KR101572414B1

KR101572414B1 - A Context Information-based Routing Scheme with Energy-based Message Prioritization for Delay Tolerant Networks

Info

Publication number: KR101572414B1
Application number: KR1020150063729A
Authority: KR
Inventors: 나인호
Original assignee: 군산대학교산학협력단
Priority date: 2015-05-07
Filing date: 2015-05-07
Publication date: 2015-11-27

Abstract

The present invention relates to a context information-based priority message routing scheme and a system for recommending a social network on the basis of essential attribute information and a method thereof. More specifically, the present invention relates to the context information-based priority message routing scheme and the system for recommending the social network on the basis of essential attribute information and the method thereof, which has robustness, reliability, ability for unexpected discovery, variety, privacy preservation and expandability, and can maintain stability in case of false information or attack, and can preserve privacy by performing a recommendation process of a user as increasing variety of a recommendation list catalog on the basis of the ability and behavior of the user under a specific situation at a given time, and can induce prediction as to tendency of a new user in social neighborhood relation connection, and can reduce the time for computerizing the recommendation by that a connection network among friends can induce specific interest related with an active user, and direct and close friends can be sufficient social graph material used in the prediction, and can compare distances of source node (SN), intermediate node (IN) and destination node (DN) in real-time on the basis of direction, speed and current position thereof through a routing scheme through message connection table (MCT) and the like.

Description

Technical Field [0001] The present invention relates to a social network recommendation system based on context information based priority message routing and essential attribute information, and a method thereof. [0002]

The present invention relates to a system and method for recommending a social network based on context information based priority message routing and essential attribute information, and more particularly, to a system and method for recommending social networks, Preservation, and scalability of a user, and can maintain stability when subjected to fake information or attacks, and can increase the diversity of the list of recommendation lists based on the user's ability and behavior in a given situation at a given time, It is possible to preserve privacy and to induce speculation as to the preferences of new users in social neighbors connections, and that the network of friends can be a specific interest that can be associated with an active user Can be induced and direct and close friends can be induced The social graph data can already be created, which can shorten the computerization time of the recommendation. It can also be used for real-time communication of SN, IN, DN based on the moving direction, To a social network recommendation system based on comparable context information based priority message routing and essential attribute information, and a method thereof.

In general, a delay-tolerant network (DTN) is an asymmetric networking technology that can be deployed in a networking environment in which stable communication connectivity is not guaranteed. It stores data received in a storage space, Only the data is transmitted. DTN is applied not only to space communication supporting communication between satellites but also to sensor networks and MANET (Mobile Ad-Hoc Network). In DTN networking environments, it is very important to provide a way to improve overall routing speed and performance, since the overhead is small and routing reliability is high.

On the other hand, the accuracy of predictions is the most common metric used to evaluate the performance of traditional recommendation systems. However, this does not apply to social network based recommendation systems that use social connections when making predictions or suggestions.

Other important features are considered when implementing and evaluating predictions or implications. This paper evaluates and measures various social network based recommendation systems in terms of robustness, trust, ability to make unexpected discoveries, diversity, preservation of privacy, and scalability.

Through observation and analysis, the authors of this paper suggest some suggestions to improve the implementation of various recommender systems.

The rapid increase in the amount of information on the web has made it difficult for Internet users to get the information they want. This problem is exacerbated if the user does not use appropriate navigation tools.

Over the past decade, various Recommendator Systems (RSs) have been proposed to address this problem.

RSs in highly acclaimed sites such as Amazon, Netflix, TripAdvisor, Yahoo, and YouTube have been very important to their success. The important thing is to provide users with interesting items based on their existing preferences, transactions and profiles, so that they can make sound decisions.

Integration of social networks has opened up a new field of research in recommender systems. Because of the numerous social networking websites such as Facebook and LinkedIn and Twitter, it is highly desirable to use a system that integrates information from these resources to provide personalized recommendations for individuals, groups, or communities.

Incorporating knowledge from social networks, such as social influences or social interactions, comes from the fact that users are often assisted by opinions or referrals from their friends.

The first step in selecting a suitable RS algorithm is to determine which features of the application to focus on. Knowing the important characteristics to be considered and understanding their effects will help designers to implement appropriate recommendation system approaches and algorithms.

The need for careful selection of characteristics is also important in other respects. The SNRS uses knowledge gained from the social network to improve the recommendation process. This knowledge includes explicit and implicit social interactions, social influences, trust, and social behavior patterns.

Conventional SOMAR (Social Mobile Activity Recommender) is a social network-based recommender system that recommends activities based on user's social network, mobile phone data and sensor data in the ubiquitous category.

Conventional FilmTrust is a web recommender that links and combines social networks and movie ratings. In this system, the user can read the rating of the movie or movie, and record the impression. It is used to credit ratings within social networks based on assessing similarities.

GLOSS (Group Learning Sharing Own Contribution Search) is a survey system that integrates social networks and provides recommendations based on the weight of credit. This can find some similar users, modify the credit weights, and discover potentially trustworthy users using feedback mechanisms.

The conventional MyPopCorn is a Facebook film RS application method using a non-price social graph. This requires explicit feedback on the movie from the user. Its recommendations are divided into two. One is provided by a traditional user-based RS in which the neighbor is assumed to be all users in the database. The other is provided based on the neighbors' active user friends.

Conventional SNS is a system that predicts and recommends items that are not yet seen by active users but are very interesting and hard to find. It uses social network interactions and access records of items providing recommendations.

However, the above-mentioned conventional inventions have a problem in privacy preservation and scalability, and can not maintain the stability of the RS when fake information or attacks are encountered, and can not maintain the reliability of the RS based on the user's ability and behavior in a specific situation at a given time The diversity of the list can not be increased, and social network recommendation systems have come to the front of users. Therefore, the necessity of introducing a new system to solve these problems has emerged urgently.

