Content distribution method based on small data packets

Abstract

The invention provides a content distribution method based on small data packets, and relates to the technical field of wireless communication. Comprising the following steps: dividing a plurality of mobile nodes into a plurality of node sets in a distributed manner; updating the plurality of node sets in real time by using a node set updating formula to obtain a plurality of updated node sets; setting a reward function in each node set based on the plurality of updated node sets; determining a real-time reward value of the corresponding node according to the reward function; updating the Q value of the corresponding node in real time according to the real-time reward value to obtain an updated Q value; obtaining an optimized relay node by using the updated Q value; and transmitting the to-be-propagated information by using the optimized relay node until the to-be-propagated information is transmitted to the target node, and if the to-be-propagated information is not propagated to the target node, returning to the Q value updating step. According to the invention, the problem that efficient opportunity network routing cannot be realized due to insufficient perception of context information of nodes and messages by a traditional algorithm in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to a content distribution method based on small data packets. Background Art

[0002] With the rapid development of modern science and technology, wireless communication technology has made significant progress in the past few decades and has become one of the most mature technologies in the world. The TCP / IP protocol provides a variety of functions for many application services based on it, including but not limited to email, file transfer, information sharing, etc. To ensure the stable operation of the TCP / IP protocol, a stable end-to-end link must be established between the source node and the destination node. However, due to the constraints of factors such as node energy and geographical location, it is not easy to maintain a stable end-to-end link in the actual environment.

[0003] In order to deal with the problem of communicating in a network that lacks stable end-to-end links, scholars have proposed a new network paradigm, namely delay-tolerant networks. Its application scenarios have expanded from space to the ground, ocean, and even the seabed, including wild animal tracking, post-disaster rescue, remote area communications, social mobile networks, etc. In these scenarios, the network topology changes frequently and there is a lack of pre-existing infrastructure, so end-to-end paths are relatively scarce. As a derivative form of delay-tolerant networks, opportunistic networks are networks that use the opportunity of node encounters to exchange information.

[0004] Routing is an essential basic function of opportunistic networks and a hot topic in research. Efficient routing algorithms can reduce communication delays, reduce energy loss, and improve data delivery efficiency. The core of opportunistic network routing is to select appropriate relay nodes for message forwarding and transmission. Traditional opportunistic network routing cannot reasonably utilize the context information of nodes in the network and cannot accurately select appropriate relay nodes. In recent years, with the development of artificial intelligence related technologies, it is combined with opportunistic network routing to achieve more accurate context information perception and more efficient message routing.

[0005] Content distribution is an important application of message communication in opportunistic networks. According to the size of the content to be distributed, it is mainly divided into small data packets and large data packets. The existing opportunistic network routing methods have the following main shortcomings:

[0006] For the transmission of small data packets, traditional algorithms have insufficient awareness of the contextual information of nodes and messages and cannot achieve efficient opportunistic network routing. Summary of the invention

[0007] In order to overcome the deficiencies of the prior art, the purpose of the present invention is to provide a content distribution method based on small data packets. The present invention solves the problem that traditional algorithms in the prior art are insufficiently aware of the context information of nodes and messages and cannot achieve efficient opportunistic network routing.

[0008] To achieve the above object, the present invention provides the following solutions:

[0009] A content distribution method based on small data packets, comprising:

[0010] Distributively dividing multiple mobile nodes into multiple node sets;

[0011] Using a node set update formula to update the multiple node sets in real time to obtain multiple updated node sets;

[0012] Based on the updated plurality of node sets, a reward function is set in each node set, wherein the reward function includes a transmission reward function and an encounter reward function;

[0013] Determine the real-time reward value of the corresponding node according to the reward function;

[0014] The Q value of the corresponding node is updated in real time according to the real-time reward value to obtain an updated Q value, wherein the corresponding node is a source node of current information and a current relay node;

[0015] Use the updated Q value to obtain the optimized relay node;

[0016] The optimized relay node is used to transmit the information to be transmitted until the information to be transmitted is transmitted to the target node. If the information is not transmitted to the target node, the process returns to the step of "updating the Q value of the corresponding node in real time according to the real-time reward value to obtain an updated Q value".

