JP5522800B2 - Method for selecting communication device for executing predetermined function at density based on value of communication device, communication device and program for the method - Google Patents

Description

  The present invention relates to a method of determining a node that serves as a server from a large number of nodes according to a situation. That is, it does not consist of communication between a single fixed server node and many other client nodes, but focuses on providing applications while dynamically changing servers. As a network in which such a type of application is realized, a multi-hop network such as a P2P network or a sensor network can be cited. Moreover, a CDN (Contents Delivery Network) can be cited as an analogy to the idea of dynamically changing a server.

  The P2P network is a virtual network (also referred to as an overlay network) configured between end hosts constructed on the Internet mechanism. In this virtual network, a free (proprietary) rule protocol that is not bound by Internet restrictions (address, name resolution, etc.) can be defined and operated. In the P2P network, services such as file sharing such as Gnutella and Winny are provided based on communication between end hosts.

  In a sensor network, sensor devices (sensor nodes or simply nodes) that measure the environment are not only used by a single node, but also a large number of devices are distributed in the environment to provide a wide range of environmental information. It is also used for acquisition. At that time, not all of the nodes are operated at the same time. Instead, only the necessary and sufficient number of nodes that cover the measurement area are operated, and the information measured by these nodes is processed. A mechanism to transfer information to an appropriate node is necessary to collect it on the server to perform. Non-Patent Documents 1 and 2 describe routing techniques in these sensor networks.

  CDN (Contents Delivery Network) is a network technology that enables high-speed access to Web content, etc., and copies of content are distributed to servers located as close to the network as possible from users. It is to keep. Since content is distributed across the network, not only is the response to the user faster, but the traffic to transfer the content is also in a local range (that is, between the user and the nearby server holding the replica). ), Which is also useful for avoiding congestion.

  Without providing a management device for managing the entire network for the above network, a communication device that performs a predetermined function is given to the communication device among the communication devices that constitute the network, or In Japanese Patent Application No. 2009-253739, a method for selecting and arranging at a density determined by a value acquired by a communication device has been filed.

  The application is a method for autonomously decentralizing “dynamically arranging nodes operating as servers at regular intervals (one in several nodes) according to circumstances”. Arranging a server means enabling a server function of a node (assuming that all nodes have a server function). The situation is, for example, a load state of a peripheral node. In the application, if the load on the peripheral node is high, the server density (how many nodes are present) is increased accordingly. The lower the density, the lower the density. At this time, a node group (connected graph) in a “similar situation” is considered as one cluster, and the server density is controlled for each cluster. The application does not determine whether each node randomly becomes a server with an appropriate probability. This is because even if the density of the server can be controlled on average, it is always random, and therefore a large spot is produced in the density. The application determines whether each node should be a server in an autonomous and distributed manner, and does not require the existence of a server that controls the whole (a server that controls server density) or a specific initial state of each node, Servers can be arranged at approximately equal intervals.

C. Intanagonwiwat, et al. "Directed Diffusion for Wireless Sensor Networking", IEEE / ACM Transactions on Networking, VOL. 11 No. 1, pp2, 2003 F. Ye, et al. “GRADient Broadcast: A Robust Data Delivery Protocol for Large Scale Sensor Networks”, ACM Wireless Networks, Vol. 11 No. 3, pp. 285-298 (14), 2005

  However, in the above-mentioned application (Japanese Patent Application No. 2009-253739), each node changes the potential variable autonomously and creates a spatial distribution of a plurality of mountain-like potentials, thereby making the “server arrangement” as described above. Realize (the node located at the top of the mountain becomes the server). However, in the application, each node must initialize the value of the potential variable immediately after startup and whenever the situation changes. This can cause the following problems:

  The time at which all nodes are activated is not necessarily the same. When the activation time is greatly shifted for each node, the potential value (absolute value) of each node becomes very small. Therefore, although a spatial distribution of mountain-like potentials is formed, the undulations of the potential may be too small to be identified.

