Communication management device, communication management method, and communication management system

Abstract

The aim is to improve resource utilization efficiency across the network. [Solution] The communication management device 1 includes a calculation unit 11 that calculates entropy, which is an expected value of the amount of information obtained when communication occurs by each of the multiple communication terminals 2, from the probability of communication occurrence for each of the multiple communication terminals 2 calculated based on the communication history of the multiple communication terminals 2; an allocation unit 13 that allocates wireless resources to each of the multiple communication terminals 2 based on the code length of the code determined for each of the multiple communication terminals 2; a learning unit 14 that learns the relationship between each of the multiple communication terminals corresponding to the code length of the determined code and the wireless resources allocated to each of the multiple communication terminals using a machine learning model; and a setting unit 15 that sets communication management information including the learned machine learning model to the multiple communication terminals 2.

Description

[Technical Field]

[0001] The present invention relates to a communication management device, a communication management method, and a communication management system. [Background technology]

[0002] In recent years, with the increase in the number of communication terminals, including IoT terminals, there is a demand for technology that can efficiently encode communication data transmitted from each communication terminal and improve the resource utilization efficiency of the entire network.For example, Patent Document 1 discloses a device that stores an initial state calculated from the bit sequence of an input data packet, performs arithmetic coding by applying the concept of its own information content based on the frequency of occurrence of each symbol, generates a code string with an optimal code length assigned and a final state, and transmits them.

[0003] The technology described in Patent Document 1 targets the entropy of a single communication terminal and does not take into consideration reflecting the actual communication patterns of multiple communication terminals as a whole. FIG. 9 is a schematic diagram showing the configuration of a conventional communication management system. As shown in FIG. 9, in the conventional example, instead of individually and in detail evaluating the probability of occurrence of a communication event for each of multiple communication terminals 250, the probability of occurrence of a communication event for all communication terminals 250 is assumed to be the same, and a network is designed based on the entropy of a single communication terminal 250. The actual communication entropy for each communication terminal 250, i.e., the entropy which is the expected value of the amount of information of a communication event occurring at that communication terminal 250, cannot be accurately reflected under the uniform assumption. As a result, the network design is redundant compared to the actual communication pattern.

[0004] In particular, when allocating wireless resources, the probability of occurrence of communication events for multiple communication terminals is assumed to be uniform, and wireless resources are allocated equally to each of the multiple communication terminals, which limits the improvement of the overall resource utilization efficiency of the network. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 2023-107751 Summary of the Invention [Problem to be solved by the invention]

[0006] According to conventional techniques, radio resources are allocated without taking into consideration the probability of a communication event occurring for each of a plurality of communication terminals, which limits the improvement of the overall resource utilization efficiency of the network.

[0007] The present invention has been made to solve the above-mentioned problems, and has as its object to improve the resource utilization efficiency of the entire network. [Means for solving the problem]

[0008] In order to solve the above-mentioned problems, the communication management device of the present invention comprises: a calculation unit configured to calculate entropy, which is an expected value of the amount of information obtained when communication occurs by each of a plurality of communication terminals, from the communication occurrence probability of each of the plurality of communication terminals calculated based on the communication history of the plurality of communication terminals; a determination unit configured to determine a code length of a code corresponding to the communication occurrence probability of each of the plurality of communication terminals, wherein the average code length of the code lengths corresponding to the communication occurrence probability of the plurality of communication terminals has the entropy as a lower limit; an allocation unit configured to allocate wireless resources to each of the plurality of communication terminals based on the code length of the code determined for each of the plurality of communication terminals; a learning unit configured to learn the relationship between each of the plurality of communication terminals corresponding to the code length of the determined code and the wireless resources allocated to each of the plurality of communication terminals using a machine learning model; and a setting unit configured to set communication management information including the learned machine learning model to the plurality of communication terminals.

[0009] In the communication management device according to the present invention, the determination unit may determine the code length using Huffman coding.

[0010] Furthermore, the communication management device according to the present invention may further include a first acquisition unit configured to acquire the communication history collected in a core network that controls communications between the plurality of communication terminals.

[0011] In order to solve the above-mentioned problems, the communication management method of the present invention includes a calculation step of calculating entropy, which is an expected value of the amount of information obtained when communication occurs by each of a plurality of communication terminals, from the communication occurrence probability of each of the plurality of communication terminals calculated based on the communication history of the plurality of communication terminals; a determination step of determining a code length of a code corresponding to the communication occurrence probability of each of the plurality of communication terminals, wherein the average code length of the code lengths corresponding to the communication occurrence probability of the plurality of communication terminals is set to a lower limit of the entropy; an allocation step of allocating wireless resources to each of the plurality of communication terminals based on the code length of the code determined for each of the plurality of communication terminals; a learning step of learning the relationship between each of the plurality of communication terminals corresponding to the determined code length of the code and the wireless resources allocated to each of the plurality of communication terminals using a machine learning model; and a setting step of setting communication management information including the learned machine learning model to the plurality of communication terminals.