[0017] Preferably, the distributed division of the plurality of mobile nodes into a plurality of node sets comprises:

[0018] Obtain the movement rules of each mobile node and the historical encounter data of each mobile node;

[0019] The multiple mobile nodes are divided into multiple node sets according to the movement rules of the various mobile nodes and the historical encounter data of the various mobile nodes.

[0020] Preferably, the expression of the node set is:

[0021]

[0022] Among them, the SCS iMS(j) is a node set, MS(i) is the encounter set of node i, |MS(i)| is the cardinality of the set, and MS(i)∩MS(j) is the intersection of the encounter sets of nodes i and j.

[0023] Preferably, the expression of the node set update formula is:

[0024]

[0025] The SC i is the updated node set.

[0026] Preferably, updating the Q value of the corresponding node in real time according to the real-time reward value to obtain an updated Q value includes:

[0027] If a node encounter occurs at the source node and no information transmission occurs, determining a first reward value according to an encounter reward function;

[0028] If the source node encounters a node and information transmission occurs, determining a second reward value according to the transmission reward function;

[0029] The Q value of the corresponding node is updated in real time according to the first reward value and the second reward value to obtain an updated Q value.

[0030] Preferably, the expression of the encounter reward function is:

[0031]

[0032] Among them, R c (d,i) is the reward generated by the message with the target node d when the current node is connected to node i, R c (d,i) is the reward generated by the message with the target node d when the current node is connected to node i, g is a constant greater than 0, B size (i) is the cache size of node i, B occupied (i) refers to the size of the current cache area.

[0033] Preferably, the expression of the transmission reward function is:

[0034]

[0035] Among them, R t (d,i) is the transmission reward function, and hop refers to the number of hops the message has experienced so far.

[0036] The present invention discloses the following technical effects:

[0037] The present invention provides a content distribution method based on small data packets, comprising: dividing multiple mobile nodes into multiple node sets in a distributed manner; updating the multiple node sets in real time using a node set update formula to obtain multiple updated node sets; setting a reward function in each node set based on the updated multiple node sets, the reward function including a transmission reward function and an encounter reward function; determining a real-time reward value of a corresponding node according to the reward function; updating the Q value of the corresponding node in real time according to the real-time reward value to obtain an updated Q value, wherein the corresponding node is a source node of current information and a current relay node; obtaining an optimized relay node using the updated Q value; transmitting the information to be transmitted using the optimized relay node until the information to be transmitted is transmitted to a target node, and if the information is not transmitted to the target node, returning to the step of "updating the Q value of the corresponding node in real time according to the real-time reward value to obtain an updated Q value". The present invention introduces Q learning into the routing of small data packets, realizes more accurate and intelligent relay node selection, and realizes the goals of directional spraying, adaptive initial copy number, etc., improves the delivery success rate of small data packets, and reduces network load at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0039] Figure 1 A flow chart of a content distribution method based on small data packets provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0042] like Figure 1 As shown, the present invention provides a content distribution method based on small data packets, comprising:

[0043] Step 100: Distributively divide a plurality of mobile nodes into a plurality of node sets;

[0044] Step 200: using a node set update formula to update the multiple node sets in real time to obtain multiple updated node sets;

[0045] Step 300: Based on the updated plurality of node sets, a reward function is set in each node set, wherein the reward function includes a transmission reward function and an encounter reward function;

[0046] Step 400: Determine the real-time reward value of the corresponding node according to the reward function;

[0047] Step 500: updating the Q value of the corresponding node in real time according to the real-time reward value to obtain an updated Q value, wherein the corresponding node is a source node of current information and a current relay node;

[0048] Step 600: Use the updated Q value to obtain the optimized relay node; when the corresponding Q value of the encountered node is greater than its own, it is selected as the relay node to achieve directional propagation of the message.

[0049] Step 700: Use the optimized relay node to transmit the information to be transmitted until the information to be transmitted is transmitted to the target node. If it is not transmitted to the target node, return to the step of "updating the Q value of the corresponding node in real time according to the real-time reward value to obtain an updated Q value".

[0050] Furthermore, the distributed division of the plurality of mobile nodes into a plurality of node sets includes:

[0051] Obtain the movement rules of each mobile node and the historical encounter data of each mobile node;

[0052] The multiple mobile nodes are divided into multiple node sets according to the movement rules of the various mobile nodes and the historical encounter data of the various mobile nodes.

