JP6492113B2 - Traffic prediction apparatus and traffic prediction method - Google Patents

Description

  The present invention relates to a traffic prediction device and a traffic prediction method.

  With the increase in functionality of mobile terminals and the increase in bandwidth of mobile communication networks, it has become an environment where mobile traffic is likely to occur regardless of location. In addition, since the usage forms of mobile terminals are diversifying, local upload / download traffic (hereinafter referred to as “event traffic”) by using various mobile applications in the venue and its surroundings at the time of a large-scale event. .) Is expected to occur.

  Event traffic has local time of occurrence and place of occurrence, and the tendency of apps used for each event is different. For this reason, in response to new / additional deployment of permanently installed communication resources such as optical fibers and base stations, the usage rate of communication resources during normal times is reduced, resulting in excessive capital investment.

  Therefore, by appropriately deploying communication resources that can be temporarily prepared, such as mobile base station vehicles and Wi-Fi (registered trademark) AP, communication quality on the day of the event can be suppressed while suppressing the cost required for communication resources. Can be improved.

  In this case, in order to estimate the amount of temporary communication resources that should be prepared and to consider the location of communication resources that maximizes the capacity of event traffic, it is possible to predict event traffic volume in advance at the spatio-temporal level. It becomes important.

  Conventionally, a large-scale event such as a comic market is a target of event traffic countermeasures, and the amount of event traffic at the next event and the deployment of communication resources have been implemented based on past experience and know-how. As an approach for predicting communication traffic based on past know-how, there is an approach for predicting future communication traffic volume using extrapolation based on past observed communication traffic data (Patent Document 1, Patent Document). 2).

Japanese Patent Laying-Open No. 2015-231187 JP2015-216585A

  However, until now large-scale events have been mainly held at the same place, but due to the diversification of mobile services and the enhancement of mobile communication network functions that make it possible, Events that have not been implemented in the past and event traffic that accompanies them are expected to increase regardless of location. In considering traffic countermeasures for such new events (hereinafter referred to as “future events”), past know-how such as traffic observation values has not yet been obtained. It is difficult to apply traffic prediction directly.

  The present invention has been made in view of the above points, and an object thereof is to make it possible to predict traffic in a future event at a spatio-temporal level.

  Therefore, in order to solve the above-described problem, the traffic prediction apparatus uses, as agent data set in advance for each agent, the locations of a plurality of agents each of which aggregates a plurality of users of communication terminals in a simulation space. The mobile behavior unit to be changed based on the unit time, and for each agent, for each agent, determine whether to communicate based on the communication occurrence probability set in the storage unit, for the agent determined to communicate, A communication action unit that determines a communication app to be used based on a selection ratio set in the storage unit for each communication app, and one or more communication ranges in the simulation space for each unit time. Determined by the communication action unit for each agent including A communication environment control unit that determines a bandwidth to be allocated to each agent based on the requested bandwidth of the communication application and the amount of traffic that can be used per unit time set in the communication range. .

  Traffic in future events can be predicted at the spatio-temporal level.

It is a 1st figure which shows the structural example of embodiment of this invention. It is a 2nd figure which shows the structural example of embodiment of this invention. It is a figure which shows the hardware structural example of the simulator apparatus in embodiment of this invention. It is a figure which shows the function structural example of the simulator apparatus in embodiment of this invention. It is a figure which shows the image regarding the movement range and destination candidate of an agent. It is a figure which shows the image regarding a communication range. It is a flowchart for demonstrating an example of the process sequence which a MAS engine performs. It is a flowchart for demonstrating an example of the process sequence which a purpose management part performs. It is a flowchart for demonstrating an example of the process sequence which a route calculation part performs. It is a flowchart for demonstrating an example of the process sequence which a movement action part performs. It is a flowchart for demonstrating an example of the process sequence which a communication action part performs. It is a flowchart for demonstrating an example of the process sequence which a movement environment control part performs. It is a flowchart for demonstrating an example of the process sequence which a communication environment control part performs. It is a figure which shows the information around the event venue made into the object in the traffic prediction of this Embodiment. It is a figure which shows the total DL traffic amount at the time of using each model in a simulation range.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In this embodiment, in the case of a communication event on a mobile communication network, when a large-scale event such as a sports event or a comic market is newly planned, communication resources to be prepared in advance for accommodating traffic are allocated. The present invention relates to a technique for predicting the amount of mobile traffic around an event venue with a specified time and space granularity for the purpose of estimation.

  The present embodiment relates to the mechanism of the collective behavior model 13 and the environment model 14 and an apparatus for operating the model. The collective behavior model 13 regards a representative mobile user group (hereinafter simply referred to as “group”) at the time of an event as one moving object, and algorithmizes the group movement behavior and communication behavior. In addition, the environmental model 14 includes a movement range of a group such as a road and a building, a communication range in which various communication methods such as Wi-Fi (registered trademark) and LTE (Long Term Evolution) can be used. Alternatively, external factors used as conditions for communication are modeled. In this embodiment, a group is defined as one agent model using a multi-agent based simulation (MAS) framework for simulating individual movements, and based on the behavior result of each group, the mobile of the target event This technology predicts traffic volume.

  1 and 2 are diagrams showing a configuration example of an embodiment of the present invention. This embodiment is a future event for predicting traffic by local upload / download using various mobile apps (hereinafter referred to as “event traffic”) in events such as sports competitions and comic markets. Software that simulates the generation of traffic, hardware that operates the software, and storage that stores input / output data required for the simulation.

  On the software side, the collective behavior model 13, the environment model 14, and the input data storage unit 15 are connected via an API 12 (Application Program Interface) to the MAS engine 11 that controls the entire simulation, and the MAS engine 11 calculates the entire simulation. And the calculation result is output to the output data storage unit 16 via the API 12.

  In the following description, for the sake of simplicity, an example is shown in which all software and input / output data storage media are mounted on a single hardware. However, the collective behavior model 13, As long as data can be transmitted and received among the environment model 14, the API 12, the MAS engine 11, the input data storage unit 15, and the output data storage unit 16, the mounting form is not limited. For example, a storage medium for holding all software and input / output data may be mounted on one hardware, or each software and an input / output data storage medium may be mounted on separate hardware.

  FIG. 3 is a diagram illustrating a hardware configuration example of the simulator device according to the embodiment of the present invention. The simulator device 10 of FIG. 3 includes a drive device 100, an auxiliary storage device 102, a memory device 103, a CPU 104, an interface device 105, and the like that are connected to each other via a bus B.