[0012] In addition, the communication management method of the present invention may further include a second acquisition step executed by each of the multiple communication terminals to acquire the trained machine learning model included in the communication management information set in the setting step, a third acquisition step to acquire identification information of the own terminal, an inference step to provide the acquired identification information of the own terminal to the trained machine learning model as information indicating the communication terminal to be inferred corresponding to the code length of an unknown code, perform calculations on the trained machine learning model, and output the wireless resources allocated to the own terminal, and a communication step to perform communication using the wireless resources allocated to the own terminal output in the inference step.

[0013] In the communication management method according to the present invention, the determining step may determine the code length using Huffman coding.

[0014] Furthermore, the communication management method according to the present invention may further include a first acquisition step of acquiring the communication history collected in a core network that controls communications between the plurality of communication terminals.

[0015] In order to solve the above-mentioned problems, the communication management system of the present invention is a communication management system comprising the above-mentioned communication management device and the plurality of communication terminals, wherein each of the plurality of communication terminals comprises a second acquisition unit configured to acquire the trained machine learning model included in the communication management information set by the communication management device, a third acquisition unit configured to acquire identification information of the own terminal, an inference unit configured to provide the acquired identification information of the own terminal to the trained machine learning model as information indicating the communication terminal to be inferred corresponding to the code length of an unknown code, perform calculations on the trained machine learning model, and output the wireless resources allocated to the own terminal, and a communication unit configured to perform communication using the wireless resources allocated to the own terminal output by the inference unit. [Effects of the Invention]

[0016] According to the present invention, a code length of a code corresponding to a communication occurrence probability of each of a plurality of communication terminals is determined, a relationship between each of the plurality of communication terminals corresponding to the determined code length and the wireless resources allocated to each of the plurality of communication terminals is learned using a machine learning model, and communication management information including the learned machine learning model is set to the plurality of communication terminals, thereby improving resource utilization efficiency of the entire network. [Brief explanation of the drawings]

[0017] [Figure 1] FIG. 1 is a block diagram showing the configuration of a communication management system including a communication management device according to an embodiment of the present invention. [Figure 2] FIG. 2 is a block diagram showing the configuration of a communication terminal according to this embodiment. [Figure 3] FIG. 3 is a diagram for explaining an outline of a communication management system including a communication management device according to this embodiment. [Figure 4] FIG. 4 is a diagram for explaining an outline of a communication management system including a communication management device according to this embodiment. [Figure 5] FIG. 5 is a diagram illustrating the configuration of the learning unit of the communication management device according to the present embodiment. [Figure 6] FIG. 6 is a block diagram showing the hardware configuration of the communication management device according to this embodiment. [Figure 7] FIG. 7 is a block diagram showing a hardware configuration of a communication terminal according to this embodiment. [Figure 8] FIG. 8 is a sequence diagram showing an outline of the operation of the communication management system according to the present embodiment. [Figure 9] FIG. 9 is a diagram for explaining a conventional communication management system. DETAILED DESCRIPTION OF THE INVENTION

[0018] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will now be described in detail with reference to FIGS.

[0019] [Communication Management System Configuration] First, an overview of a communication management system including a communication management device 1 according to an embodiment of the present invention will be described with reference to FIG.

[0020] The communication management system according to this embodiment includes a communication management device 1, a communication terminal 2, a base station 3, and a core network 4. As an example, the communication management system is provided in a 5G mobile communication network, but may also be a network that uses fixed lines. As shown in FIG. 1, the communication management device 1 is connected to the core network 4 via a network NW such as a LAN, a WAN, or the Internet.

[0021] The communication terminal 2 is realized as a mobile communication terminal such as a smartphone, a tablet computer, a laptop computer, a wearable device, an industrial robot, etc. The communication terminal 2 includes a subscriber identity module (SIM) 208, and the contract profile of the SIM 208 includes identifier information such as an international mobile subscriber identity (IMSI). The communication terminal 2 is uniquely identified by the IMSI.

[0022] The communication terminal 2 is also configured as an IoT device to which a terminal IP address that uniquely identifies the terminal is assigned. In this embodiment, there are n communication terminals 2 (n is a positive integer of 2 or more). The communication terminals 2 connect to an external data network (not shown) from a core network 4 via the base station 3 in which each communication terminal 2 is located.