[0053] Specifically, since most mobile nodes in opportunistic networks are vehicle-mounted devices or handheld devices, such as mobile phones and tablets, the movement of nodes conforms to the movement of vehicles or pedestrians. It is not completely random, but has certain regularity and periodicity. In addition, nodes also have social attributes. Nodes belonging to the same "community" geographically may meet frequently. Messages are transmitted between nodes belonging to the same "community", and the probability of successful delivery is greater. In addition to dividing communities geographically in advance, the "social" attributes of nodes can also be characterized based on the movement patterns of mobile nodes or the history of their encounters.

[0054] Furthermore, we define social circle similarity to indicate the similarity of the encounter sets of two nodes. If two nodes have encountered many identical nodes, that is, have many "common friends", then the social circle similarity between them is considered to be very high. The specific definition method (node ​​set expression) is as follows:

[0055]

[0056] Among them, the SCS i (j) is the node set, MS(i) refers to the encounter set of node i, that is, the nodes that node i has met before. |MS(i)| refers to the cardinality of the set, that is, the size of the set. MS(i)∩MS(j) refers to the intersection of the encounter sets of nodes i and j, that is, the nodes that both nodes have met before. The more similar the social circles of two nodes are, the more likely they are to successfully deliver messages.

[0057] Furthermore, the expression of the node set update formula (the specific method of updating the social circle of a node) is:

[0058]

[0059] The SC i is the updated node set. When node i meets node j, if SCS i (j) is greater than the threshold, node j will be added to the social circle of node i. Threshold is a number between 0 and 1, which can be adjusted according to the scenario requirements. Each node creates an empty social circle at the beginning of the simulation, and continuously updates its social circle according to the above method during the simulation. The node social circle attribute is taken into account when setting the reward function of the ANSAW-Q algorithm and the number of initial copies of the message.

[0060] Further, the updating of the Q value of the corresponding node in real time according to the real-time reward value to obtain an updated Q value includes:

[0061] If a node encounter occurs at the source node and no information transmission occurs, determining a first reward value according to an encounter reward function;

[0062] If the source node encounters a node and information transmission occurs, determining a second reward value according to the transmission reward function;

[0063] The Q value of the corresponding node is updated in real time according to the first reward value and the second reward value to obtain an updated Q value. During the entire routing process, the node updates the corresponding Q value in a distributed manner using the two reward values ​​and stores it in its own Q table.

[0064] Specifically, there are two ways to trigger the Q value update, one is to establish a connection between two nodes, and the other is to complete the transmission of messages between two nodes.

[0065] When two nodes meet and establish a connection, the nodes will update their Q values ​​according to the set connection reward function. This is because nodes that meet frequently have more transmission opportunities and are more likely to deliver messages to the target node. Reasonable updates to the corresponding Q values ​​when nodes meet can reflect this social relationship to a certain extent, so that relay nodes can be selected more intelligently when the routing algorithm performs directional spraying.

[0066] Furthermore, when nodes meet, the node Q value update expression is:

[0067]

[0068] The above formula indicates that the Q value is updated when two nodes meet, where Q S (d,i) represents the Q value corresponding to transmitting the message of target node d to the encounter node i of S from the perspective of the current node S. α represents the learning rate, which is a parameter in Q learning. γ is the discount factor, which indicates the importance of future rewards. In reinforcement learning, the weights of current rewards and rewards obtained by future actions are different. Therefore, the discount factor γ between 0 and 1 is used to adjust the weight of future rewards. The larger the value of γ, the more importance is attached to future rewards. When γ = 1, future rewards are equally important as current rewards. R c (d,i) is the connection reward function. It refers to the maximum Q value obtained by transmitting a message with the target node d to all its neighboring nodes from the perspective of node i. i Represents the set of neighbor nodes of node i.

[0069] When two nodes complete the transmission of a message, the node will update its Q value according to the transmission reward function. In the current state, performing an action (i.e. selecting a relay node for message transmission) will get a reward, and the Q value will be updated according to this reward and state transfer. The expression for node Q value update during node transmission is:

[0070]

[0071] The above formula indicates that the Q value is updated after the message is successfully transmitted between two nodes. The only difference from the previous formula is that R t (d,i), which refers to the transfer reward function.

[0072] Specifically, in reinforcement learning, the setting of rewards is crucial. Specifically, the value of reward R directly affects the update of Q value, thereby affecting the routing decision.