United States Patent Publication No. 2009-0003267 U.S. Published Patent Application No. 2004-0203787 Korean Patent Publication No. 2005-0025052 Korean Patent Publication No. 2008-0054589 Korean Patent No. 1143167

[1] A. Vasilakos, Y. Zhang, and T. Spyropoulos, Delay Tolerant Networks: Protocols and Applications, Boca Raton, FL: CRC Press, [2] A. E. Al-Fagih and H. S. Hassanein, Routing Schemes for Delay-Tolerant Networks: An Applications Perspective. Technical Report 2012-588, Kingston, Canada: Telecommunications Research Lab, Queen's University, 2012. [3] I. Tumar, A. Sehgal, and J. Schonwalder, "Impact of mobility patterns on the performance of a disruption tolerant network with multi-radio energy conservation" in Proceedings of the 25th IEEE International Conference on Advanced Information Networking and Applications, Biopolis, Singapore, March 22-25, 2011, pp. 69-76. http://dx.doi.org/10.1109/AINA.2011.83 [4] C. P. Mayer, Hybrid Routing in Delay Tolerant Networks, Karlsruhe, Germany: KIT Scientific Publishing, 2012. [5] B. N. Schilit and M. M. Theimer, "Disseminating active map information to mobile hosts," IEEE Network, vol. 8, no. 5, pp. 22-32, Sep. 1994. 313011 [6] J. O. Bird, Basic Engineering Mathematics, 6th ed., New York, NY: Routledge, 2014. [7] R. A. Cabacas, H. Nakamura, and I. Ra, "Energy consumption analysis of delay tolerant network routing protocols," International Journal of Software Engineering and Its Applications, vol. 8, no. 2, pp. 1-10, 2014. http: //dx.doi.org/10.14257/ijseia.2014.8.2.01 [8] R. Cabacas and I. Ra, "Reducing message overhead using community-based message transmission for delay tolerant networks," Proceedings of the International Conference on Convergence Content, Jeju, Korea, June 26-28, 2014, pp. 61-62. [9] T. Spyropoulos, K. Psounis, and CS Raghavendra, "Spray and wait: an efficient routing scheme for intermittently connected mobile networks," Proceedings of ACM SIGCOMM Workshop on Delay-Tolerant Network- ing, August 22-26, 2005, pp. 252-259.http: //dx.doi.org/10.1145/1080139.1080143 [10] H. Karl and A. Willig, Protocols and Architectures for Wireless Sensor Networks, Hoboken, NJ: Wiley, 2005. [11] W. Li, Y. Hu, X. Fu, S. Lu, and D. Chen, "Cooperative positioning and tracking in disruption tolerant networks," IEEE Transactions on Parallel and Distributed Systems, IEEE early access articles, Mar. 2014. http://dx.doi.org/10.1109/TPDS.2014.2310471 [12] I. Joe and S. B. Kim, "A message priority routing protocol for delay tolerant networks (DTN) in disaster areas," Future Generation Information Technology. Lecture Notes in Computer Science, vol. 6485, T. H. Kim, Y. H. Lee, B. H. Kang, and D. Klezzak, Eds. Heidelberg: Springer Berlin, 2010, pp. 727-737. http://dx.doi.org/10.1007/978-3-642-17569-5 72 [13] E. C. R. de Oliveira and C. V. N. de Albuquerque, "NECTAR: a DTN routing protocol based on neighborhood contact history," Proceedings of the ACM Symposium on Applied Computing, Honolulu, HI, March 9-12, 2009, pp. 40-46. http://dx.doi.org/10.1145/1529282.1529290 [14] N. Encarnacion, "Transmission priority decision scheme based on remaining energy for body sensor networks," M.S. Thesis, Kunsan National University, Gusan, Korea, 2013. [15] A. Kerenan, J. Ott, and T. K. et al., "The ONE simulator for DTN protocol evaluation," Proceedings of the 2nd International Conference on Simulation Tools and Techniques, Rome, Italy, March 2-6, 2009, pp. 1-10.http: //dx.doi.org/10.4108/icst.simutools2009.5674 [16] A. Vahdat and D. Becker, "Epidemic routing for partiallyconnected ad hoc networks. Technical Report CS-2000-06, "Available http://www.cs.duke.edu/vahdat/ps/epidemic.pdf [17] A. Lindgren, A. Doria, and O. Schel ㅄ en, "Probabilistic routing in intermittently connected networks, "in Proceedings of the 1st International Workshop on Service Assurance with Partial and Intermittent Resources, Fortaleza, Brazil, August 1-6, 2004, pp. 239-254. http://dx.doi.org/10.1007 / 978-3-540-27767-5 24

SUMMARY OF THE INVENTION The present invention has been made in order to solve the above-mentioned problems, and it is an object of the present invention to provide a method and apparatus for using DTN based on energy based message priority and based on context information based priority message routing and essential attribute information, The ability to discover, diversity, privacy, and scalability. It can maintain the stability of the RS when fake information or attacks are encountered. Based on the user's ability and behavior in a specific situation at a given time, Priority message routing based on context information that can preserve privacy by implementing active users 'recommendation process in user' s location and situation while enhancing diversity of list, social network recommendation system based on essential attribute information, and The purpose of the method is to provide.

In addition, the present invention can induce a guess about the preferences of a new user in a social neighbors connection, and the network of friends can induce a certain interest that can be related to an active user, and direct and close friends can predict There is already sufficient social graph data to be used, and this provides a social network recommendation system and method based on context information based priority message routing and mandatory attribute information that can shorten the computerization time of recommendation There is a purpose.

In addition, the present invention provides priority message routing based on context information that can compare their distances in real time based on moving directions, speeds, and current positions of SN, IN, and DN through routing techniques through updated MCTs, A social network recommendation system based on information, and a method thereof.

In order to solve the above problem, according to the present invention, a source node (SN), an intermediate node (IN), and a destination node (DN) (x axis position), PosY (Y axis position), STime (time), ETime (elapsed time), VelX (x axis velocity), VelY A contact table (CT) for recording context information of each node during a period; And records the updated context information of the message destinations (DNSs) existing in the storage of the SN among the storage nodes, and stores the ID, the destination information including the context information, the elapsed time (ET), the node ID A message connection table (MCT) for storing a temporary list of all SN messages evaluated to be delivered to the IN; And an energy-based message prioritization module for determining a message prioritization rule according to a message prioritization rule with a message connection table (MCT). And the CT has information about how each node contacts the other nodes on the network so that the two nodes can contact each other for the first time in order to obtain information about all the message destinations the SN has in recent storage The SN and IN store new contact records in the CT, update the contact records when there is an existing record of contact between nodes, and update connection messages for all message destinations in the SN storage for MCT update. And retrieves the latest context information for the message destination from the IN's CT, De S, I, and updates the constant contact situation and context information of the node D in J, delay time of the last contact is calculated by the equation (1) below,

[Equation 1]

For the forwarder node selection, update the MCT by comparing their distances based on the moving direction, speed, and current position of SN, IN, DN.

The next step after obtaining the latest context information for the destination of all messages in the SN is to estimate the DN location by calculating the estimated location of DN at the current contact time using the following equation: .