  A program for realizing processing in the simulator device 10 is provided by a recording medium 101 such as a CD-ROM. When the recording medium 101 storing the program is set in the drive device 100, the program is installed from the recording medium 101 to the auxiliary storage device 102 via the drive device 100. However, the program need not be installed from the recording medium 101 and may be downloaded from another computer via a network. The auxiliary storage device 102 stores the installed program and also stores necessary files and data.

  The memory device 103 reads the program from the auxiliary storage device 102 and stores it when there is an instruction to start the program. The CPU 104 executes functions related to the simulator device 10 in accordance with a program stored in the memory device 103. The interface device 105 is used as an interface for connecting to a network.

  FIG. 4 is a diagram illustrating a functional configuration example of the simulator device according to the embodiment of the present invention. In FIG. 4, the collective behavior model 13, the environment model 14, the API 12, and the MAS engine 11 are realized by a process that the CPU 104 executes one or more programs installed in the simulator device 10. The input data storage unit 15 and the output data storage unit 16 can be realized by using, for example, a storage device that can be connected to the auxiliary storage device 102 or the simulator device 10 via a network.

  In the simulator apparatus 10, the user of the mobile terminal (communication terminal) in the event to be predicted is simulated by various parameters and models. A mobile terminal user that is simulated by various parameters and models is called an “agent”.

  The collective behavior model 13 is an agent behavior algorithm in a prediction target event in simulation. The agent performs the movement behavior and the communication behavior calculated per unit time based on the collective behavior model 13 and parameters obtained from the input data.

The agent performs a movement action and a communication action according to the action purpose according to the scenario set in the agent data D2 stored in the input data storage unit 15. In the agent data D2, the following parameters are set for each agent.
Agent ID: Unique identifier for each agent Attribute: Group ID that is assigned to each group grouped according to the sex, age, hobbies, etc. of the agent and does not change during the simulation time
・ Speed: Walking speed of the agent ・ Stop limit: Time until the action purpose is switched when the agent is stopped ・ Start time: Time when the agent starts action ・ End time: Scheduled time when the agent finishes action ・ Start Location: Destination candidates / arrival places where the agent starts the action: Destination candidates / route points when the agent finishes the action: Action objectives to be implemented before reaching the arrival place (information for identifying the via points) The list of via points includes the following parameters for each action purpose.
・ Priority: Used to change the behavioral purpose of an agent with multiple behavioral goals when it is determined that the currently set behavioral goal cannot be processed for a certain period of time due to congestion or other factors. Parameter (The smaller the parameter, the higher the priority)
・ Task type: Action purpose type ・ Scheduled start time: Scheduled start time for executing action purpose processing ・ Scheduled dwell time: Action purpose is achieved when arriving at a destination candidate that can process the action purpose Time to stay until (scheduled end time takes precedence)
・ Scheduled end time: Time to stop staying during execution of the action purpose process and shift to the next purpose ・ Process completion flag: The initial value is 0 and becomes 1 when the action purpose process is completed. Each agent is placed in the simulation space and has a dynamically changing dynamic state. The dynamic state includes a combination of a simulation time zone, an area type, and a moving state. The intra-simulation time zone is a current time zone in the simulation space (for example, a time zone of 1 hour interval such as 1 o'clock, 2 o'clock,...). The area type is a type (for example, indoor, outdoor, event venue, etc.) of a place (area) including the current location of the agent (hereinafter referred to as “current location”). The movement state indicates the movement state of the agent and is classified into one of movement, stop, and stay. The movement indicates a state where the agent is moving on the simulation. The stop indicates that the agent is stopped on the simulation (because of the environment such as a signal, or due to the agent such as a matrix / congestion). The stay indicates that the agent is staying at the destination candidate. The movement means that a coordinate value indicating the current location of the agent changes in a two-dimensional coordinate system (movement range described later) in the simulation space.

  The collective behavior model 13 includes a purpose management unit 131, a movement behavior unit 132, a communication behavior unit 133, a route calculation unit 134, and the like. The purpose management unit 131 controls the setting and change of the action purpose for each agent, and the change of the movement action and the communication action accompanying the change of the action purpose. The purpose management unit 131 also uses the route calculation unit 134 to generate an agent movement route. Each agent moves using a place that satisfies the action purpose set for each agent as the destination.

  The movement action unit 132 determines the movement action of the agent during the simulation time.

  The communication behavior unit 133 determines the communication behavior of the agent.

  The route calculation unit 134 calculates a travel route to the destination corresponding to the action purpose set for the agent.

  The environment model 14 is a mechanism for simulating an event space on the simulation and returning results for the mobile behavior and communication behavior performed by the agent. The environment model 14 includes the movement behaviors implemented by the collective behavior model 13 according to the movement range of agents simulating roads and buildings, and the communication range of agents simulating LTE, Wi-Fi (registered trademark), and the like. Determine whether communication behavior is possible.

  The agent moves on a plane as a moving range and a destination candidate expressed in two dimensions in the simulation space. FIG. 5 shows an image related to the movement range of the agent and the destination candidates. The destination candidate means a place that can be the destination of the agent (that is, a place that can satisfy the action purpose of the agent). Information on the movement range and the destination candidate is included in the movement range / destination data D4 stored in the input data storage unit 15.

  The destination candidates are handled as ranges in the two-dimensional coordinate system, and the number of people who can stay areaLimit for each destination candidate is set in advance in the movement range / destination data D4. The number of people means the number of users of mobile terminals, not the number of agents. The destination candidates where the agents corresponding to the number of people who can stay are limited are obstacles when other agents pass through or reach the destination. By doing so, congestion and queues are reproduced. The movement range may be temporarily prohibited from moving according to a preset time. In addition, a type of action purpose that can be processed is set as a task type for the destination candidate. As described above, an action purpose is set for the agent, and the action purpose is achieved by reaching a destination candidate that can be processed and staying for a certain period of time. An agent's action is completed when all action objectives have been achieved.

  Note that the movement range / destination data D4 is set based on the movement range and the destination candidates at the target event venue.

  The environment model 14 includes a mobile environment control unit 141 and a communication environment control unit 142. The movement environment control unit 141 determines whether or not the movement behavior of the agent that the movement behavior unit 132 is to implement is acceptable.

  The communication environment control unit 142 determines whether or not the communication behavior of the agent that the communication behavior unit 133 is to implement is acceptable. Whether or not the communication action is possible is determined based on whether or not the position of the agent is included in a preset communication range.