[0023] The base station 3 is composed of a wireless base station compatible with the 5G system, and relays communications between the communication terminal 2 present in the communication area and the core network 4. The base station 3 is connected to the core network 4 via a network such as a backhaul link. The base station 3 relays data from the communication terminal 2 when the communication terminal 2 communicates using a resource block of the time zone x frequency band of the wireless resources assigned to the communication terminal 2 based on the communication management information set by the communication management device 1.

[0024] The core network 4 provides centralized control, routing, management, and security for communications relayed by the base station 3. The core network 4 includes a UPF (User Plane Function) 40 in the U-plane. The core network 4 also includes nodes in the C-plane, such as an AMF (Access and Mobility Management Function) and a UDM (Unified Data Management), which are not shown. Functional nodes in the U-plane and C-plane other than the UPF 40 that the core network 4 includes are not shown in the figure.

[0025] The UPF 40 is a user plane function that processes packets between the base station 3 and a data network such as the Internet. The UPF 40 includes a communication interface 40a for communicating with the communication management device 1. The UPF 40 collects communication history of the communication terminal 2 that performs traffic processing.

[0026] [Communication management device functional block] As shown in FIG. 1, the communication management device 1 includes a first acquisition unit 10, a calculation unit 11, a determination unit 12, an allocation unit 13, a learning unit 14, a setting unit 15, and a first storage unit 16.

[0027] The first acquisition unit 10 acquires communication histories collected by the core network 4 that controls communications between multiple communication terminals 2. The first acquisition unit 10 acquires communication histories from all communication terminals 2 for which traffic processing is performed on the UPF 40. The first acquisition unit 10 can acquire communication histories for any set period, such as one month. The communication history is a communication log that records communication events such as the transmission and reception of packets for each IMSI, and in particular, is a communication log that records whether or not packets have been transmitted or received.

[0028] The calculation unit 11 calculates entropy, which is an expected value of the amount of information obtained when communication occurs with each of the multiple communication terminals 2, from the communication occurrence probability of each of the multiple communication terminals 2 calculated based on the communication histories of the multiple communication terminals 2. The calculation unit 11 calculates the probability of communication occurring for each IMSI from the communication history for each IMSI for one month acquired by the first acquisition unit 10. For example, when communication histories indicating the number of communications performed for n=100 communication terminals 2 for one month are collected, it is assumed that the communication terminal 2 with IMSI_1 performed communications 10 times per month. In this case, the communication occurrence probability of IMSI_1 is calculated to be 10 / 100=0.1. The calculation unit 11 similarly calculates the communication occurrence probabilities for the other communication terminals 2.

[0029] The calculation unit 11 calculates the sum of the communication occurrence probabilities for each of the 100 devices and normalizes the sum so that the total value is 1. For example, if the sum of the communication occurrence probabilities for the 100 communication terminals 2 is 2.5, the value of the communication occurrence probability for each communication terminal 2 is divided by 2.5. In this case, the communication occurrence probability for IMSI_1 is 0.1 / 2.5=0.04.

[0030] Based on Shannon's source coding theorem, the calculation unit 11 calculates entropy H(S), which is the average information amount of the expected value of the information amount, for all events from the probability of each event occurring, using the following equation (1).

number

[0031] In the above equation (1), ρ i indicates a communication occurrence probability, which is the occurrence of communication at each communication terminal 2, and M indicates the total number of communication terminals 2. Furthermore, an event (communication event) is whether or not communication occurs at each communication terminal 2, and the calculation unit 11 calculates the communication occurrence probability ρ i Based on the self-information of each event, lnρ i Then, calculate the self-information lnρ i The entropy H(S) is calculated by taking the probability-weighted average of

[0032] The determination unit 12 is configured to determine the code length of the code according to the communication occurrence probability of each of the multiple communication terminals 2. The determination unit 12 determines the code length so that the average code length of the code lengths according to the communication occurrence probability of the multiple communication terminals 2 is the lower limit of the entropy. Here, the average code length L bar, which is the probability-weighted average of the code lengths assigned to each event, is expressed by the following equation (2).

number

[0033] In the above equation (2), lρ i indicates the code length for the probability of communication occurrence for each communication terminal 2. Furthermore, according to Shannon's source coding theorem, the average code length L bar in the above equation (2) has a relationship with the entropy H(S) in the above equation (1) as expressed in the following equation (3).

number

[0034] In the above equation (3), e is an arbitrary value e>0. The above equation (3) indicates that an event, which is a symbol from the information source, can be coded with an average code length L bar that is very close to the entropy H(S). The decision unit 12 constructs a decision tree using instantaneous codes such as Huffman codes based on the above equations (2) and (3), and assigns an integer code length to each communication terminal 2 representing each event according to its communication occurrence probability. An instantaneous code is a code that allows the division of symbols from the information source to be instantly identified when a coded sequence of symbols from the information source is transmitted in time series.