(v = d / t where v is the speed of the object, d is the distance, and t is the time)

&Quot; (2) "

In order to determine the trajectory of the destination, the selector node of the present invention finds the trajectory using a slope-intercept form y = mx + b, Y intercept is calculated as b = y - mx, the estimated position of DN is the first coordinate, the second coordinate is the latest known position of DN, and the slope and y intercept are determined , The y-intercept of the slope and DN is calculated by the following equation (3) to determine the trajectory of the DN.

&Quot; (3) "

The sender node selection of the present invention determines the trajectory and also the estimated location of the DN to determine whether or not the IN is moving toward the location of the DN and calculates the trajectory of the IN / SN via Equation (4) and SN is determined by using four time intervals to estimate its position at a specific time, unlike the position of the DN determined using the elapsed time (ET).

&Quot; (4) "

&Quot; (5) "

In the present invention, when the trajectory of the nodes is determined, the routing technique applies the y-intercept and the slope computed in the previous step to the following equation (6) to calculate possible intersections of IN / SN and DN, And intersects with each other within a certain radius of the transfer range of the nodes so as to calculate the intersection of IN / SN and DN.

&Quot; (6) "

The sender node selection of the present invention is based on the speed of IN / SN (v = d (n)) to calculate the time required for IN / SN to reach the intersection and to estimate the location of DN by IN time to reach the intersection. / t) to determine the time interval or time required for the object to reach a particular point, and calculates the time required for each node to arrive at the intersection using the estimated IN / SN and DN calculations .

&Quot; (7) "

The sender node selection of the present invention determines an estimated position for a destination when IN arrives at the intersection point calculated in Equation 8 using the current position of the contact time of SN and IN which are estimated positions for the destination, IN < / RTI > is within the mutual delivery range that can deliver the message.

&Quot; (8) "

The sender node selection of the present invention uses the computed intersection point to determine whether a message is finally delivered from the SN to IN to calculate the distance between IN and DN when the IN intersection reaches the intersection point, / SN and determines the distance between IN / SN and DN using Pythagorean Theorem in Equation (9).

&Quot; (9) "

The sender node selection of the present invention is such that if the calculated distance of IN and DN is less than SN then the message is added to the queue of SNs to be sent to IN and the final distance DFinal, (ET) is recorded in the MCT as the best connection of each message. Since the nodes in the DTN store messages in different destinations in their buffer for a certain period of time, It is characterized by limiting the transmission of messages to the IN when it has low probability of contacting the destination.

A node with energy level of 50% or more of the present invention accepts packets of all priorities, a node with an energy level of 50% to 25% accepts packets with a medium priority or a higher priority, Receive only high packets.

In the present invention, a source node, an intermediate node (IN), a destination node (DN), a node ID (Node Id), a PosX (x axis position), PosY Including the summary vector and the routing table, including the location (Y axis position), STime (time), ETime (elapsed time), VelX (x axis velocity), VelY (y axis velocity) Records updated contact context information (CT) and updated context information on message destinations (DNSs) existing in the storage of the SN among the respective nodes, and stores the ID, the destination information including the context information, the elapsed time ET), a node ID of an IN selected as a forwarder node, a message connection table (MCT) storing a temporary list of all SN messages evaluated to be delivered to the IN, and a message prioritization rule The energy that determines priority Based message prioritization module for exchanging context information after nodes are connected and for acquiring information on all message destinations that the SN has in recent storage, When the contact is made, the SN and IN store the new contact record in the CT and update the contact record when there is an existing record in which the nodes are in contact with each other; Update the connection message for all message destinations in the SN storage, retrieve the latest context information for the message destination from the CT of IN, and update the constant contact status of nodes S, I, J and the context information of node D MCT update step to compare MCTs based on the moving direction, speed, and current position of SN, IN, and DN to determine whether the message is a reliable forwarder node that delivers the message to the destination The next step after acquiring the latest context information on the destination of all messages in the SN is a DN position estimation step of calculating an estimated position of the DN at the current contact time; In order to determine the trajectory of the destination, the trajectory using the slope-intercept form through y = mx + b is found, and the change of the value of the y coordinate with respect to the change of the value at the x- , The y intercept is calculated as b = y - mx, the estimated position of DN is the first coordinate, the second coordinate is the latest known position of DN, and the slope and DN a trajectory determination step of the DN calculating the y-intercept; In order to determine whether IN is moving towards the location of the DN, the trajectory also determines the estimated location of the DN, calculates the trajectory of the IN / SN, and, unlike the location of the DN determined using the elapsed time (ET) Determining the trajectory of the IN / SN utilizing the four time intervals to estimate its position at a particular time; Once the trajectory of the nodes is determined, the routing technique computes the y-intercept and slope computed in the previous step to calculate the intersection points of the IN / SN and DN, A step of calculating an intersection point of IN / SN and DN to determine whether or not they cross each other within a specific radius; Since the speed of IN / SN (v = d / t) is valid to calculate the time required for IN / SN to reach the intersection and estimate the position of DN by IN time to reach the intersection, Determining the time required for the interval or object to reach a particular point, and computing the time required for each node to arrive there using IN / SN and DN calculations using the estimated intersection; In the first routing step, the estimated location of the destination is determined when IN arrives at the calculated intersection, using the current location of the contact time of SN and IN, which is the estimated location for the destination, Checking whether they are within the mutual delivery range; IN When the intersection is reached, an important step in determining whether a message is finally delivered from the SN to the IN to finally calculate the distance between IN and DN is to use the calculated intersection to determine the distance between DN and IN / SN Determining the distance between IN / SN and DN using the Pythagorean Theorem; The calculated final distance DFinal for IN and SN is compared if the calculated distance of IN and DN is less than SN, then the message is added to the queue of SNs to be sent to IN; The node ID of the IN and the elapsed time ET are recorded in the MCT as the best connection of each message and the nodes in the DTN are allowed to transmit messages having different destinations to a certain period of time The present invention includes the step of restricting the transmission of the message to the IN when it has low probability of contacting the destination compared to the SN.

A node with an energy level of 50% or more accepts packets of all priorities, a node with an energy level of 50% to 25% accepts packets with a medium priority or a higher priority, Receiving only a packet with a high number of packets.

The present invention as described above can maintain the stability of the RS when fake information or attack is generated, with its robustness, trust, ability to perform unexpected discoveries, diversity, privacy preservation, and scalability. There is a remarkable effect that the privacy can be preserved by performing active recommendation process of the user in the position and situation of the user while enhancing diversity of the recommended list based on the user's ability and behavior.