  FIG. 6 is a diagram illustrating an image related to a communication range. An agent can perform a communication action when the agent is present on a plane that is a two-dimensional representation in the simulation space and serves as a communication range. A communication range ID is set as an identifier for each communication range, and the communication range can be expressed in any plane for each communication method such as Wi-Fi (registered trademark) and LTE (additional types of communication methods are also possible ). Each communication range has an upper limit for upload (UL) / download (DL) traffic that can be used per unit time and an upper limit for the number of connected terminals. The number of connected terminals has already reached the upper limit. New communication cannot be performed within the communication range. Information regarding the communication range is included in the communication range data D3 stored in the input data storage unit 15. In FIG. 6, communication ranges L1, L2, and L3 are illustrated as LTE communication ranges, and W1 is illustrated as a Wi-Fi (registered trademark) communication range. The agent A1 is located at a position included in the communication range L2 and the communication range W1.

  The communication environment control unit 142 manages the number of connected terminals in all communication ranges in the simulation space and the requested bandwidth of the agent.

  The MAS engine 11 controls the entire simulation (there is a commercially available product having an equivalent function, and these can also be used). For example, the MAS engine 11 manages the passage of time in the simulation space, and for each agent, the mobile behavior and communication behavior according to the environment model 14 that simulates the event space on the simulation and the collective behavior model 13 that simulates the behavior of the agent. Is calculated for each unit time by the movement action unit 132 and the communication action unit 133.

  The API 12 exchanges necessary data between the MAS engine 11, the collective behavior model 13 and the environment model 14, and outputs an agent action log L1 as output data. 4, the API 12 includes a data conversion unit 121, an action log output unit 122, and the like. The data conversion unit 121 performs a conversion process on the communication application data D1, the movement range / destination data D4, and the like based on an aggregation number S that is the number of persons aggregated for one agent. The action log output unit 122 outputs an action log L1. The aggregation number S is set in advance and stored in the input data storage unit 15.

  Hereinafter, the process procedure which the simulator apparatus 10 performs is demonstrated. FIG. 7 is a flowchart for explaining an example of a processing procedure executed by the MAS engine.

  At the start of the implementation of event traffic prediction, the MAS engine 11 reads the communication application data D1, the agent data D2, the communication range data D3, and the movement range / destination data D4 from the input data storage unit 15 and in the simulation space. The time t is initialized (t = 0) (S101). In addition, the agent status, which is a parameter indicating the behavioral situation in the simulation of each agent, is initialized to “Before Behavior Start” for all agents. Further, the data conversion unit 121 is called to perform conversion processing on the communication application data D1, the movement range / destination data D4, and the like. Details of the conversion process will be described later.

  Subsequently, the MAS engine 11 sets the status of each agent whose start time is t and the agent status is “Before Action Start” to “being active”, and sets or changes the action purpose of each agent. The purpose management unit 131 is executed (S102). At this time, the agent ID of each agent is input to the purpose management unit 131, and the purpose management unit 131 sets an action purpose for each corresponding agent and determines the movement route of each agent.

  Subsequently, the MAS engine 11 initializes the action flags of all agents whose status is “being active” to 0 (S103). Subsequently, the MAS engine 11 selects one agent whose status is “being active” and whose action flag is 0 (S104). Subsequently, the MAS engine 11 causes the movement action unit 132 to determine the movement action of the selected agent (hereinafter referred to as “target agent”) (S105). At this time, the agent ID of the target agent is input to the movement action unit 132. Subsequently, the MAS engine 11 sets the action flag of the target agent to 1 (S106).

  When steps S104 to S106 are executed for all agents whose status is “being active” and whose action flag is 0 (Yes in S107), the MAS engine 11 determines the communication action of each agent. The communication action unit 133 is executed (S108).

  Subsequently, the MAS engine 11 outputs the output of the action log L1 that is information indicating the movement action and the communication action adopted by each agent whose status is “being active” and whose action flag is 1 to the action of the API 12. The log output unit 122 is executed (S109). Subsequently, the MAS engine 11 advances the current time t in the simulation space by one unit time (S110).

  Steps S102 to S110 are repeated until all agents finish their actions (S111).

  FIG. 8 is a flowchart for explaining an example of a processing procedure executed by the purpose management unit. When the agent is requested from the MAS engine 11 at the time t when the agent starts to act (step S102 in FIG. 7), the purpose managing unit 131 requests the behavioral behavior change from the moving behavior unit 132 as described later. If so, execute the process.

  In step S201, the purpose management unit 131 acquires agent data D2 corresponding to the agent ID input from the request source.

  Subsequently, the purpose management unit 131 determines whether or not the request source is the MAS engine 11 (S202). That is, in step S102 of FIG. 7, it is determined whether or not the process related to the agent that starts the action should be executed.

  When the request source is the MAS engine 11 (Yes in S202), it is included in the transit point (behavior purpose list) included in the agent data D2 of the agent to be processed (hereinafter referred to as “target agent”). Among the action objectives that are not selected, the first action objective is selected in the order of the list of waypoints among the unselected action objectives (S203).

  Subsequently, the destination management unit 131 sets a destination candidate closest to the current location as the destination of the target agent among the destination candidates that can process the selected task type for the action purpose (S204). The first current location (movement start location) of each agent is specified by the start location set in the agent data D2 of each agent.

  Subsequently, the destination management unit 131 sets the start location set in the agent data D2 of the target agent as the movement source, sets the destination selected in step S204 as the movement destination, and moves between the movement source and the movement destination. The route calculation unit 134 is caused to calculate the movement route (S205). Note that the migration source, the migration destination, and the agent ID of the target agent are input to the route calculation unit 134.

  On the other hand, when the movement action unit 132 described later is a request source (No in S202), that is, when a change of action purpose is requested from the movement action unit 132 regarding an agent that has already started action, the purpose management unit 131 indicates whether the current time t has passed the end time set in the agent data D2 of the target agent, or whether the scheduled end time for all the action purposes included in the route points of the target agent has passed. Determination is made (S206).

  If the current time t has passed the end time or the scheduled end time of all the action objectives (Yes in S206), the purpose management unit 131 uses the arrival location set in the agent data D2 of the target agent as the target Set as the destination of the agent (S207).