[0035] 3 is a diagram showing a process in which the determination unit 12 determines the code and code length of an event according to the communication occurrence probability of each communication terminal 2. In FIG. 3, the communication occurrence probability ρ i First, the determination unit 12 determines the communication occurrence probability ρ i Next, ii) 0 and 1 are associated with a pair of events with a low probability of communication occurrence. In FIG. 3, IMSI_3 and IMSI_4 are associated with 0 and 1. After that, the decision unit 12 iii) calculates the sum of the probabilities (0.1+0.2=0.3) for the pair of events of IMSI_3 and IMSI_4, and again calculates the communication occurrence probability ρ i The determination unit 12 repeats i) to iii) and finds a sequence of 0s and 1s by tracing the corresponding sequence of 0s and 1s in reverse order, and determines this as the code of each event.

[0036] As shown in Figure 3, for the event of whether or not communication occurs by each of IMSI_1 to IMSI_4, the code c is determined as 101 for IMSI_1, 11 for IMSI_2, 100 for IMSI_3, and 0 for IMSI_4. Also, by determining the code c for each event, the code length lρ iare determined to be 3 bits, 2 bits, 3 bits, and 1 bit, respectively. In this way, the deeper an event is in the decision tree, the longer the code length. By determining the code and code length for each communication terminal 2, each code and code length is associated with the IMSI, which is identification information for each communication terminal 2. Information associating each code and code length with each IMSI is stored in the first storage unit 16.

[0037] The allocation unit 13 allocates radio resources to each of the plurality of communication terminals 2 based on the code length of the code determined for each of the plurality of communication terminals 2. Specifically, the allocation unit 13 sequentially allocates the code length of the code determined for each communication terminal 2 to resource blocks constituting the radio resources. Fig. 4 is a diagram for explaining the radio resources allocated to each communication terminal 2 by the setting unit 15. Fig. 4 shows a two-dimensional plane in which the radio resources are defined by the time axis and the frequency axis. As shown in Fig. 4, the radio resources are further allocated to a time period t N and frequency band f N The total code length determined for multiple communication terminals 2 is L tot =N, and the radio resources are variable resources that allow dynamic allocation.

[0038] The allocation unit 13 sequentially allocates the code lengths of the codes determined for IMSI_1 to IMSI_n to resource blocks of the radio resources, and associates the time period t and frequency band f with the IMSI of each communication terminal 2. For example, if the code length of the code of IMSI_1 is 3 bits, resource blocks of time period t1 × frequency band f1, time period t2 × frequency band f2, and time period t3 × frequency band f3 are allocated. Similarly, resource blocks are sequentially allocated to IMSI_2 to IMSI_n according to their code lengths. In this way, the setting unit 15 specifies which resource blocks each communication terminal 2 can use in the radio resources shared by multiple communication terminals 2. Information on the radio resources allocated to each communication terminal 2 by the allocation unit 13 is stored in the first storage unit 16.

[0039] The learning unit 14 is configured to use a machine learning model to learn the relationship between each of the multiple communication terminals 2 corresponding to each of the code lengths of the determined codes and the radio resources allocated to each of the multiple communication terminals 2. More specifically, the learning unit 14 uses the machine learning model to learn the relationship between the IMSI values, which are identification information of each of the multiple communication terminals 2 associated with each of the code lengths of the determined codes, and the radio resources allocated to each of these IMSIs. In other words, the learning unit 14 learns the relationship between the different entropy values of the multiple communication terminals 2 and the radio resource allocation information. The learning unit 14 can learn the machine learning model through supervised learning.

[0040] FIG. 5 shows the structure of a neural network model employed as an example of a machine learning model used by the learning unit 14. The neural network model includes an input layer x, a hidden layer h, and an output layer y. The learning unit 14 provides the IMSIs of each of the multiple communication terminals 2 corresponding to each code and code length to the input layer of the neural network model, applies an activation function to the weighted sum of the inputs, and passes the output determined by threshold processing to the output layer. The output node of the output layer outputs a predicted output of the model for the time period t and frequency band f of the radio resources allocated to each of the multiple communication terminals 2. The neural network model is constructed as a multi-label classification model in which the time period t and frequency band f are selected, respectively.