In addition, the present invention can induce a guess about the preferences of a new user in a social neighbors connection, a particular interest that can be related to an active user can be induced in a network of friends, There can be enough social graph data already in place and this has a special effect that can shorten the time to computerize referrals.

Also, the present invention has an effect of comparing the distance of SN, IN, and DN based on the moving direction, speed, and current position of the SN in real time through the MCT or the like.

1 is a diagram showing the format of a contact table (CT) according to the present invention.
FIG. 2 is a view showing source codes of energy-based message prioritization rules according to the present invention.
FIG. 3 is a diagram illustrating that the S and J nodes are additionally updated when they are in contact with each other several times in different time zones according to an embodiment of the present invention.
4 is a diagram illustrating updating of the constant contact status of nodes S, I, J and the context information of node D according to an embodiment of the present invention.
5 is a view showing simulation results according to an embodiment of the present invention.
6 is a view showing simulation results according to another embodiment of the present invention.

For a better understanding of the present invention, a preferred embodiment of the present invention will be described with reference to the accompanying drawings. The embodiments of the present invention may be modified into various forms, and the scope of the present invention should not be construed as being limited to the embodiments described in detail below. The present embodiments are provided to enable those skilled in the art to more fully understand the present invention. Therefore, the shapes and the like of the elements in the drawings can be exaggeratedly expressed to emphasize a clearer description. It should be noted that in the drawings, the same members are denoted by the same reference numerals. Further, detailed descriptions of well-known functions and configurations that may be unnecessarily obscured by the gist of the present invention are omitted.

Routing component

Hereinafter, essential components of the routing technique according to the present invention will be described.

1. Contact table (CT)

As shown in FIG. 1, according to a social network recommendation system based on essential attribute information including robustness, trust, ability to perform unexpected discoveries, diversity, privacy preservation, and scalability, Message delivery that includes Node ID, PosX, PosY, STime, ETime, VelX, and VelY Interchange some important information for. The information includes the summary vector (the message list existing in the nodes storage) and the routing table of FIG. The present invention implements the format of the contact table (CT) of Figure 1 to record the context information of the nodes during the contact period.

2. Message Connection Table (MCT)

In a social network recommendation system based on mandatory attribute information according to the present invention, the MCT records updated context information for existing message destinations (DNSs) in the storage of the SN. It also shares responsibility for delivering the message to the destination (ID, context information), the elapsed time (ET), and the node ID of the IN selected as the forwarder node. For all SN messages evaluated to be delivered to the IN, the MCT stores a temporary list of such messages. And these messages will be searched for checking by energy based message priority rules.

3. Energy-based message prioritization module

As shown in FIG. 2, a node having an energy level of 50% or more with a message connection table (MCT) accepts packets of all priorities, and a node with an energy level of 50% to 25% , Less than 25% Energy-level nodes receive only high-priority packets In a social network recommendation system based on attribute information, energy-based message prioritization rules require that certain types of messages be transmitted and transmitted to the destination Decide if you have a higher chance. The determinant of the prioritization rule is the residual energy remaining in IN at contact.

A message prioritization rule prioritizes important packets in a body sensor network (BSN) for monitoring a patient. The transfer capability of IN depends on the residual energy, especially as the node energy is reduced, its communication capability is also reduced. Therefore, when a node reaches a low energy level it is consumed in energy-critical messages in applications (apps). The energy-based message prioritization rules are executed after all messages have been delivered to the IN.

Routing phase

Hereinafter, the routing step according to the present invention will be described.

The routing method proposed in the social network recommendation system based on the essential attribute information showing the stability of the RS when fake information or attack is attacked is divided into three routing steps. A CT exchange or update, an MCT update, and a choice of forwarder nodes. The code representations presented in Table 1 below are used in the routing algorithm of the proposed technique. Routing steps are covered in detail in the following subheadings.

1. CT exchange and update

In the social network recommendation system based on essential attribute information for the stability of RS when fake information or attack is attacked, the routing step starts the first step by context information exchange after nodes are connected. To preserve privacy by enforcing the active user's recommendation process within the user's location and context, CT has information about how each node contacts other nodes on the network. CMP uses CT to obtain information about all message destinations that the SN has in recent storage. When two nodes accidentally come into contact with each other for the first time, SN and IN store a new contact record in the CT, and the contact record is updated when there is an existing record in which the nodes contact each other.

3 further updates when the S and J nodes become in contact with each other several times in different time zones. This allows the node's CT to be updated in certain contact situations and thereby update the current context information for the nodes on the network.

2. MCT Update

After the context information is exchanged in the social network recommendation system based on the essential attribute information for enhancing diversity of the recommendation list list according to the present invention, the routing technique updates the connection message for all message destinations in the SN storage. This step retrieves the current context information for the message destination from the CT of IN.

FIG. 4 shows that the constant contact state of the nodes S, I, J and the context information of the node D are updated. The delay time of the last contact is calculated by the following equation (1).

3. Select the forwarder node

According to the present invention, selecting a sender node in a social network recommendation system (CMP) based on essential attribute information that can preserve privacy by performing an active user recommendation process in a user's location and situation is a key step of the present invention , It is a step of determining whether or not the trust conveying the message to the destination is a forwarding agent node. As shown in the updated MCT, the routing scheme compares their distances based on the moving direction, speed, and current position of SN, IN, DN. The concept applied here is taken from the equation of the straight line. In general, it is reasonable to assume that a straight line is the shortest path from one point to another and that nodes under normal conditions follow a straight path. The routing step is further subdivided into a series of further steps.

1) DN location estimation

The next step after obtaining the current context information of all the destinations of the message in the SN according to the social network recommendation system based on the essential attribute information of the present invention is to calculate the estimated location of DN at the current contact time.

Equation (2) shows how to calculate the estimated position of the DN.

2) Determine DN trajectory

In order to determine the trajectory of the destination according to the social network recommendation system based on the essential attribute information of the present invention, a line indicating a trajectory using a slope-intercept form (y = mx + b) is first described have.

According to the present invention, a change in the value of the y-coordinate is used for the change of the value in the x-coordinate in order to find the gradient.

Also, since the x and y positions of the node are available, the y intercept should be computed as b = y - mx.

The estimated position of DN computed from the previous step is the first coordinate and the second coordinate is the most recently known position of DN and is used to determine the slope and y intercept. Equation 3 shows the formula for calculating the slope and the y intercept of DN.

3) Determine the trajectory of IN / SN

According to the social network recommendation system based on the essential attribute information of the present invention, its trajectory also uses the same set of equations to determine the estimated location of DN, in order to determine whether IN moves towards the location of the DN .