If the current time t has not passed the end time and the scheduled end time of any action purpose has not passed (No in S206), the purpose management unit 131 sets the agent data D2 of the target agent. Among the action objectives included in the waypoints being made, it is determined whether or not there is an action objective that satisfies all of the following conditions (a) to (c) (S208).
(A) The processing completion flag is 0.
(B) The current time t has passed the scheduled start time.
(C) The current time t has not reached the scheduled end time.

  If there is one or more corresponding behavioral objectives (Yes in S208), the purpose management unit 131 selects any one of the relevant behavioral objectives (S209). When there are a plurality of corresponding action purposes, the action purpose with the lowest priority is selected. When there are a plurality of action objectives with the lowest priority, the action objective closest to the current location is selected from among the plurality of action objectives that can process the corresponding task type. When there are a plurality of corresponding action purposes, the action purpose with the earliest scheduled end time is selected from the plurality of action purposes. When there are a plurality of action objectives with the earliest scheduled end time, for example, one action objective is randomly selected from the plurality of action objectives. The selected action purpose becomes a new action purpose of the target agent.

  Subsequently, the destination management unit 131 sets a destination candidate nearest to the current location of the target agent as a destination of the target agent among destination candidates that can process the task type of the selected behavioral purpose ( S210).

  On the other hand, if there is no corresponding action purpose in step S208 (No in S208), the purpose management unit 131 returns a failure to set the action purpose to the mobile action unit 132 (S211). As will be apparent from the description below, in this case, the movement state of the target agent is maintained in the original state.

  Subsequent to step S207 or S210, the purpose management unit 131 uses the current location of the target agent as the movement source, the destination of the target agent as the movement destination, and inputs the movement source and the movement destination to the route calculation unit 134. The route calculation unit 134 is made to calculate a new movement route of the target agent (S212). The agent ID of the target agent is also input to the route calculation unit 134.

  FIG. 9 is a flowchart for explaining an example of a processing procedure executed by the route calculation unit. The processing procedure of FIG. 9 is called in step S205 or S212 of FIG.

  In step S301, the route calculation unit 134 presets the input movement source and destination (coordinate values indicating each), the movement range and destination information defined in the movement range / destination data D4, and presetting. The route calculation from the movement source to the movement destination is executed based on the determined policy. The preset policy is, for example, searching for a route with the shortest distance. The route calculation (or route search) may be performed using a known method.

  Subsequently, the route calculation unit 134 associates information indicating the calculated movement route (hereinafter referred to as “movement route information”) with the input agent ID, for example, the memory device 103 or the auxiliary storage device 102. (S302). If the travel route information is already associated with the agent ID, the travel route information is overwritten with new travel route information.

  FIG. 10 is a flowchart for explaining an example of a processing procedure executed by the mobile action unit. In response to a request from the MAS engine 11, the mobile behavior unit 132 determines the current location at t + 1 with respect to the current time t in the simulation space at the target agent based on the travel route information of the target agent. Note that the processing procedure of FIG. 10 is called in step S105 of FIG.

  In step S401, the movement behavior unit 132 confirms the movement state of the dynamic attribute of the agent (target agent) associated with the input agent ID. When the movement state is “movement” or “stop” (No in S402), the movement action unit 132 is set in the current location of the target agent and the agent data D2 of the target agent in the current time t in the simulation. Using the speed, a range in which the target agent can move per unit time is calculated (S403). For example, a range of a circle centered on the current location of the target agent and having a radius of speed × unit time is calculated as a movable range.

  Subsequently, the movement action unit 132 temporarily selects a position where the remaining distance is minimum with respect to the route to the destination indicated by the movement route information of the target agent in the movable range (S404). Subsequently, the mobile behavior unit 132 activates the mobile environment control unit 141 and inputs the agent ID of the target agent and the position information (coordinates) indicating the temporarily selected position to the mobile environment control unit 141, and the target agent Position information (coordinates) indicating the final destination of the movement is determined (S405).

  The mobile behavior unit 132 updates the simulation internal time zone and area type of the dynamic attribute of the target agent as necessary at this timing. That is, if the time zone within the simulation of the target agent does not match the time zone of the current time t, the time zone within the simulation of the target agent is updated to the time zone to which the current time t belongs. If the area type of the position related to the confirmed position information is different from the area type of the target agent, the area type of the target agent is updated to the area type of the position related to the position information.

  Subsequently, the movement action unit 132 includes the position indicated by the position information (coordinates) returned from the movement environment control unit 141 in the destination of the target agent or other destination candidates having the same task type as the destination. It is determined whether or not (S406). When the position information is included in the destination or the other destination candidates (Yes in S406), the mobile action unit 132 adds the scheduled residence time of the current action purpose of the target agent to t, The earlier of the scheduled end time of the target action of the target agent is set to the target agent as the stay end time, and the behavior state of the target agent is changed to “stay” (S407).

  If the position information (coordinates) returned from the movement environment control unit 141 is not included in the destination of the target agent or other destination candidates having the same task type as the destination (No in S406), the movement action unit 132, a difference dst_new between the distance dst_old from the current location at the target agent t to the destination of the target agent and the distance dst_new from the location indicated by the location information returned from the mobile environment control unit 141 to the destination of the target agent It is determined whether -dst_old is equal to or greater than a predetermined threshold th_dst (S408).

  When the difference is equal to or greater than the threshold th_dst (Yes in S408), the movement environment control unit 141 sets the movement state of the target agent to “movement” if the movement state of the target agent is “stop” (S409). ).

  When the difference is smaller than the threshold th_dst (No in S408), if the movement state of the target agent is “movement” (Yes in S409), the movement action unit 132 sets the current time t as the stop state start time, The stop state start time is linked to the agent ID of the target agent, and the movement state of the target agent is changed to “stop”. (S410).

  On the other hand, if the movement state of the target agent is “stopped” (No in S410), the moving action unit 132 determines whether the current time t in the simulation space exceeds the stop state start time + stop limit for the target agent. Is determined (S412). If the current time t exceeds the stop state start time + stop limit (Yes in S412), the mobile action unit 132 determines that the target agent has been in the stop state for a long time, and calls the purpose management unit 131 to call the target agent. The action purpose is changed (S413). At this time, the agent ID of the target agent is input to the purpose management unit 131. When the change (setting) of the action purpose is successful (Yes in S414), the mobile action unit 132 changes the action state of the target agent to “movement” (S415).