[0041] For example, when the code length determined by the determination unit 12 is 8 bits, the output nodes x1, x2, . . . , x indicating the allocation of the time slot t are N For example, the values of x1 to x8 are 1, and the other output nodes x9 to x N takes 0. Similarly, the output nodes o1, o2, . . . , o N For example, the values of o1 to o8 are 1, and the other output nodes o9 to o Ntakes the value 0. The number N of output nodes corresponding to time period t and the number N of output nodes corresponding to frequency band f are the same as the total value of the code lengths determined for the multiple communication terminals 2 to be managed. As explained in the allocation of wireless resources in FIG. 4, each communication terminal 2 is allocated resource blocks that do not overlap with each other, so the output node values for the multiple communication terminals 2 are different from each other.

[0042] The learning unit 14 introduces an objective function E of the following equation (4) to learn the parameters of the neural network model so that the predicted values of the time zone t and frequency band f indicating the allocation of radio resources from the neural network model to the IMSI of the communication terminal 2 identified by the code length of the code determined by the determination unit 12 become the values of the time zone t and frequency band f of the correct label.

number

[0043] In the above equation (4), y1, y2, , y N ,o1,o2,···,o N indicates the predicted output value of each output node. Also, Y1, Y2, , Y N ,O1,O2,···,O N is a correct label of the training data, and is prepared separately in advance. The training data is an IMSI value, which is identification information of the communication terminal 2 associated with the code length, and a time period Y and a frequency band O of the radio resource, which are correct labels assigned to the IMSI value. The learning unit 14 adjusts the weight parameters of the neural network model so that the objective function E in the above equation (4) is minimized, that is, becomes 0. The learning unit 14 can optimize the objective function E using an error backpropagation method or the like. As the objective function, cross-entropy loss can be used instead of squared error.

[0044] The setting unit 15 is configured to set communication management information including the trained machine learning model to a plurality of communication terminals 2. The setting unit 15 can broadcast the communication management information to the IMSIs of the plurality of communication terminals 2 via the network NW. The setting unit 15 can transmit the communication management information to the plurality of communication terminals 2 via the core network 4 and the base station 3. The communication management information can include resource-related information such as resource block mapping information in addition to the trained machine learning model.

[0045] The first storage unit 16 stores the trained machine learning model constructed by the learning unit 14. The first storage unit 16 can store the IMSIs of multiple communication terminals 2 to be managed. The first storage unit 16 also stores allocation information of resource blocks of wireless resources for each communication terminal 2 shown in FIG. 4. Furthermore, the first storage unit 16 stores information that associates the IMSIs of multiple communication terminals 2 with corresponding codes and code lengths.

[0046] [Communication terminal function block] 2 is a functional block diagram showing the configuration of a communication terminal 2 according to the present embodiment. Each of the multiple communication terminals 2 has the same configuration. As shown in FIG. 2, the communication terminal 2 includes a second acquisition unit 20, a third acquisition unit 21, an inference unit 22, a communication unit 23, and a second storage unit 24.

[0047] The second acquisition unit 20 acquires the trained machine learning model included in the communication management information set by the communication management device 1. The second acquisition unit 20 can acquire the trained machine learning model by reading it from the second storage unit 24 in which the communication management information transmitted by the setting unit 15 of the communication management device 1 is stored.

[0048] The third acquisition unit 21 acquires the identification information of the terminal itself. The third acquisition unit 21 acquires the IMSI stored in the SIM 208 of the terminal itself.

[0049] The inference unit 22 provides the IMSI, which is the identification information of the own terminal acquired by the third acquisition unit 21, to the trained machine learning model as information indicating the communication terminal 2 to be inferred that corresponds to the code length of the unknown code, performs calculations on the trained machine learning model, and outputs the radio resources allocated to the own terminal. The communication terminal 2 to be inferred that has the code length of the unknown code is a communication terminal 2 that has some code length, that is, some entropy value, and is a terminal for which radio resources to be allocated to the communication terminal 2 have not been requested.

[0050] The inference unit 22 provides the IMSI of the terminal itself as an input to be inferred to an input node of the model in which learned parameters are set in the neural network model having the structure shown in Fig. 5. Then, the inference unit 22 performs a weighted product-sum operation on the input value, applies an activation function, and outputs an output value including values of the time period t and the frequency band f that indicate the resource block allocated to the IMSI of the terminal itself.

[0051] The communication unit 23 performs communication using the radio resources allocated to the terminal itself output by the inference unit 22. The communication unit 23 encodes data within the allocated resource blocks, appropriately modulates and maps the data, and performs communication using the resource blocks of the allocated time period t and frequency band f.