  The change in the situation corresponds to a change in the cluster variable C (the cluster variable C is a parameter for determining the interval between the peaks). In the conventional method, whenever the variable C changes, the value of the potential variable is initialized with a random number. In other words, this means that the potential distribution that has been generated so far is completely discarded and the process is started from the beginning. Therefore, if the variable C changes, any node will abandon the server function once. If the value of the variable C changes frequently, the potential variable (that is, the spatial distribution of the potential) cannot reach a steady state, and the server cannot be determined.

  Therefore, the present invention relies on the initial value of the potential variable of the communication device for the communication device that performs a predetermined function among the communication devices that constitute the network without providing a management device that manages the entire network. It is an object of the present invention to provide a method for arranging the communication device and a communication apparatus and program therefor without any problem.

  In order to achieve the above object, a communication device according to the present invention stores potential variables and auxiliary variables of neighboring devices connected by a communication link, and stores the potential variables and auxiliary variables every time the neighboring device receives the potential variables and auxiliary variables. First means for updating the potential variable and auxiliary variable of the adjacent device, and second means for storing the potential variable, auxiliary variable and density control variable of the own device and transmitting the potential variable and auxiliary variable of the own device to the adjacent device And a third means for updating the potential variable and auxiliary variable of the own device based on the potential variable and auxiliary variable of each adjacent device, and the potential variable, auxiliary variable, and density control variable of the own device by a predetermined calculation formula, Determine whether or not to execute a given function based on the device's potential variable and the potential variable of each neighboring device 4 means, and the fourth means is that one of the conditions for executing the predetermined function is that the potential variable of its own device is larger or smaller than the potential variable of all adjacent devices, The predetermined calculation formula includes a term that is 0 and a random term when the potential variable is equal to or greater than a certain threshold when the potential variable of the device itself is updated.

  Further, the predetermined calculation formula is to distribute the potential variable of each communication device in a network composed of a plurality of the communication devices so that the maximum value and the minimum value repeat, and the maximum value and the minimum value It is also preferable that the repetition density varies depending on the density control variable.

であり、
隣接装置ｎのポテンシャル変数ｕ_ｎ ^ｔ、自装置のポテンシャル変数ｕ_ｉ ^ｔから、自装置の更新後の補助変数ｖ_ｉ ^ｔ＋Ｉｃを求める計算式は、

であり、

であり、δ（ｔ）は−１≦δ（ｔ）≦１を満たす乱数であることも好ましい。 If A, ΔT, Δx, h, and k are positive constants,
From the auxiliary variable v _n ^t of the neighboring device n, the potential variable u _i ^{t of} the own device, the auxiliary variable v _i ^t of the own device, and the density control variable c _i ^t of the own device, the potential variable u _i ^{t + 2Ic} after updating of the own device The calculation formula for

And
From the potential variable u _n ^t of the adjacent device n and the potential variable u _i ^t of the own device, a calculation formula for ^obtaining the updated auxiliary variable v _i ^{t + Ic} of the own device is as follows:

And

It is also preferable that δ (t) is a random number that satisfies −1 ≦ δ (t) ≦ 1.

であり、
隣接装置ｎのポテンシャル変数ｕ_ｎ ^ｔ、自装置のポテンシャル変数ｕ_ｉ ^ｔから、自装置の更新後の補助変数ｖ_ｉ ^ｔ＋Ｉｃを求める計算式は、

であり、

であり、δ（ｔ）は−１≦δ（ｔ）≦１を満たす乱数であることも好ましい。 When A, ΔT, Δx, h, and k are positive constants, two auxiliary variables v _n ^t and w _n ^{t of} the adjacent device n, the potential variable u _i ^{t of} the own device, and two auxiliary variables of the own device ^{Formulas for obtaining} v _i ^t and w _i ^t , the own device's density control variable c _i ^t , the updated potential variable u _i ^{t + 2 Ic} and the auxiliary variable w _i ^{t + 2 Ic} are respectively

And
From the potential variable u _n ^t of the adjacent device n and the potential variable u _i ^t of the own device, a calculation formula for ^obtaining the updated auxiliary variable v _i ^{t + Ic} of the own device is as follows:

And

It is also preferable that δ (t) is a random number that satisfies −1 ≦ δ (t) ≦ 1.