  On the other hand, when the movement state of the target agent is “staying” (Yes in S402), the moving behavior unit 132 confirms the staying end time associated with the agent ID of the target agent (S416). When the current time t in the simulation space has passed the stay end time (Yes in S417), the mobile behavior unit 132 sets 1 to the processing completion flag for the current behavioral purpose of the target agent (S418). Subsequently, the mobile behavior unit 132 activates the purpose management unit 131 and attempts to reset the behavior purpose of the target agent (S419). If the resetting is successful (Yes in S420), the movement behavior unit 132 changes the movement state of the target agent to “movement” (S421). When the purpose management unit 131 is activated, the agent ID of the target agent is input to the purpose management unit 131.

  FIG. 11 is a flowchart for explaining an example of a processing procedure executed by the communication action unit. The processing procedure of FIG. 11 is called in step S108 of FIG. The communication behavior unit 133 uses the collective behavior model communication application data to select communication applications and generate traffic at each location after the movement behavior unit 132 determines the destinations of all agents at the current time t in the simulation time. Make a decision. The collective behavior model communication application data is generated by the data conversion unit 121 based on the communication application data D1 stored in advance in the input data storage unit 15 and is stored in the auxiliary storage device 102 or the memory device 103, as will be described later. It is remembered. For example, there are as many types of various parameters of the communication application data D1 as there are combinations of agent attributes and dynamic states. Therefore, there are as many types of collective behavior model communication application data as there are combinations of agent attributes and dynamic states. The value of each parameter set in the communication application data D1 is a value per user (per person), whereas the value of each parameter set in the communication application data for collective behavior model is 1 agent. It is a per value.

  In step S501, the communication action unit 133 checks the communication flag associated with the agent ID of each agent. The communication flag is a flag indicating whether or not the agent is communicating, and is set to 0 when the agent starts to act. Note that 0 indicates that communication is not being performed, and 1 indicates that communication is being performed.

  Subsequently, for each agent whose communication flag is 0, the communication behavior unit 133 uses the communication app according to the communication occurrence probability set in the collective behavior model communication app data regarding the attribute and dynamic state of the agent. Is determined (S502). The communication occurrence probability is a probability of performing communication (probability of using any communication application).

  Subsequently, the communication action unit 133 sets a communication flag for the agent to 1 for each agent that is determined to use the communication application in step S502, and the collective action model regarding the attribute and dynamic state of the agent. The communication application to be used by each agent is determined according to the application selection ratio set in the communication application data for use (S503). The application selection ratio is a ratio at which each communication application is selected. For each communication application, the traffic volume communicated by one use and the requested bandwidth per unit time are set in the communication application data for collective behavior model for each DL / UL. For each agent that uses a communication application, the remaining traffic amount is generated as a variable, and the communication that is set in the collective behavior model communication application data for the communication application is communicated with one use (DL / UL) traffic. The amount is set to the remaining traffic amount.

  Subsequently, the communication behavior unit 133 generates a list of combinations of the agent ID of the agent whose communication flag is 1 and the requested bandwidth per unit time (S504). Subsequently, the communication action unit 133 inputs the list generated in step S504 to the communication environment control unit 142, and the list of available bands at the current time t in the simulation space generated by the communication environment control unit 142. Is received (S505). The list of available bandwidths is a list including available bandwidths per unit time assigned to the agent related to each element of the agent ID and requested bandwidth list.

  Subsequently, the communication behavior unit 133 updates the remaining traffic amount of each agent based on the list of available bands received from the communication environment control unit 142 (S506). That is, the remaining traffic volume of each agent is updated by the remaining traffic volume−usable bandwidth × unit time.

  Subsequently, the communication action unit 133 sets 0 to the communication flag of the agent whose remaining traffic amount is 0 or less (S507). As a result, the communication behavior of the agent ends.

  FIG. 12 is a flowchart for explaining an example of a processing procedure executed by the mobile environment control unit. The processing procedure of FIG. 12 is called in step S405 of FIG. The movement environment control unit 141 manages, as coordinate information, information related to the agent movement range and destination candidates expressed in two dimensions, and current position information related to all agents in the simulation space.

  In step S601, the movement environment control unit 141 acquires the current location of the agent (target agent) associated with the input agent ID. Subsequently, the moving environment control unit 141 determines whether there is a prohibited passage on the destination indicated by the temporarily selected position information (coordinates) input from the moving action unit 132 or on the route to the destination from the current location. It is determined whether or not (S602). For example, it is determined whether or not movement is restricted due to the presence of a signal or the like. The range of the traffic prohibited area such as a signal is set in advance with respect to the movement range.

  If there is no corresponding prohibited road zone (Yes in S602), the movement environment control unit 141 determines whether or not the movement destination or the movement range serving as a route to the movement destination is blocked by another agent. Is determined (S603). That is, it is determined whether or not movement is restricted due to congestion or the presence of a matrix. The determination can be performed based on whether or not the current location of another agent is included in the movement range.

  When the movement destination or the movement range serving as a route to the movement destination is not blocked by another agent (Yes in S603), the movement environment control unit 141 determines that the destination candidate of the movement destination is another agent. It is determined whether or not it is blocked by (S604). That is, it is determined whether or not there are restrictions on movement due to use of destination candidates by other agents.

  When the destination candidate is not blocked by another agent (Yes in S604), the movement environment control unit 141 outputs the position information (coordinates) input from the movement action unit 132 to the movement action unit 132 as it is. (S605).

  If the destination candidate is blocked by another agent (No in S604), the mobile environment control unit 141 determines whether there is another destination candidate where the target agent can stay within the movable range. It is determined whether or not (S606). Note that the other destination candidates need to be destination candidates that can process the task type of the current action purpose of the target agent. In addition, the movable range refers to a circular area centered on the current location of the target agent and having a radius as a distance to the destination indicated by the input movement information.

  When there is a corresponding destination candidate (Yes in S606), the movement environment control unit 141 outputs position information (coordinates) of a position where the destination candidate can stay in the destination candidate to the movement action unit 132 (S607). When a plurality of destination candidates are applicable, the nearest destination candidate may be selected from the current location of the target agent.

  On the other hand, when the determination result in step S603, step S604, or step S606 is negative, the movement environment control unit 141 moves the position information (coordinates) of the place closest to the movement destination within the movable range to the movement action. The data is output to the unit 132 (S608). The meaning within the movable range is the same as the meaning in step S606.

  Subsequent to step S605, S607, or S608, the mobile environment control unit 141 updates the current location of the target agent with the position information output to the mobile action unit 132 in step S605, S607, or S608 (S609). .