Equations (4) and (5) show the calculation of the trajectory of IN / SN. In this case, unlike the location of the DN determined using the elapsed time (ET), four time intervals were utilized to estimate its location at a particular time. The time interval has little effect on the trajectory of the node because its important factor is its speed.

4) In / SN and DN intersection calculation

If the trajectory of the nodes is determined according to the social network recommendation system based on the essential attribute information of the present invention, the routing technique identifies possible intersections of IN / SN and DN. The y-intercept and slope calculated in the previous step to calculate the intersection point are used in Equation (6). The primary goal of this step is to determine at what particular intersection within a certain radius of the range of delivery of the nodes to each other.

5) Estimate the location of DN

The step of calculating the time required for the IN / SN to reach the intersection according to the social network recommendation system based on the essential attribute information of the present invention and estimating the location of the DN based on the IN time for reaching the intersection, IN to have the shortest time to reach the estimated intersection. This step will also prove that IN delivers the message at the optimal time.

Since the speed of IN / SN (v = d / t) is valid, the time interval or the time required for the object to reach a certain point can be determined.

The time required for each node to arrive at the intersection using the estimated IN / SN and DN calculations in the previous step is shown in Equation (7) below.

In the first routing step, the estimated position for the destination means the current position of the contact time of SN and IN. In contrast, this particular step determines the estimated location for the destination when IN arrives at the intersection calculated in equation (8). Further, the step checks whether the DN and IN are within the mutual delivery range in which the message can be delivered.

6) IN When the intersection is reached, calculate the distance between IN and DN

The social network recommendation system based on the essential attribute information of the present invention is an important step for determining whether a message is finally delivered from the SN to the IN using the intersection points calculated from the third and sixth steps, Determine the distance.

Using the Pythagorean Theorem in Equation 9, the distance between IN / SN and DN is determined when it reaches the intersection.

In the social network recommendation system based on the essential attribute information of the present invention, the calculated final distance DFinal for IN and SN is compared. If the calculated distance of IN and DN is smaller than SN, the message is sent to IN And is added to the queue of the SN to be processed. This step can be either the source of the message that evaluates other nodes in contact or the IN sharing the message delivery responsibility to the destination.

In addition, this means that while IN is going to the destination, it evaluates the other node, whether it is a suitable forwarder node, while touching another node.

The final distance (DFinal), the node ID of the IN and the elapsed time (ET) are recorded in the MCT as the best connection for each message. Since the nodes in the DTN store messages having different destinations in their buffers for a certain period of time, the present invention restricts the transmission of messages to the IN when it has a low possibility of touching a destination compared with the SN. In the routing scheme according to the present invention, only a very small number of messages are given to the IN, and in particular, the destination of the evaluated message has a possibility to contact the IN.

When all messages are evaluated and added to the SN queue, the energy based prioritization rule checks all messages and allows only message types according to energy-based prioritization rules.

4. Simulation Settings

1) ONE simulator

According to the present invention, an ONE simulator based on a social network recommendation system based on essential attribute information having robustness, trust, ability to perform unexpected discoveries, diversity, privacy preservation, , Routing simulation, visualization and reporting in one program. The present invention utilizes diffusive and predictive routing techniques available in the ONE simulator for accurate comparison.

2) Simulation setup and scenario

ONE simulator settings and scenarios are used to assess diffuse and predictive performance.

The ONE simulator scenario includes a mobile communication terminal holder in an urban area. We use the Helsinki region to imitate the same routine of moving people, cars, and trams everyday. Transportation, such as cars and trains, has movement through the map that can be installed using the map Route-File in the ONE simulator. The energy settings are listed in Table 2 and the simulation settings are described in Table 3, respectively.

The ONE simulator scenario includes a node with a random initial energy for the start of the simulation. The initial energy of the node is set within the range of 3,600 mAh to 4,800 mAh battery level. The ONE simulator scenario implies that when a mobile product uses a specific mobile service supported by DTN, the mobile product replicates the scenario in the real world, not at the same energy level. The number of newly created messages in the ONE simulator scenario is estimated to be 250.

The following Table 2 and Table 3 detail the indicators used in the performance evaluation of the routing protocol.

We define the following for simulation. Delivery ratio is the ratio of the number of messages received by DNs to the number of newly created messages.

The overhead ratio is the difference in the relayed delivery message divided by the number of delivered messages.

Average latency is a basic performance metric that refers to the average delivery time for messages from the SN to the destination.

Residual energy is the average residual energy of the nodes at the end of the simulation.

The Delivered messages per message types refers to the number of messages delivered for each message type. The metric in accordance with the present invention will prove that the messages of high importance are delivered more successfully than the messages that are not.

Simulation result

In the social network recommendation system based on the essential attribute information of the present invention, the mobile configuration consists of DTN nodes (pedestrian, car, train) made up of different kinds of different speeds, And preserving privacy by implementing active users' recommendation processes in the context of the location and situation, and showing favorable results in the stability of RS when fake information or attacks are encountered. 5 is a result of the execution of the first simulation scenario. And the success of the delivery to the CMP is 20% higher as compared to the rapidly diffusing and predictable as shown in Figure 5 (a).

The social network recommendation system based on essential attribute information of the present invention observe whether changes in speed and map-based motion affect the results of the present invention. It can be seen from FIG. 5 (b) that there is a large difference in the overhead ratio. The number of relayed messages in CMP is the lowest when the routing algorithm chooses the best forwarder node for each message. However, as compared with the prediction of FIG. 5 (c), there is a slight difference in the average delay of the CMP. It can be seen that faster delivery of messages can be achieved with various speeds of the nodes according to the present invention and privacy can be preserved by actively performing the recommendation process of users in the user's position and situation.

For the simulation scenario based on the social network recommendation system based on the essential attribute information of the present invention, all the routing protocols are completely combined with the message types to be delivered to the destination. However, CMP is not compatible with MT4 and MT5 message types Set priorities.

As shown in FIG. 5 (d), even if the queue is determined according to the message type and the size, it can be seen that the message is distributed fairly for all message types at a high energy level. Figure 5 (e) shows the average residual energy of the nodes at the end of the simulation. The residual energy of nodes using CMP is shown to be higher when compared to the other two routing schemes as shown in FIG. 5 (e).

The specific scenario of the social network recommendation system based on the essential attribute information so as to trust the user based on the user's ability and behavior in a specific situation at a given time from the viewpoint of random start energy is that at the beginning of the day, It is imitating the situation of the application platform. Users have a variety of battery levels and are using application services.