  The fourth means may determine whether the value change caused by updating the potential variable of the device itself is large or small based on a predetermined threshold, and setting the small change as one of the conditions for executing the predetermined function. preferable.

  It is also preferable to determine the density control variable based on the value of the monitoring target inside or outside the device itself.

  In order to achieve the above object, the method according to the present invention is a method for selecting a communication device that performs a predetermined function in a system including a plurality of communication devices each managing a potential variable, an auxiliary variable, and a density control variable. A first step in which each communication device advertises a potential variable and an auxiliary variable to adjacent devices connected by a communication link; and when each communication device receives a potential variable and an auxiliary variable from an adjacent device, If the potential variable and auxiliary variable of the device have already been saved, the potential variable and auxiliary variable of the neighboring device saved are updated, and if not saved, the potential variable and auxiliary variable of the neighboring device have been saved. The second step of saving the variables, and each communication device has the potential variables and auxiliary variables of each neighboring device, And the third step of updating the potential variable and auxiliary variable of the own device with a predetermined calculation formula based on the potential variable, auxiliary variable, and density control variable of each of the communication devices. The fourth step of determining whether or not to execute the predetermined function based on the potential variable of the adjacent device has one of the conditions for determining whether or not to execute the predetermined function in the fourth step. Is greater than or less than the potential variable of all adjacent devices, and the predetermined calculation formula is 0 when the potential variable is equal to or greater than a certain threshold when the potential variable of the own device is updated. Includes terms and random terms.

  Further, the predetermined calculation formula is such that the potential variable distribution of each communication device has a shape in which peaks and valleys repeat, where the value of the potential variable is high, and the density at which the peaks and valleys are generated is It is also preferred to vary with density control variables.

According to the program of the present invention,
A computer is functioned as the communication device.

  In the present invention, a potential spatial distribution equivalent to the existing method is generated without depending on the initial value of the potential variable of each node. Since it does not depend on the initial value, it is only necessary to set the potential variable to an appropriate value such as 0 immediately after startup. Further, it is not necessary to set a random number in the potential variable every time the variable C changes as in the existing method. Therefore, even if the variable C changes, the spatial distribution of potential that has been formed so far is maintained, and continuously changes from there to a new spatial distribution. Accordingly, since the server also moves with the change of the spatial distribution, it is possible to avoid a situation in which the server cannot be determined.

1 shows a communication system according to the present invention. It is the simplified block diagram of the communication apparatus by this invention. It is a figure explaining the flow of a process. It is a figure which shows distribution of a potential variable.

  The best mode for carrying out the present invention will be described in detail below with reference to the drawings. FIG. 1 is a diagram showing a communication system according to the present invention. The communication system according to the present invention includes a plurality of communication apparatuses according to the present invention represented by squares in FIG. In FIG. 1, a line connecting each communication device is a communication link. In the network configured by the communication device of the present invention, for example, a P2P network or a wireless ad hoc network has a network configuration, that is, a communication link between communication devices, or a communication device participating in the network. , May change over time. In the following description, an adjacent device refers to a communication device connected by a communication link when viewed from a certain communication device. That is, the communication device corresponding to the rectangle filled with the vertical line in FIG. 1 is an adjacent device of the communication device corresponding to the rectangle filled with the diagonal line in FIG.

  FIG. 2 shows a simplified configuration of a communication device according to the present invention. As shown in FIG. 2, the communication device according to the present invention includes an adjacent device variable management unit 1, a variable calculation unit 2, an own device variable management unit 3, and an execution determination unit 4. In the first embodiment of the present invention described below, each communication device manages a potential variable u, an auxiliary variable v, and a density control variable (cluster variable) C. Here, the potential variable is a value used by each communication device to determine whether or not to execute a predetermined function, and the auxiliary variable is a variable used for calculating the potential variable. A density control variable is a variable that changes the density of a communication device that performs a predetermined function. Here, in the first embodiment, the density control variable is a value set in advance in each communication device.