  FIG. 13 is a flowchart for explaining an example of a processing procedure executed by the communication environment control unit. The processing procedure of FIG. 13 is called in step S505 of FIG.

  In step S701, the communication environment control unit 142 lists, for each agent associated with each agent ID input from the communication action unit 133, a communication range including the current location of the agent, and only one communication range is listed. The connection destination of the selected agent is determined as the communication range. For agents that are not included in any communication range, the communication range is not listed. In addition, a communication range in which there is no vacancy (residual) in the number of connected terminals is not listed.

  Subsequently, when there is an agent that lists a plurality of communication ranges related to different communication methods, the communication environment control unit 142 selects a communication range of a connection destination for each agent according to a preset policy. (S702). The preset policy is, for example, that Wi-Fi (registered trademark) is preferentially connected over LTE.

  Subsequently, the communication environment control unit 142 selects one communication range according to a preset policy for an agent in which a plurality of communication ranges that can be used in the same communication method are listed (S703). The predetermined policy is, for example, that a communication range having a relatively large remaining with respect to the upper limit of the number of connected terminals is preferentially selected.

  In steps S702 and S703, the communication range is preferentially selected for an agent that has already been connected to one of the communication ranges at the time (t-1) one unit time before the current time.

  As a result of execution of steps S701 to S703, for each communication range, a list of agent IDs and request bands of agents connected to the communication range is generated.

  Subsequently, the communication environment control unit 142 may connect in steps S701 to S703 so that the number of agents to be connected is equal to or less than the upper limit value for a communication range in which the number of agents to be connected exceeds the upper limit value of the number of connected terminals. Some of the determined agents are excluded from connection targets (S704). At this time, an agent that is not connected to the communication range is preferentially selected as an exclusion target at a time (t−1) one unit time before the current time t. That is, a situation where an agent that has already been in communication suddenly becomes unable to communicate is avoided.

  Subsequently, the communication environment control unit 142 allocates a bandwidth per unit time to each agent determined to be connected to the communication range for each communication range (S705). Specifically, for each communication range, the upper limit of the bandwidth of the communication range is distributed according to a preset method and assigned as an available bandwidth per unit time of each agent. The preset method is, for example, to apportion the upper limit of the bandwidth of the communication range by the requested bandwidth of each agent connected to the communication range. Bandwidth allocation is performed similarly for both UL and DL.

  Note that the available bandwidth of an agent that has a requested bandwidth but could not connect to the communication range is set to 0 for both DL and UL. That is, the usable bandwidth of the agent not included in any communication range or the agent excluded in step S704 is set to zero.

  Subsequently, the communication environment control unit 142 outputs a list in which the requested bandwidth portion in the combination list of the input agent ID and the requested bandwidth is replaced with the available bandwidth to the communication action unit 133 (S706).

  Next, details of step S109 in FIG. 7 will be described. In step S109, the API 12 mediates the exchange of data among the collective behavior model 13, the environment model 14, the MAS engine 11, and the input data storage unit 15, and outputs the execution result of the MAS engine 11 as the behavior log L1. To do. The action log L1 includes, for each agent, an agent ID, a time, a current location (coordinates), a dynamic state, an ID of a connected communication range, a used communication method, a used communication application, and a communication traffic amount (allocation). The available data) (DL / UL) and the like, and is stored in the output data storage unit 16 every unit time.

  Next, the conversion process executed by the data conversion unit 121 will be described. The conversion process is executed in step S101 in FIG.

First, the data conversion unit 121 calculates the following equation (1) using the stayable number of people areaLimit and the aggregation number S at each destination candidate set in the movement range / destination data D4.
areaLimit '= areaLimit / S (1)
The areaLimit ′ calculated by calculating Expression (1) is used as the number of agents that can stay at each destination candidate on the simulation. The areaLimit is set for each destination candidate. Therefore, areaLimit ′ is calculated for each destination candidate.

  areaLimit 'is used, for example, for determining whether or not an agent can stay in each destination candidate. Specifically, in steps S604 and S606 in FIG. 12, the mobile environment control unit 141 determines whether the destination candidate is areaLimit 'for the destination candidate, whether the destination candidate is blocked by another agent. The determination is based on whether or not the destination candidate is staying.

  Subsequently, the data conversion unit 121 performs three-step conversion on the communication application data D1 to generate collective behavior model communication application data, and the collective behavior model communication application data is stored in the auxiliary storage device 102 or the memory. Store in device 103. The communication application data D1 includes the number of combinations of agent attributes and dynamic states for the application usage probability, the traffic volume that is communicated by using each communication application once, the requested bandwidth per unit time of each communication application, and the like. Only the value per person is set. The collective behavior model communication application data is data in which a result obtained by converting a value per person set in the communication application data D1 into a value per agent is set. Therefore, the following conversion processing is executed by the number of combinations of agent attributes and dynamic states. The application usage probability means a probability that a communication application is used, and is a concept corresponding to communication occurrence probability × application selection ratio.

As the first stage (conversion (1)), the data conversion unit 121 sets the communication traffic data (DL / UL) T _DL and T _UL per person that are set in the communication application data D1 for each communication application. Using the requested bandwidth per unit time (DL / UL) B _DL and B _UL , the communication time Δt is calculated by the following equation (2).

Subsequently, the data converting unit 121, Delta] t is the same a number of communication application (app using _{probability: p 1, ..., p a} ; communication traffic _{(DL): T DL_1, ...} , T DL_a; communication traffic (UL ): T _{UL — 1} ,..., T _{UL — a} ) are _grouped as one communication application.

Each parameter is calculated by the following equation, where p ′ is the application use probability of the communication application and Δt ′ (= Δt) is the communication time and T ′ _DL and T ′ _UL are the communication traffic (DL / UL). Is done.
p ′ = (p ₁ +... + p _a )
_{T 'DL = (p 1 T} DL_1 + ... + p a T DL_a) / p'
_{T 'UL = (p 1 T} UL_1 + ... + p a T UL_a) / p'
Note that the first stage (conversion (1)) may be omitted. Whether or not to execute the first stage (conversion (1)) may be set in advance by the user.

