function of the network topology is maximized, taking into account all the number of action cases that all players can have. However, the proposed strategy represents a method of selecting an action from which each player's own utility function can be maximized, from among actions selectable by each player.

The NC-Centralized strategy has the highest network utility, and the Non-NC Centralized strategy has the lowest network utility.

Since the Non-NC Centralized strategy needs to be optimized by considering link interdependencies as constraints, the optimal network topology cannot achieve a higher shortening gain than other network coding-based strategies that create multiple outgoing links.

The computational complexity of each strategy can be measured by the size of the search space determined by the size of the action available for link formation.

일 수 있다.In the Non-NC Centralized strategy, each node creates at most one outgoing link, so if there are many nodes, Nv links can be selected. Therefore, since there are Nv nodes in the network topology, the maximum selectable number is

Can be

개의 이웃 노드가 링크를 만들 수 있고 각각의 링크는 활성화 비활성화 2가지 선택을 가지고 있으므로, 각각의 노드는 최대로

선택을 가질 수 있다. 따라서, 네트워크 토폴로지에 Nv의 노드가 있으므로, 최대로 선택가능한 수는

일 수 있다. In the NC-Centralized strategy,

Since two neighbor nodes can create a link, and each link has two choices of active and inactive, each node is maximized.

You can have a choice. Therefore, since there are Nv nodes in the network topology, the maximum selectable number is

Can be

개의 이웃 노드 페어가 있을 수 있고 각각의 노드 페어는 4개의 액션 선택을 가질 수 있다. 따라서, 각각의 노드 페어는 액션을 독립적으로 선택할 수 있으므로, 탐색 공간(search space)의 크기는

일 수 있다. 여기서, 탐색 공간은 네트워크 토폴로지의 전체 효용 함수가 최대화 되는 액션을 선택할 때, 모든 플레이어가 가질 수 있는 액션조합의 경우의 수에 대응할 수 있다. In the proposed algorithm strategy of Figure 5, to the maximum

There may be four neighbor node pairs, and each node pair may have four action selections. Therefore, each node pair can independently select an action, so the size of the search space

Can be Here, the search space may correspond to the number of action combinations that all players can have when selecting an action in which the entire utility function of the network topology is maximized.

로서 대규모의 네트워크에 실제로 적용될 수 있다. The NC-Centralized strategy can be difficult to apply in practice due to its high computational complexity. However, in the case of the proposed algorithm of Figure 5, the computational complexity is

As it can be applied to large-scale networks.

The calculation complexity is described in Table 1 below, which can also be confirmed through FIG. 11.

The embodiments described above may be implemented with hardware components, software components, and / or combinations of hardware components and software components. For example, the devices, methods, and components described in the embodiments include, for example, a processor, controller, arithmetic logic unit (ALU), digital signal processor (micro signal processor), microcomputer, field programmable gate (FPGA). It can be implemented using one or more general purpose computers or special purpose computers, such as arrays, programmable logic units (PLUs), microprocessors, or any other device capable of executing and responding to instructions. The processing device may run an operating system (OS) and one or more software applications running on the operating system. In addition, the processing device may access, store, manipulate, process, and generate data in response to the execution of the software. For convenience of understanding, a processing device may be described as one being used, but a person having ordinary skill in the art, the processing device may include a plurality of processing elements and / or a plurality of types of processing elements. It can be seen that may include. For example, the processing device may include a plurality of processors or a processor and a controller. In addition, other processing configurations, such as parallel processors, are possible.

The software may include a computer program, code, instruction, or a combination of one or more of these, and configure the processing device to operate as desired, or process independently or collectively You can command the device. Software and / or data may be interpreted by a processing device, or to provide instructions or data to a processing device, of any type of machine, component, physical device, virtual equipment, computer storage medium or device. , Or may be permanently or temporarily embodied in the transmitted signal wave. The software may be distributed on networked computer systems, and stored or executed in a distributed manner. Software and data may be stored in one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer readable medium. The computer-readable medium may include program instructions, data files, data structures, or the like alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the embodiments or may be known and usable by those skilled in computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs, DVDs, and magnetic media such as floptical disks. -Hardware devices specially configured to store and execute program instructions such as magneto-optical media, and ROM, RAM, flash memory, and the like. Examples of program instructions include high-level language code that can be executed by a computer using an interpreter, etc., as well as machine language codes produced by a compiler. The hardware device described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

As described above, although the embodiments have been described with limited drawings, those skilled in the art can apply various technical modifications and variations based on the above. For example, the described techniques are performed in a different order than the described method, and / or the components of the described system, structure, device, circuit, etc. are combined or combined in a different form from the described method, or other components Or even if replaced or substituted by equivalents, appropriate results can be achieved

Claims (12)

Decomposing the entire network composed of a plurality of nodes having a multicast flow into a plurality of subgraphs pairing two adjacent nodes having a unicast flow;
Determining a link between two nodes present in each of the subgraphs by performing a link formation game using a utility function for each of the decomposed subgraphs; And;
Forming a topology of the entire network composed of the plurality of nodes based on a link between two nodes for each of the determined subgraphs
Network topology forming method comprising a. According to claim 1,
The decomposing step,
A method of forming a network topology that decomposes the entire network into a plurality of subgraphs through network coding that combines all packets transmitted to a destination node to generate a single link-independent single packet. According to claim 1,
The decomposing step,
Each of the two nodes includes virtual sub-nodes according to a destination node, and a method of forming a network topology comprising virtual sub-nodes having the same destination node. According to claim 1,
The utility function,
A method of forming a network topology considering a link formation cost required to form an outgoing link in each of a plurality of nodes constituting the entire network. According to claim 4,
The link formation cost,
A method of forming a network topology that decreases when a two-way link between two adjacent nodes is connected than when a one-way link between two adjacent nodes is connected among a plurality of nodes constituting the entire network. According to claim 1,
Determining a link between two nodes present in each of the subgraphs,
Initializing all link states and storing the current first link state;
Performing a link formation game for each of the decomposed subgraphs to detect a second link state that satisfies a utility value for evaluating the performance of a network topology;
Identifying a link between two final nodes based on a result of comparing the stored first link state with the detected second link state.
Network topology forming method comprising a. The method of claim 6,
The step of identifying the link between the final two nodes is:
If the result of comparing the stored first link state and the detected second link state is the same, the network topology forming method determines the detected second link state as a link between the final two nodes. The method of claim 6,
The step of identifying the link between the final two nodes is:
If the comparison between the stored first link state and the detected second link state is not the same, the detected second link state is changed to the current first link state and stored, and the second link state is detected again Thereby determining the link between the two final nodes. The method of claim 6,
Determining a link between two nodes present in each of the subgraphs,
A method of forming a network topology for determining whether an outgoing link of each of the two nodes is active through Nash Equilibrium (NE).
delete delete delete

Images