[0052] The second storage unit 24 stores communication management information. The second storage unit 24 stores the IMSI, which is identification information of the terminal itself.

[0053] [Hardware configuration of communication management device] Next, an example of a hardware configuration for realizing the communication management device 1 having the above-described functions will be described with reference to FIG.

[0054] 6, the communication management device 1 can be realized by, for example, a computer including a processor 102, a main memory device 103, a communication interface 104, an auxiliary memory device 105, and an input / output (I / O) 106 connected via a bus 101, and a program that controls these hardware resources. The communication management device 1 can also include a display device 107 connected via the bus 101.

[0055] The processor 102 is realized by a CPU, a GPU, an FPGA, an ASIC, or the like.

[0056] The main memory device 103 pre-stores programs for the processor 102 to perform various controls and calculations. The processor 102 and the main memory device 103 implement the functions of the communication management device 1, such as the first acquisition unit 10, the calculation unit 11, the determination unit 12, the allocation unit 13, the learning unit 14, and the setting unit 15 shown in FIG.

[0057] The communication interface 104 is an interface circuit for connecting the communication management device 1 to various external electronic devices via a network.

[0058] The auxiliary storage device 105 is composed of a readable / writable storage medium and a drive for reading and writing various information such as programs and data from and to the storage medium. The auxiliary storage device 105 can use a semiconductor memory such as a hard disk or flash memory as the storage medium.

[0059] The auxiliary storage device 105 has a program storage area for storing the communication management program executed by the communication management device 1. The auxiliary storage device 105 realizes the first storage unit 16 described in FIG. 1. The auxiliary storage device 105 has a program storage area for storing a learning program for the communication management device 1 to perform supervised learning. The auxiliary storage device 105 also has an area for storing the IMSI of the communication terminal 2 to be managed. Furthermore, the auxiliary storage device 105 may have, for example, a backup area for backing up the above-mentioned data, programs, etc.

[0060] The input / output I / O 106 is an input / output device that inputs signals from external devices and outputs signals to external devices.

[0061] The display device 107 is configured by an organic EL display, a liquid crystal display, or the like.

[0062] [Hardware configuration of communication terminal] Next, an example of a hardware configuration for realizing the communication terminal 2 having the above-described functions will be described with reference to FIG.

[0063] As shown in Fig. 7, the communication terminal 2 can be realized by, for example, a computer including a processor 202, a main memory device 203, a communication interface 204, an auxiliary memory device 205, and an input / output (I / O) 206 connected via a bus 201, and a program that controls these hardware resources. The communication terminal 2 also includes a display device 207 and an SIM 208 connected via the bus 201. The communication terminal 2 has the same hardware configuration as the communication management device 1 shown in Fig. 6. The following description will focus on the different configurations.

[0064] The processor 202 is realized by a CPU, a GPU, an FPGA, an ASIC, etc. The processor 202 includes one or more processors, and may include a baseband processor that performs baseband processing, and a main processor that processes application execution and performs calculations on trained machine learning models.

[0065] The auxiliary storage device 205 has a program storage area for storing an inference program executed by the communication terminal 2. The auxiliary storage device 205 realizes the second storage unit 24 described in Fig. 2. Furthermore, for example, the auxiliary storage device 205 may have a backup area for backing up the above-mentioned data, programs, etc.

[0066] The SIM 208 is configured by a physical SIM or an eSIM (Embedded SIM), and stores an IMSI, an authentication key, and the like assigned to the subscriber of the communication terminal 2. The SIM 208 realizes the second storage unit 24 described in FIG.

[0067] [Operation of the communication management system] Next, the operation of the communication management system having the above-described configuration will be described with reference to the sequence diagram of FIG.

[0068] 8 is an operation sequence showing an overview of the operation of a communication management system including a communication management device 1 and a communication terminal 2. First, the UPF 40 collects communication histories of multiple communication terminals 2 to be managed (step S1). Next, the first acquisition unit 10 of the communication management device 1 acquires the communication histories from the UPF 40 via the network NW (step S2). The first acquisition unit 10 can acquire the communication histories of each communication terminal 2 for a preset period, for example, one month, from the UPF 40.

[0069] Next, the calculation unit 11 uses the above formula (1) to calculate the entropy H(S), which is the average information amount of the expected value of the information amount, for all events of the multiple communication terminals 2 from the communication occurrence probability, which is the occurrence probability of an event related to whether or not communication occurs by each communication terminal 2 (step S3).