As the second stage (conversion (2)), the data conversion unit 121 has a conditional probability that the communication use by the communication application has not occurred until just before the application use probability of each communication application. Convert to communication app usage probability. Here, it is assumed that the total number of communication applications collected through the first stage is n. _P 1 a conditional application using probability in each communication application, ..., _{p n,} the communication time Δt _1, ..., and Δt _n. In the case where the communication application is not used, it is assumed that p ₀ = 1−p ₁ −... −p _n and Δt ₀ = 1. This time, the application used per unit time probability ^{_{p - 1, ..., p -}} n is more than, is calculated by the formula below. Note that p ⁻ is expressed as a symbol in which − is added immediately above p in Equation 2.

なお、単位時間あたりのアプリ使用確率ｐ⁻ _１，…，ｐ⁻ _ｎに対し、発生トラヒック量Ｔ⁻ _ＤＬ、Ｔ⁻ _ＵＬ、及び単位時間あたりの要求帯域Ｂ⁻ _ＤＬ、Ｂ⁻ _ＵＬについては、以下の式で算出される。
Ｔ⁻ _ＤＬ＝Ｂ⁻ _ＤＬ＝Ｂ_ＤＬ
Ｔ⁻ _ＵＬ＝Ｂ⁻ _ＵＬ＝Ｂ_ＵＬ
第３段階（変換（３））として、データ変換部１２１は、単位時間あたりのアプリ使用確率に変換したｎ個の通信アプリの設定に関して、一つの通信アプリとみなしてパラメータをまとめる。パラメータをまとめるアプローチは２通りあり、いずれが採用されるかについては予めユーザによって選択可能である。

Incidentally, the application uses the probability ^p per unit time _- 1, ..., p ^- _n contrast, generated traffic ^T _- ^{DL, T} _{- UL,} and requested bandwidth per unit time ^B _- ^{DL, B} _- for _{the UL,} the following It is calculated by the following formula.
T ^- _DL = B ^- _DL = _BDL
T ^- _UL = B ^- _UL = _BUL
As the third stage (conversion (3)), the data conversion unit 121 regards the settings of n communication applications converted into application use probabilities per unit time and collects parameters as a single communication application. There are two approaches for collecting parameters, and which one is adopted can be selected in advance by the user.

[Approach 1]
An expected value related to the amount of traffic generated per unit time per person is calculated in the same manner as in the first stage based on the parameters calculated in the second stage, and is multiplied by the aggregation number S. The application usage probability per unit time of the collected communication applications is p ⁻ , the communication traffic volume (DL / UL) is T ⁻ ' _DL , T ⁻ ' _UL , and the required bandwidth per unit time (DL / UL) is B ^−. _Assuming that “ _DL , B ⁻ ” _UL , the data conversion unit 121 calculates the value of each parameter by the following equation.
_{^{p - '= (p - 1}} + ... + p - n)
_{^{T - 'DL = B -'}} DL = S (p - 1 T - DL_1 + ... + p - n T - DL_n) / p - '
_{^{T - 'UL = B -'}} UL = S (p - 1 T - UL_1 + ... + p - n T - UL_n) / p - '
The data conversion unit 121 decomposes the application use probability p ⁻ ′ into a communication occurrence probability and an application selection ratio. In approach 1, the application use probability p ⁻ ′ is calculated for a case where a communication applications are grouped into n and n communication applications are grouped into one communication application. Therefore, the application selection ratio of the one communication application is 1. Therefore, the application use probability p ⁻ ′ is used as it is as a communication occurrence probability. Accordingly, the data conversion unit 121, and the communication probability, and application selection ratio (= 1), communication traffic _{^{T - 'DL, T -'}} UL and requested bandwidth B per unit time ^- _'DL, B ^- 'Set _UL and communication application data for collective action model. Such communication application data for collective behavior model is used by the communication behavior unit 133 in the processing procedure of FIG.

[Approach 2]
It is considered that the aggregate number S uses one of n communication apps independently according to the app usage probability per unit time per person, and the application usage probability per unit time, communication traffic volume, and required bandwidth in all combinations Is calculated based on the parameters calculated in the second stage.

The application usage probability per unit time of the collected communication applications is p ⁻ , the communication traffic volume (DL / UL) is T ⁻ ' _DL , T ⁻ ' _UL , and the required bandwidth per unit time (DL / UL) is B ^−. _'DL, B ^-' When _UL, data conversion unit 121, the values of these parameters, the combination _k 0 in the number of people using each communication _application, ..., _k n (k ₀ is using a communication app The number of people who did not do this is used as a variable.
_{^{T - 'DL = B -'}} DL = k 1 T - DL_1 + ... + k n T - DL_n
_{^{T - 'UL = B -'}} UL = k 1 T - UL_1 + ... + k n T - UL_n

なお、ｋ_０，…，ｋ_ｎの組み合わせは、全部で_Ｓ＋ｎＣ_ｎ通りとなる。

_{Incidentally,} k 0, ..., a combination of _{k n} becomes _S + n _{C n} as a total.

Further, the data conversion unit _121, k 0, ..., app uses probabilities for the combination of _{k n} also be calculated by the following equation. Note that p ⁻ is expressed as a symbol in which − is added immediately above p in Equation 4.

アプローチ２において、データ変換部１２１は、ｋ_０，…，ｋ_ｎの組み合わせ_Ｓ＋ｎＣ_ｎだけ通信アプリがあるとみなして、通信発生確率、各通信アプリのアプリ選択比率、通信トラヒック量、要求帯域を集団行動モデル用通信アプリデータに設定する。ここで、通信発生確率は、数４の全通りの算出結果の総和である。各通信アプリの選択比率は、通信アプリごとの数４の算出結果の比率である。斯かる集団行動モデル用通信アプリデータが、図１１の処理手順において通信行動部１３３によって用いられる。

In approach 2, the data conversion unit 121 considers that there are communication applications for only combinations _{S + n} C _{n of} k ₀ ,..., K _n , and sets the communication occurrence probability, the application selection ratio of each communication application, the communication traffic amount, and the required bandwidth. Set in communication application data for collective behavior model. Here, the communication occurrence probability is the sum of all the calculation results of Equation 4. The selection ratio of each communication application is the ratio of the calculation results of Equation 4 for each communication application. Such communication application data for collective behavior model is used by the communication behavior unit 133 in the processing procedure of FIG.

  Next, the results of traffic prediction according to this embodiment will be described below. In the traffic prediction, event traffic generation by 2,000 mobile terminal users is simulated using a user behavior model (a model in which one mobile terminal user is one agent (that is, a model of the aggregate number S = 1)). Compared with the approximate result obtained by simulating 1000 mobile agents using two mobile terminal users as one agent by the collective behavior model.