[0073] Furthermore, the expression of the encounter reward function is:

[0074]

[0075] Among them, R c (d,i) is the reward generated by the message with the target node d when the current node is connected to node i. CT(i) refers to the number of times the current node encounters node i, and β is a weight parameter between 0 and 1, which can be used to adjust the relative weight of the encounter frequency and the cache occupancy. size (i) is the cache size of node i, B occupied (i) refers to the size of the current cache area. g is a constant greater than 0. Its function is to generate an additional reward when the social circle of node i contains the message target node d. The size of g can be adjusted within a reasonable range, but it is necessary to ensure that R c (d,i) cannot be greater than the reward the node receives when it encounters the target node.

[0076] Furthermore, the expression of the transmission reward function is:

[0077]

[0078] Among them, R t (d,i) is the transmission reward function. When the message is transmitted, the reward is obtained. Hop refers to the number of hops that the message has experienced so far. The larger the hop, the more times the message is forwarded and copied, and the more chances it has to be delivered to the target node. We hope to give priority to transmitting "new" messages with fewer hops, so when setting the transmission reward function, the larger the hop, the smaller the reward value.

[0079] When the encounter node i happens to be the target node of the current message, after the transmission is successful, a reward value of 100 will be obtained, which means successful delivery and a considerable reward. It can also be set to other values, as long as the reward obtained after successful delivery to the target node is much greater than the reward obtained in other cases. If node i is not the target node of the message, the reward value will be determined by combining the number of hops of the message and the connection reward function between nodes.

[0080] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0081] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A content distribution method based on small data packets, characterized in that: include: Distributively dividing multiple mobile nodes into multiple node sets; Using a node set update formula to update the multiple node sets in real time to obtain multiple updated node sets; Based on the updated plurality of node sets, a reward function is set in each node set, wherein the reward function includes a transmission reward function and an encounter reward function; Determine the real-time reward value of the corresponding node according to the reward function; The Q value of the corresponding node is updated in real time according to the real-time reward value to obtain an updated Q value, wherein the corresponding node is a source node of current information and a current relay node; Use the updated Q value to obtain the optimized relay node; The optimized relay node is used to transmit the information to be transmitted until the information to be transmitted is transmitted to the target node. If the information is not transmitted to the target node, the process returns to the step of "updating the Q value of the corresponding node in real time according to the real-time reward value to obtain an updated Q value".

2. A content distribution method based on small data packets according to claim 1, characterized in that: The distributed division of the plurality of mobile nodes into a plurality of node sets comprises: Obtain the movement rules of each mobile node and the historical encounter data of each mobile node; The multiple mobile nodes are divided into multiple node sets according to the movement rules of the various mobile nodes and the historical encounter data of the various mobile nodes.

3. The method for distributing content based on small data packets according to claim 1, characterized in that: The expression of the node set is: Among them, the SCS i MS(j) is a node set, MS(i) is the encounter set of node i, |MS(i)| is the cardinality of the set, and MS(i)∩MS(j) is the intersection of the encounter sets of nodes i and j.

4. The method for distributing content based on small data packets according to claim 2, characterized in that: The expression of the node set update formula is: The SC i is the updated node set.

5. The method for distributing content based on small data packets according to claim 2, characterized in that: The step of updating the Q value of the corresponding node in real time according to the real-time reward value to obtain an updated Q value includes: If a node encounter occurs at the source node and no information transmission occurs, determining a first reward value according to an encounter reward function; If the source node encounters a node and information transmission occurs, determining a second reward value according to the transmission reward function; The Q value of the corresponding node is updated in real time according to the first reward value and the second reward value to obtain an updated Q value.

6. A method for distributing content based on small data packets according to claim 5, characterized in that: The expression of the encounter reward function is: Among them, R c (d, i) is the reward generated by the message with the target node d when the current node is connected to node i, R c (d, i) is the reward generated by the message with the target node d when the current node is connected to node i, g is a constant greater than 0, B size (i) is the cache size of node i, B occupied (i) refers to the size of the current cache area.

7. The method for distributing content based on small data packets according to claim 5, characterized in that: The expression of the transfer reward function is: Among them, R t (d, i) is the transmission reward function, and hop refers to the number of hops the message has experienced so far.

Images