From the simulation results shown in Figure 6, CMP still outperforms diffusion and prediction in terms of message delivery and energy efficiency. Along with all other simulation scenarios, the delivery message shown in Figure 6 (a) reaches 60% when the number of newly created messages is 250.

In Figure 6 (c), the average latency latency is shown to be higher when compared to the previous scenario because of the unacceptance of other messages for delivery due to low energy, It is a direct transaction (exchange) with a low overhead rate. The number of selected sender nodes is smaller and therefore results in less message delivery and delay. The results shown in Figure 6 (d) illustrate the average of CMP in transmitting an MT3 message to an MT5 message.

Due to the random start energy level between nodes in order to conserve privacy by enforcing the active user's recommendation process within the user's location and context, a greater number of important messages are already given priority as soon as the energy level of the node reaches 50% to 25% They are ranked. This also shows that the number of dead nodes at the end of the simulation is higher, as shown in Figure 6 (e).

Thus, in the present invention (CMP), we introduced an efficient context-based routing scheme with an energy-based priority message for DTNs. The present invention consists of several elements that perform efficient routing of messages to reach their destination. It involves using other nodes as anchor nodes with contact records for the message destination.

The evaluation of the DN position and the measurement of the likelihood of IN to contact can be accomplished largely using the formula of the line equation. Energy based message delivery rules have been implemented to efficiently use the energy of nodes in accepting messages during low energy levels.

Implementation of the CMP approach in a social network recommendation system based on mandatory attribute information has been compared with diffusive and predictive extension of the two existing routing protocols by the ONE simulator. A first scenario according to the present invention is to set up a mobile communication product having various characteristics in an urban environment. The second scenario includes a random starting energy level that demonstrates the effectiveness of energy efficiency and message priority rules. Most results show better efficacy on all aspects (especially delivery and delivery costs). However, the proposed scheme has buffering time in the number of significant messages exchanged or low overhead for any transaction such as slightly higher latency. These transactions are inevitably well balanced when compared to other routing techniques.

In order to validate the effect of the approach according to the invention, more intensive simulations require contact time, contact frequency, and rate of change of nodes. Comparisons with conventional context-based routing schemes similar to the proposed approach will give insight into the shortcomings and limitations. Finally, other contextual information should be reviewed to improve the routing implementation of the proposed approach.

The present invention has industrially compared and evaluated the SNRS (Social Network Recommendation System) based on a series of characteristics. For example, the characteristics of how robustness, trust, ability to make unexpected discoveries, diversity, privacy preservation, scalability, and how these affect the success of referrals in social network recommendation systems based on essential attribute information And how it has increased industrial applicability in the social network referral system industry.

a. Toughness

And the stability of the Recommendation System (RS) when fake information or an attack occurs in a social network recommendation system based on essential attribute information according to the present invention. This attack usually consists of a form of profile injection designed to raise the value of any one of several items. Robustness measures the behavior of the system before and after attacks to determine how it affects social network recommendation systems, which are typically based on mandatory attribute information.

The present invention has been tested to determine the effect of attacking the model in the CF algorithm. The average prediction shift is one of the methods commonly used to evaluate the robustness of RS.

The social network recommendation system based on the essential attribute information according to the present invention is likely to be industrially applicable by explaining changes in predicted evaluation of an item before and after an attack on a general average prediction or a prediction that is an attack target .

b. responsibility

The reliability of the social network recommendation system based on the essential attribute information according to the present invention measures the spontaneity that trusts the user based on the user's ability and behavior in a specific situation at a given time. People are generally willing to keep a mental map of their trust level for their friends' advice. The work in the social network recommendation system based on the essential attribute information uses the credit rating in the social network based on the estimation of similarity. It is based on the concept that a correlation exists between credit and user similarity. The present invention demonstrates this correlation in observational studies of actual online communities.

In a social network recommendation system based on essential attribute information, work in SOMAR or SNS does not include reliability among users in a social network, rather than importance of social network interactions in the recommendation process.

More specifically, the social network component of FilmTrust requires users to perform a trustworthiness assessment of each person added as a friend. With trust values collected from social network recommendation systems based on mandatory attribute information, people use TidalTrust, a credit network estimation algorithm, as a basis for personal estimation estimation for each user. The accuracy of the evaluation recommended in the experiment is better between the simple average evaluation and the evaluation by the general RS algorithm.

c. Unexpectedly lucky find (Serendipity)

An unexpected luck finding in a social network recommendation system based on mandatory attribute information according to the present invention is a measure of how surprised it is for successful recommendation. It is the amount of such information that is classified as new or basically unclear to the user in the referral. Some work on recommendation for unexpected good discovery shows that such lucky finds are more in the referral list for other items in other categories than listings for similar items. In addition, the present invention proposes a recommendation method for enhancing diversity of a list of recommended lists in a social network recommendation system based on essential attribute information.

d. Diversity

In the social network recommendation system based on the essential attribute information according to the present invention, diversity is the quality of the result list that helps solve ambiguity well. Diversity generally applies to a series of items related to how different items are to each other.

e. Privacy Preservation

Privacy in a social network recommendation system based on essential attribute information according to the present invention is a very important issue to a user. Users are reluctant to provide personal details because they are afraid of being misused. RS administrators are concerned with legal issues associated with protecting user privacy.

The use of OSN data will preserve the user's privacy. A social network recommendation system based on essential attribute information preserves privacy by performing actively recommendation process in the user 's location and situation. The two elements of transparency and anonymity help to preserve privacy in SNRS, leaving no trace of delivering valuable data to third parties within the mobile phone category.

In the social network recommendation system based on the essential attribute information, the users themselves question the basis of the recommendation. They appreciate that they better accept referrals if they understand what items are presented to them. The present invention has evaluated the role of transparency in the accuracy of social network recommendation systems based on essential attribute information. They explain that RSs are largely unused for high-risk decisions due to a lack of transparency. The availability of transparency is based on the area or function of the RS.

The transparency according to the present invention is very helpful to RS in recommending travel plans, investment, real estate and the like. However, most SNRSs act like a series of black boxes, as mentioned above. It does not allow the user to know how it was done without leaving a recommendation process in the social network recommendation system based on the essential attribute information. These SNRSs are primarily low-level area types and are prone to lack of transparency. However, MyPopCorn users can see how the proposed entry is derived from user preferences or similarities with other users and friends.