[0070] Next, the determination unit 12 determines the code length of the code according to the communication occurrence probability of each of the multiple communication terminals 2 (step S4). The determination unit 12 constructs a decision tree using instantaneous codes such as Huffman codes based on the above equations (2) and (3), and determines the code and code length for each event. Subsequently, the allocation unit 13 allocates wireless resources to each of the multiple communication terminals 2 based on the code length of each event determined in step S4 (step S5). The allocation unit 13 sequentially allocates the time slot t and frequency band f of the resource blocks constituting the wireless resources, and the code length associated with each communication terminal 2 (IMSI), to the wireless resources of FIG. 4. The allocation information is stored in the first storage unit 16.

[0071] Next, the learning unit 14 uses a machine learning model to learn the relationship between each of the multiple communication terminals 2 corresponding to each of the code lengths determined in step S5 and the radio resources allocated to each of the multiple communication terminals 2 (step S6). In step S6, the learning unit 14 trains the neural network model of FIG. 5 using training data in which the resource blocks (time slot t and frequency band f) allocated to the IMSIs (IMSI_1 to IMSI_n) are assigned as correct labels to the IMSI values (IMSI_1 to IMSI_n) of the communication terminals 2 corresponding to each of the determined code lengths. The learning unit 14 provides the IMSI values of the communication terminals 2 corresponding to each code length as input to the neural network model, calculates model predicted values for the time slot t and frequency band f, and optimizes parameters using the objective function E of equation (4) above or cross-entropy loss to minimize errors with the correct labels for the time slot t and frequency band f values. The trained machine learning model is stored in the first storage unit 16.

[0072] Next, the setting unit 15 sets the communication management information including the trained machine learning model constructed in step S6 to the multiple communication terminals 2 (step S7). The setting unit 15 can transmit the communication management information by broadcast to all of the multiple communication terminals 2. Thereafter, the second acquisition unit 20 of the communication terminal 2 acquires the trained machine learning model included in the communication management information (step S8).

[0073] Next, the third acquisition unit 21 of the communication terminal 2 acquires the IMSI of the communication terminal 2 from the SIM 208 (step S9). Subsequently, the inference unit 22 of the communication terminal 2 provides the value of the IMSI of the communication terminal 2 acquired in step S9 to the trained machine learning model acquired in step S8 as information of the communication terminal 2 to be inferred, which has the code length of an unknown code, performs calculations on the trained machine learning model, and outputs the resource blocks (time period t and frequency band f) allocated to the IMSI of the communication terminal 2 (step S10).

[0074] Thereafter, the communication unit 23 performs data communication using the resource blocks of the time period t and frequency band f determined in step S10 (step S11). Each communication terminal 2 encodes data within the allocated resource blocks, modulates it appropriately, and maps it. The mapped data is transmitted through the designated resource blocks. Each of the multiple communication terminals 2 to be managed performs the processes from step S8 to step S11.

[0075] As described above, the communication management device 1 according to the present embodiment determines the code length of a code according to the communication occurrence probability of each of the plurality of communication terminals 2, and allocates wireless resources to each of the plurality of communication terminals 2 based on the determined code length. Then, the relationship between each of the plurality of communication terminals 2 corresponding to each of the determined code lengths and the wireless resources allocated to each of the plurality of communication terminals 2 is learned using a machine learning model, and communication management information including the learned machine learning model is set to the plurality of communication terminals 2. This improves resource utilization efficiency throughout the network. In particular, in contrast to conventional techniques that treat entropy based on the communication occurrence probability as uniform and allocate wireless resources equally to each communication terminal 2, the communication management device 1 according to the present embodiment takes into account the communication occurrence probability that differs for each communication terminal 2, thereby effectively reducing the number of frequency bands when allocating wireless resources.

[0076] In the embodiment described above, the communication management system is described as a system conforming to the 5G standard, but the communication standard may be 3G, 4G / LTE, 6G, etc. Furthermore, the communication management system is not limited to a mobile communication network and may be a network using a fixed line as described above. In this case, the communication history may be collected via a wireless router or a wireless access point.

[0077] In the embodiment described above, the determination unit 12 determines the code length using Huffman coding. However, other coding methods such as Shannon-Fano coding and arithmetic coding can also be used to determine the code length.

[0078] In the embodiment described above, the learning unit 14 learns a machine learning model that receives the IMSI of the communication terminal 2 as an input and outputs the radio resources allocated to the IMSI. However, the input to the machine learning model can be identification information other than the IMSI, such as the IP address of the communication terminal 2.

[0079] In the embodiment described above, the communication management device 1 is configured as a device provided outside the wireless access network or the core network 4. However, the communication management device 1 can be installed in the wireless access network, such as in the base station 3.