  The adjacent device variable management unit 1 holds the potential variable and auxiliary variable of the adjacent device in association with the identifier of the adjacent device. Specifically, the neighboring device variable management unit 1 holds the neighboring device each time it receives a signal including a potential variable and an auxiliary variable of the neighboring device that the neighboring device advertises based on a predetermined schedule. If the potential variable and auxiliary variable are updated to the received values, a communication device and a communication link are newly set, and an advertisement is received from the communication device for the first time, the identifier of the communication device included in the advertisement The potential variable and the auxiliary variable included in the advertisement are stored in association with. If a signal including a potential variable and an auxiliary variable is not received from a certain neighboring device for a predetermined period, it can be considered that the neighboring device is no longer adjacent to the neighboring device. 1 deletes the potential variable and auxiliary variable of the communication device that was the neighboring device.

  The own device variable management unit 3 holds the potential variable, auxiliary variable, and density control variable of the own device, and advertises the potential variable and auxiliary variable of the own device to neighboring devices based on a predetermined schedule.

  The variable calculation unit 2 includes the potential variable and auxiliary variable of each adjacent device held by the adjacent device variable management unit 1, and the potential variable, auxiliary variable, and density control of the own device held by the own device variable management unit 3. Based on the variable, the value of the potential variable and the auxiliary variable of the own device is newly determined at the timing specified by a predetermined schedule. The own device variable management unit 3 updates the potential variable and auxiliary variable of the managed device itself to the values determined by the variable calculation unit 2. The own device variable management unit 3 outputs the value before the update to the execution determination unit 4 before updating the value of the potential variable to the value determined by the variable calculation unit 2, and the execution determination unit 4 Is held as the previous value of the potential variable of its own device.

  The execution determination unit 4 has the potential variable immediately before the own device held at the timing specified by the predetermined schedule, and the latest potential variable of the own device held by the own device variable management unit 3. Then, it is determined whether or not to execute a predetermined function from the potential variable of each adjacent device managed by the adjacent device variable management unit 1. When it is determined that the predetermined function is not executed at the time of determination, the execution of the predetermined function is started when it is determined to be executed, and when the predetermined function is determined not to be executed, Do not start. Similarly, when a predetermined function is being executed at the time of determination, if it is determined to be executed, the predetermined function is continued to be executed, and if it is determined not to be executed, a predetermined function is Stops execution of the function.

  Next, the operation of the system in the first embodiment will be described. In the initial state, each communication device sets a predetermined value, for example, 0 for the potential variable and the auxiliary variable.

  In the present embodiment, the variable calculation unit 2 periodically calculates and updates the potential variable and the auxiliary variable of the own device at the timing of the update interval Ic, and the own device variable management unit 3 sets 1 to m × Ic. Times, the potential variable of the own device is advertised, and the execution determination unit 4 determines execution of a predetermined function once for n × Ic. Note that m and n are integers. FIG. 3 shows the flow of processing when m = 1 and n = 4. Note that the potential variable and the auxiliary variable may be updated at different timings. In other words, the potential variable may be updated at time T + x · Ic (x is an integer), and the auxiliary variable may be updated at time T + x · Ic + Ic / 2.

Next, the update process of the potential variable and auxiliary variable of the device itself in the variable calculation unit 2 will be described. Note that the following formulas (1) and (2) are based on the assumption that the variable u has been calculated at a certain time t. u _i ^t is the value of the variable u of the node calculated by the node i at time t. u _n ^t is the value of the variable u received by the node i from the adjacent node between the time t−Ic and the time t + Ic. v _i ^{t + Ic} is the value of the variable v of the node calculated by node i at time t + Ic. v _n ^t is the value of the variable v received by the node i from the adjacent node between time t and time t + 2Ic. c _i ^t is the value of the density control variable C at time t of the node i. First, the variable calculation unit 2 includes the potential variable u _i ^t , the auxiliary variable v _i ^t , the density control variable c _i ^t held by the own device variable management unit 3, and the adjacent held by the adjacent device variable management unit 1. From the device auxiliary variable v _n ^t , the updated potential variable u _i ^{t + 2Ic} of the ^device is ^obtained by the following equation (1). Thereafter, the own device variable management unit 3 updates the held potential variable u _i ^t to the obtained u _i ^{t + 2Ic} . Subsequently, the variable calculation unit 2 uses the potential variable u _i ^t held by the own device variable management unit 3 and the auxiliary variable u _n ^{t of the} neighboring device held by the neighboring device variable management unit 1 to The updated auxiliary variable v _i ^{t + Ic} is obtained by the following equation (2).