  Each model (user behavior model, collective behavior model) starts activities from the station after 5:00, stays in the waiting area until 10:00, and has a main purpose task assuming goods purchase in the product sales area after 10:00. A scenario was set in which sub-tasks were performed in order, assuming breaks, cosplay tours, excursions in the product sales area, etc., and when all tasks were completed or the event end time was exceeded, the user moved to the station and ended the action.

  FIG. 14 shows information around the target event venue. Since the aggregate number S in the collective behavior model is 2, the number of coordinates that can be retained in the destination candidate is ½ that in the case where the user behavior model is used.

  There are two types of communication apps: app A that communicates for 3 seconds and app B that communicates for 2 seconds. In the collective behavior model, the communication occurrence probability and the amount of traffic generated by approach 1 conversion of communication app data (3) Is set.

  FIG. 15 is a diagram illustrating the total DL traffic amount when each model is used within the simulation range. From FIG. 15, it can be seen that the approximate result can be calculated with a small number of agents by using the collective behavior model for the simulation result by 2,000 agents using the user behavior model.

  As described above, according to the present embodiment, it is possible to predict the peak value of event traffic that is expected to occur when an event (future event) that has not been conducted so far is targeted at the spatio-temporal level. It becomes.

  As a result, conventionally, in event traffic countermeasures for future events, communication quality deteriorates due to underestimation of the amount of traffic generated at the time of the event, or excessive costs are required for deployment of communication resources due to overestimation. Expected to reduce the possibility of

  In addition, even when the amount of communication resources that can be deployed is insufficient, communication quality degradation can occur on the day of the event by planning in advance the schedule for changing and controlling the communication resource based on the prediction results of event traffic. The effect of suppressing sex is expected.

  Furthermore, in this embodiment, a plurality of people can be aggregated in one agent. That is, the prediction result when the number of persons is n can be approximated by the number of agents of n / m (1 ≦ m <n). Therefore, when targeting a large-scale event, simulating each mobile terminal user as an agent according to the scale of the event reduces the possibility of requiring enormous calculation time and obtaining a predicted value. be able to.

  In the present embodiment, the simulator device 10 is an example of a traffic prediction device. The data conversion unit 121 is an example of a conversion unit. The application selection ratio is an example of a selection ratio. The application usage probability is an example of the usage probability. The auxiliary storage device 102 or the memory device 103 is an example of a storage unit.

  As mentioned above, although the Example of this invention was explained in full detail, this invention is not limited to such specific embodiment, In the range of the summary of this invention described in the claim, various deformation | transformation・ Change is possible.

10 simulator device 11 MAS engine 12 API
13 Collective Action Model 14 Environmental Model 15 Input Data Storage Unit 16 Output Data Storage Unit 100 Drive Device 101 Recording Medium 102 Auxiliary Storage Device 103 Memory Device 104 CPU
105 interface device 121 data conversion unit 122 action log output unit 131 purpose management unit 132 movement behavior unit 133 communication behavior unit 134 route calculation unit 141 movement environment control unit 142 communication environment control unit B bus D1 communication application data D2 agent data D3 communication range Data D4 Movement range / destination data L1 Action log

Claims (8)

A mobile behavior unit that changes the location of a plurality of agents each uniting a plurality of users of communication terminals in a simulation space based on agent data set in advance for each agent;
For each unit time, for each agent, it is determined whether to communicate based on the communication occurrence probability set in the storage unit, and the agent determined to communicate is set in the storage unit for each communication application A communication action unit that determines a communication application to be used based on the selected selection ratio;
For each unit time, for one or more communication ranges in the simulation space, the requested bandwidth of the communication application determined by the communication action unit for each agent whose location is included in the communication range and the communication range are set. A communication environment control unit that determines a bandwidth to be allocated to each agent based on the amount of traffic available per unit time;
A traffic prediction apparatus comprising: For each of the plurality of communication apps, the requested bandwidth per unit time per person and the usage probability per person of the communication app are preset,
The request bandwidth of the unit time per person is converted into the request bandwidth of the unit time per agent, the use probability per person is converted into a communication occurrence probability and a communication application selection ratio per agent, A conversion unit that stores in the storage unit the requested bandwidth of the unit time per agent, the communication occurrence probability per agent, and the communication application selection ratio;
The traffic prediction apparatus according to claim 1. The conversion unit calculates the requested bandwidth of the unit time per agent, the communication occurrence probability per agent, and the selection ratio of the communication app when the plurality of communication apps are combined into one communication app. To
The traffic prediction apparatus according to claim 2, wherein: The conversion unit is configured such that the unit time required bandwidth per agent and the communication per agent for all combinations when each of the plurality of people uses any communication app according to the usage probability per person. Calculating the occurrence probability and the selection ratio of each communication app for each combination;
The traffic prediction apparatus according to claim 2, wherein: A mobile action procedure for changing the location of a plurality of agents each uniting a plurality of users of communication terminals in a simulation space based on agent data preset for each of the agents,
For each unit time, for each agent, it is determined whether to communicate based on the communication occurrence probability set in the storage unit, and the agent determined to communicate is set in the storage unit for each communication application A communication action procedure for determining a communication application to be used based on the selected selection ratio,
For each unit time, for one or more communication ranges in the simulation space, the requested bandwidth of the communication application determined in the communication action procedure for each agent whose location is included in the communication range and the communication range are set. A communication environment control procedure for determining a bandwidth to be allocated to each agent based on the amount of traffic available per unit time;
A traffic prediction method characterized in that the computer executes. For each of the plurality of communication apps, the requested bandwidth per unit time per person and the usage probability per person of the communication app are preset,
The request bandwidth of the unit time per person is converted into the request bandwidth of the unit time per agent, the use probability per person is converted into a communication occurrence probability and a communication application selection ratio per agent, The computer executes a conversion procedure for storing the requested bandwidth of the unit time per agent, the communication occurrence probability per agent, and the communication application selection ratio in the storage unit,
The traffic prediction method according to claim 5, wherein: The conversion procedure calculates the required bandwidth of the unit time per agent, the communication occurrence probability per agent, and the selection ratio of the communication app when the plurality of communication apps are combined into one communication app. To
The traffic prediction method according to claim 6. The conversion procedure includes the required bandwidth of the unit time per agent and the communication per agent for all combinations when each of the plurality of people uses any communication app according to the usage probability per person. Calculating the occurrence probability and the selection ratio of each communication app for each combination;
The traffic prediction method according to claim 6.

Images