[0080] The above describes embodiments of the communication management device, communication management method, and communication management system of the present invention, but the present invention is not limited to the described embodiments, and various modifications that a person skilled in the art can imagine are possible within the scope of the invention described in the claims. [Explanation of symbols]

[0081] 1...communication management device, 10...first acquisition unit, 11...calculation unit, 12...determination unit, 13...allocation unit, 14...learning unit, 15...setting unit, 16...first memory unit, 2...communication terminal, 20...second acquisition unit, 21...third acquisition unit, 22...inference unit, 23...communication unit, 24...second memory unit, 3...base station, 4...core network, 40...UPF, 101, 201...bus, 102, 202...processor, 103, 203...main memory device, 40a, 104, 204...communication interface, 105, 205...auxiliary memory device, 106, 206...input / output I / O, 107, 207...display device, 208...SIM, NW...network.

Claims

1. a calculation unit configured to calculate entropy, which is an expected value of an amount of information obtained when communication occurs by each of a plurality of communication terminals, from a communication occurrence probability of each of the plurality of communication terminals calculated based on a communication history of the plurality of communication terminals; a determination unit configured to determine a code length of a code according to the communication occurrence probability of each of the plurality of communication terminals, wherein an average code length of the code lengths according to the communication occurrence probability of the plurality of communication terminals has the entropy as a lower limit; an allocation unit configured to allocate radio resources to each of the plurality of communication terminals based on the code length of the code determined for each of the plurality of communication terminals; a learning unit configured to learn, using a machine learning model, a relationship between each of the plurality of communication terminals corresponding to the code length of the determined code and the radio resources allocated to each of the plurality of communication terminals; a setting unit configured to set communication management information including a trained machine learning model constructed by the training unit in the plurality of communication terminals; A communication management device comprising:

2. 2. The communication management device according to claim 1, The determination unit determines the code length using Huffman coding. A communication management device characterized by:

3. 2. The communication management device according to claim 1, The communication device further includes a first acquisition unit configured to acquire the communication history collected in a core network that controls communications between the plurality of communication terminals. A communication management device characterized by:

4. Executed by a communication management device, a calculation step of calculating entropy, which is an expected value of the amount of information obtained when communication occurs by each of the plurality of communication terminals, from the communication occurrence probability of each of the plurality of communication terminals calculated based on the communication history of the plurality of communication terminals; a determining step of determining a code length of a code according to the communication occurrence probability of each of the plurality of communication terminals, wherein an average code length of the code lengths according to the communication occurrence probability of the plurality of communication terminals is set to a lower limit of the entropy; an allocating step of allocating radio resources to each of the plurality of communication terminals based on the code length of the code determined for each of the plurality of communication terminals; a learning step of learning, using a machine learning model, a relationship between each of the plurality of communication terminals corresponding to the code length of the determined code and the radio resources allocated to each of the plurality of communication terminals; a setting step of setting communication management information including a trained machine learning model constructed by the learning step in the plurality of communication terminals; A communication management method comprising:

5. 5. The communication management method according to claim 4, Further, the method is executed by each of the plurality of communication terminals. a second acquisition step of acquiring the trained machine learning model included in the communication management information set in the setting step; a third acquisition step of acquiring identification information of the terminal; an inference step of providing the acquired identification information of the own terminal to the trained machine learning model as information indicating an inference target communication terminal corresponding to a code length of an unknown code, performing calculations on the trained machine learning model, and outputting the wireless resource allocated to the own terminal; a communication step of performing communication using the wireless resource allocated to the terminal output in the inference step; A communication management method comprising:

6. 5. The communication management method according to claim 4, The determining step determines the code length using Huffman coding. A communication management method comprising:

7. 5. The communication management method according to claim 4, The method further includes a first acquisition step of acquiring the communication history collected in a core network that controls communications between the plurality of communication terminals, the first acquisition step being executed by the communication management device. A communication management method comprising:

8. A communication management system comprising the communication management device according to any one of claims 1 to 3 and the plurality of communication terminals, Each of the plurality of communication terminals a second acquisition unit configured to acquire the trained machine learning model included in the communication management information set by the communication management device; a third acquisition unit configured to acquire identification information of the terminal; an inference unit configured to provide the acquired identification information of the own terminal to the trained machine learning model as information indicating an inference target communication terminal corresponding to a code length of an unknown code, perform calculations on the trained machine learning model, and output the wireless resource allocated to the own terminal; a communication unit configured to perform communication by using the wireless resource allocated to the terminal output by the inference unit; A communication management system comprising:

Images