ただし、

であり、δ（ｔ）は−１≦δ（ｔ）≦１を満たす乱数である。また、上式においてΣは、総ての隣接装置について足し合わせることを意味している。なお、上述した例では、式（１）は時刻ｔ＋２Ｉｃにおいて実施される計算、式（２）は時刻ｔ＋Ｉｃにおいて実施される計算を示している。また、Ａは正の定数で拡散係数などと呼ばれる事もある。ΔｘおよびΔＴは空間および時間を離散化するための小さな値である（ΔＴ＝Ｉｃとしてもよい）。これらすべての定数はすべてのノードに共通である。また、Ａ、Δｘ、ΔＴの値のバランスによって、変数ｕの値のネットワーク内での空間分布（つまりポテンシャルの空間分布）がクラスタ内で定常状態になるまでの時間が変わる。

However,

And δ (t) is a random number that satisfies −1 ≦ δ (t) ≦ 1. In the above equation, Σ means that all adjacent devices are added together. In the example described above, equation (1) indicates the calculation performed at time t + 2Ic, and equation (2) indicates the calculation performed at time t + Ic. A is a positive constant and is sometimes called a diffusion coefficient. Δx and ΔT are small values for discretizing space and time (may be ΔT = Ic). All these constants are common to all nodes. Further, depending on the balance of the values of A, Δx, and ΔT, the time until the spatial distribution of the value of the variable u in the network (that is, the potential spatial distribution) becomes a steady state in the cluster changes.

Equation (1) includes terms for function g and function δ. The value to be added is a small value, and a small value is realized by the constants h and k. (H and k are also common to all nodes). The constant X in the definition of the function g is a value that determines the height of the peak of the spatial distribution of potentials to be formed (X is also common to all nodes). The height of the mountain is estimated to be approximately X / sin (Δx / c _i ^t ). Note that h is preferably as small as possible, but if it is too small, it takes time to reach a steady state (h is considered to be 1/10 or less of A is appropriate). Further, k is set to a value sufficiently smaller than G _thresh * 2Ic (for example, 1/10 or less). Note that G _thresh is the same as the threshold value in the execution determination unit 4.

  In the above-mentioned example, the potential variable u of the own device is updated first by the equation (1), and then the auxiliary variable v of the own device is updated by the equation (2) using the updated potential variable. However, the auxiliary variable may be updated first, and then the potential variable may be calculated. However, the order is fixed. That is, if the potential variable is updated first according to the equation (1), the potential variable is updated first at each timing, and then the auxiliary variable is updated according to the equation (2).

ただし、

ｈ、ｋは、小さな正の定数である。 Equations (1) and (2) divide the following equation (3) in the n-dimensional Euclidean space into a partial differential equation of time first-order / space second-order and a simultaneous differential equation of differential equations of second-order space. It is considered that these are discretized expressions with respect to time and space. However, unlike the Euclidean space, the number of adjacent nodes is not fixed in the P2P network (in the n-dimensional Euclidean space, the number of adjacent points is 2n for all points), and each node has variables u and v of all adjacent nodes. And the difference between u and v of its own node are obtained and added (corresponding to the sigma symbol part of Equation (1) and Equation (2)).

However,

h and k are small positive constants.

であり、微分演算子∇^２は、ラプラシアンと呼ばれる微分演算子であり、ｎ次元ユークリッド空間では以下の式（４）の様に定義される。同様に、∇^４は以下の式（５）の様に定義される。したがって、

は以下の式（６）の様に表される。

も、式（５）から式（６）と同様にして求められる。 In equation (3),

The differential operator ∇ ² is a differential operator called Laplacian, and is defined as in the following formula (4) in the n-dimensional Euclidean space. Similarly, ∇ ⁴ is defined as the following equation (5). Therefore,

Is expressed by the following equation (6).