Anonymity is not anonymous in the OSN according to the social network recommendation system based on essential attribute information, especially in SNRS where user profile is stored or used as speculation. Anonymity is related to transparency. If the user can see how the recommendation was evaluated, it can be seen that the utilization of the user data has been made. Referral-related data may be exploited maliciously. Once a recommendation is given to a user, a social network recommendation system based on essential attribute information will allow friends to see similarities. The user data obtained from the reasoning process in the recommender system solves the problem that may be used by criminals such as harassment, house invasion, and blindness, so that privacy can be preserved by performing active user recommendation process within the user's location and situation can do.

f. Scalability

Due to the exponentially increasing number of users and items in social network recommendation systems based on mandatory attribute information, SNRS and RS suffer serious scalability problems. Social network users typically have hundreds and thousands of friends who computerize complex similarities. A social network recommendation system based on mandatory attribute information indicates that users are connected to other users but they do not always interact with each other in the same way. Users merely interact with friends in the small group of friends who are generally closest in the social network structure. SOMAR, GLOSS and MyPopCorn describe this connection in the social graph.

A social network recommendation system based on essential attribute information states that similarities among friends are on average higher than similarities between non-connected users. This can focus on the social graph as a representative relationship instead of the user's alternative social network structure. Narrowing the datasets used by SOMAR and MyPopCorn can solve scalability problems. Further, the present invention is useful for using friends who are away from the social network, focusing on instant friends based on the social network recommendation system based on essential attribute information.

g. According to the present invention, in the social network recommendation system based on the essential attribute information,

Social Network Recommendation System Based on Essential Attribute Information In the industry point of view and future challenges, considering the memory-based approach through simulation, the relative strength and stability of the model-based CF algorithm applied to the present invention have.

The social network recommendation system based on essential attribute information according to the present invention can be a basis for using model-based algorithms in consideration of SNRS, focusing on robustness. Robustness should be appreciated, especially for profile input on OSN data. Certain methods must be taken to ensure the reliability of the data used. And no bias should be added because it may affect the outcome of the prediction.

The use of important credit in the structure of the social network recommendation system based on the essential attribute information of the present invention is very necessary to improve the recommendation process. However, the SNRS according to the present invention rates the credit from the frequency of social network interactions.

And, in the past, credit was assessed without considering most of the social contexts (eg similarity of preferences, proximity of locations, community influences and reputation), but these factors are very important in improving the use of trust in SNRS.

Conventional FilmTrust also only integrates the use of real-time interactive exchanges that can be obtained from OSNs, and FilmTrust and SNS can also use social graphs as an advantage to improve their recommendation process for social graphs, It was difficult to preserve privacy by implementing an active user recommendation process.

However, the unexpected discovery possibility or the diversity of recommendation of the social network recommendation system based on the essential attribute information according to the present invention changes with the area of the RS. The possibility of unexpected discovery in certain areas such as academic research, such as GLOSS, is less important when considering the similarity or predictability of items. However, it is highly desirable for users to recommend entertainment such as movie recommendation or unexpected goodness. The dissimilarity of items can be quite helpful to some SNRSs, not to others. RSs attempting to improve items or events, such as products or films, tend to lack diversity.

The privacy preservation according to the present invention can be applied in a series of evaluated SNRSs. In addition to evaluating predictions, transparency raises both advantages and disadvantages, and transparency can be advantageously used when transparency has a favorable impact on accommodating user trust and referrals.

The present invention implicitly includes user information with recommendations, along with transparency of the data. Transparency and anonymity should be addressed by explicitly mentioning the privacy choices that users have as soon as they use RS. The user must know what data is shared or not, and the present invention is fully supported by the industry.

The social neighbors connection can lead to speculation about the preferences of the new user, and the network of friends can induce a certain interest that can be related to the active user. Direct and close friends can already have enough social graph data to use for prediction, which can shorten the time to computerize referrals. This also helps address scalability issues.

Therefore, the present invention is based on the priority information message routing based on the context information, so that robustness, trust, unexpected good discoverability, diversity, privacy preservation, and scalability in terms of the source and technology of the social network recommendation system based on essential attribute information , And this essential characteristic is as important as the accuracy of the prediction. However, most characteristics are not related to the evaluated SNRS.

In accordance with the present invention, there is clearly an exchange generated by prioritizing one property over the other, and combining all the properties may be effective for industrial development related to the social network recommendation system.

Claims

(Node ID), PosX (x-axis position), PosY (Y), and Node (ID) exist in the storage of each node for message delivery in a certain contact situation. Axis position), STime (time), ETime (elapsed time), VelX (x axis velocity), VelY (y axis velocity), and records the context information of each node during the contact period A contact table (CT);
And records the updated context information of the message destinations (DNSs) existing in the storage of the SN among the storage nodes, and stores the ID, the destination information including the context information, the elapsed time (ET), the node ID A message connection table (MCT) for storing a temporary list of all SN messages evaluated to be delivered to the IN; And
And an energy based message prioritization module for determining a message priority according to a message prioritization rule with a message connection table (MCT)
CT exchanges context information after nodes are connected for CT exchange and update, and CT has information about each node in contact with other nodes on the network, so that information about all message destinations the SN has in recent storage SN < / RTI > and IN store a new contact record in the CT, and the contact record is updated when there is an existing record in which the nodes are in contact with each other,
I update the connection message for all message destinations in the SN storage for MCT update and retrieve the latest context information for the message destination from the CT of IN to determine the constant contact situation of nodes S, I, J and the context of node D Information, and the delay time of the last contact is calculated by the following equation (1)
[Equation 1]

Based on the context information based priority message routing and mandatory attribute information, characterized by updating the MCT by comparing their distances based on the moving direction, speed, and current position of SN, IN, DN for forwarder node selection A social network referral system.

The method according to claim 1,
Forwarder node selection,
The next step after obtaining the latest context information for the destination of all messages at the SN further comprises estimating the DN location by calculating the estimated location of the DN at the current contact time with: < EMI ID = Based Priority Message Routing and Social Network Recommendation System based on Essential Attribute Information.
(v = d / t where v is the speed of the object, d is the distance, and t is the time)
&Quot; (2) "

The method according to claim 1,
Forwarder node selection,
In order to determine the trajectory of the destination, we use the slope-intercept form y = mx + b to find the trajectory and change the value of the y coordinate with respect to the change of the value in the x- , The y intercept is calculated as b = y - mx, the estimated position of DN is the first coordinate, the second coordinate is the latest known position of DN, and to determine the slope and y intercept, Wherein the trajectory of the DN is calculated by calculating the y-intercept of the slope and the DN, and a social network recommendation system based on the priority message routing and essential attribute information based on the context information.
&Quot; (3) "

The method according to claim 1,
Forwarder node selection,
To determine whether IN is moving towards the location of the DN, the trajectory also determines the estimated location of the DN,
Unlike the location of the DN determined by using the elapsed time (ET), it is possible to calculate the IN / SN trajectory through the following equations (4) and (5) And determines the trajectory of the IN / SN based on context information based priority message routing and essential attribute information.
&Quot; (4) "

&Quot; (5) "

The method according to claim 1,
Forwarder node selection,
When the trajectory of the nodes is determined, the routing technique applies the y intercept and the slope calculated in the previous step to the following equation (6) to calculate the possible intersections of IN / SN and DN to calculate a certain radius And intersection of IN / SN and DN is determined by determining whether the intersection of the IN / SN and DN intersects with each other. The priority information routing based on the context information and the social network recommendation system based on the essential attribute information.
&Quot; (6) "

The method according to claim 1,
Forwarder node selection,
To calculate the time required for IN / SN to reach the intersection and to estimate the position of DN by the IN time to reach the intersection, use the IN / SN speed (v = d / t) The time required to reach this particular point is determined and the time required for each node to arrive there using IN / SN and DN computation's estimated intersection points is calculated by Equation (7) A social network recommendation system based on priority message routing and mandatory attribute information.
&Quot; (7) "

The method according to claim 1,
Forwarder node selection,
Using the current position of the contact time of SN and IN, which is the estimated position for the destination, determine the estimated position for the destination when IN reaches the intersection point calculated in Equation (8), and if DN and IN are mutual Based on the priority information and the attribute information of the social network.
&Quot; (8) "

The method according to claim 1,
Forwarder node selection,
IN To arrive at the intersection, calculate the distance between IN and DN,
Finally, the distance between the DN and the IN / SN is determined using the calculated intersection point which determines whether the message is delivered from the SN to the IN, and the IN / SN and the IN / SN are determined using the Pythagorean Theorem And a DN is determined based on context information based on priority message routing and essential attribute information.
&Quot; (9) "

The method according to claim 1,
Forwarder node selection,
If the calculated distance of IN and DN is less than SN, the message is added to the queue of SNs to be sent to the IN,
The final distance (DFinal), the node ID of the IN and the elapsed time (ET) are recorded in the MCT as the best connection of each message, and in the DTN, Based on priority information message routing and mandatory attribute information, characterized by restricting the transmission of messages to the IN when it has a low possibility of contacting the destination compared with the SN because it is stored in their buffer. Social Network Referral System.

The method according to claim 1,
A node with energy level of 50% or more accepts packets of all priorities, a node with energy level of 50% to 25% accepts packets with middle or higher priority, and nodes with energy level of less than 25% A priority message routing based on context information, and a social network recommendation system based on essential attribute information.

(Node ID), PosX (x-axis position), PosY (Y-axis) in the storage of nodes for message delivery, Which includes the summary vector and the routing table including the location information (location), STime (time), ETime (elapsed time), VelX (x axis velocity), VelY (y axis velocity) (CT) records the updated context information for the message destinations (DNSs) existing in the storage of the SNs among the nodes (SN), and stores the ID, the destination information including the context information, the elapsed time (ET) The node ID of the IN selected as the forwarder node is included and the priority of the message is determined according to the message connection table (MCT) storing the temporary list for all SN messages evaluated to be transmitted to the IN and the message prioritization rule Energy based decision making In the method not using the first system consisting of a screen module,
When two nodes inadvertently come into contact with each other in order to exchange context information after nodes are connected and acquire information about all message destinations that SN has in recent storage, SN and IN store new contact records in CT And a CT exchange and updating step for causing the contact record to be updated when there is an existing record in which nodes are in contact with each other;
Update the connection message for all message destinations in the SN storage, retrieve the latest context information for the message destination from the CT of IN, and update the constant contact status of nodes S, I, J and the context information of node D An MCT updating step,
The sender node selection step of updating the MCT by comparing their distances based on the moving direction, speed, and current position of SN, IN, DN, to determine whether the message is a trusting forwarder node that delivers to the destination,
The next step after obtaining the most recent context information for the destination of all messages at the SN is the DN location estimate step of calculating the estimated location of the DN at the current contact time;
In order to determine the trajectory of the destination, the trajectory using the slope-intercept form through y = mx + b is found, and the change of the value of the y coordinate with respect to the change of the value at the x- , The y intercept is calculated as b = y - mx, the estimated position of DN is the first coordinate, the second coordinate is the latest known position of DN, and the slope and DN a trajectory determination step of the DN calculating the y-intercept;
In order to determine whether IN is moving towards the location of the DN, the trajectory also determines the estimated location of the DN, calculates the trajectory of the IN / SN, and, unlike the location of the DN determined using the elapsed time (ET) Determining the trajectory of the IN / SN utilizing the four time intervals to estimate its position at a particular time;
Once the trajectory of the nodes is determined, the routing technique computes the y-intercept and slope computed in the previous step to calculate the intersection points of the IN / SN and DN, A step of calculating an intersection point of IN / SN and DN to determine whether or not they cross each other within a specific radius;
Since the speed of IN / SN (v = d / t) is valid to calculate the time required for IN / SN to reach the intersection and estimate the position of DN by IN time to reach the intersection, Determining the time required for the interval or object to reach a particular point, and computing the time required for each node to arrive there using IN / SN and DN calculations using the estimated intersection;
In the first routing step, the estimated location of the destination is determined when IN arrives at the calculated intersection, using the current location of the contact time of SN and IN, which is the estimated location for the destination, Checking whether they are within the mutual delivery range;
IN When the intersection is reached, an important step in determining whether a message is finally delivered from the SN to the IN to finally calculate the distance between IN and DN is to use the calculated intersection to determine the distance between DN and IN / SN Determining the distance between IN / SN and DN using the Pythagorean Theorem;
The calculated final distance DFinal for IN and SN is compared if the calculated distance of IN and DN is less than SN, then the message is added to the queue of SNs to be sent to IN; And
The final distance (DFinal), the node ID of the IN and the elapsed time (ET) are recorded in the MCT as the best connection of each message, and in the DTN, The present invention includes limiting the transmission of a message to the IN when it has a low possibility of making contact with a destination compared with an SN, Based social network recommendation method.

12. The method of claim 11,
A node with energy level of 50% or more accepts packets of all priorities, a node with energy level of 50% to 25% accepts packets with middle or higher priority, and nodes with energy level of less than 25% Wherein the step of receiving a priority message further includes the step of receiving only the context information based on the context information.