Is also obtained in the same manner as equations (5) to (6).

  When each communication device updates the potential variable and the auxiliary variable according to the equations (1) and (2), the spatial distribution of the potential variable of each communication device is maximized at a density determined by the value of the density control variable C. The value and minimum value repeat. The larger the value of the density control variable C, the lower the density, that is, the longer the repetition cycle. FIG. 4 shows a distribution of potential variables of a line of communication devices when each communication device sets communication links with four communication devices, that is, when the network configuration is in a lattice shape, and becomes a steady state. Is shown. Thus, for example, the communication device E is an adjacent device of the communication devices D and F. If the value of the potential variable is high, FIG. 4 shows a state in which peaks and valleys are generated for every 12 communication devices. In FIG. 4, since the network is in a lattice shape, the potential variable is distributed in a sine wave shape. However, in reality, the distribution is more distorted than that shown in FIG. 4 depending on the communication link setting state. However, regardless of the setting state of the communication link, the potential variable is distributed such that the maximum value and the minimum value repeat, and the density can be controlled by the density control variable C. Note that the time until the steady state is reached is determined by the values of ΔT, Δx, and A.

Next, the determination of whether or not to execute a predetermined function in the execution determination unit 4 will be described. First, the execution determination unit 4, as described above, the front of the potential variables u _i ^t one of the own device that holds the latest potential variables u of the apparatus where the information processing apparatus variable control unit 3 holds and _i ^{t + 2Ic,} based on the update interval Ic potential variables, time rate of change of potential variables of its own device ₌ seeking ^{_{(u i t + 2Ic -u i}} t) / Ic, the absolute value of the determined time rate of change threshold _{G thresh} below It is determined whether or not.

  If it is larger than the threshold value, it can be determined that the distribution of the potential variable is not in a steady state, so that it is determined that the predetermined function is not executed. Note that this threshold should ideally be 0, but in reality it cannot be 0, and is a predetermined value close to 0, for example, (RAND_MAX / Ic) × 0.0001 or the like. Set (RAND_MAX is an appropriate real constant and is set at the time of design). If the threshold is too low, execution of the predetermined function is not easily started. If the threshold is too high, execution of the predetermined function is wastefully started in many communication apparatuses.

When the absolute value of the time change rate is equal to or less than the threshold value, the execution determination unit 4 holds the latest potential variable u _i ^{t + 2Ic of the} own device held by the own device variable management unit 3 and the neighboring device variable management unit 1 comparing the potential variables to which all the neighboring device, if the most recent of the potential variables u _{i t} ^{+ 2Ic} of the apparatus where the information processing apparatus variable control unit 3 is held by the largest, determines to perform a predetermined function, it Otherwise, it is determined that the predetermined function is not executed. In this configuration, when the local apparatus corresponds to the maximum value in the distribution of the potential variable of the network, it is determined to execute the predetermined function. However, when the local apparatus has the minimum value, it is determined to execute the predetermined function. Even in such a case, it may be determined that the predetermined function is executed at both the maximum value and the minimum value.

  In the present embodiment, the communication device that is at the top of the peak in the distribution of potential variables is executing a predetermined function. Thus, the signal may be transmitted to the communication device having the maximum potential variable. When the communication device that has received the signal does not perform the predetermined function, the communication device further transfers the signal to the communication device having the maximum potential variable among the adjacent devices. The signal arrives at the communication device that performs the above function.

  Subsequently, a second embodiment of the present invention will be described. In the first embodiment, only one auxiliary variable v is used, but in the second embodiment, two auxiliary variables v and w are used. Therefore, the neighboring device variable management unit 1 holds the potential variable of the neighboring device and two auxiliary variables, and the own device variable management unit 3 holds the potential variable, two auxiliary variables, and the density control variable of the own device, Based on a predetermined schedule, the potential variable of the own device and two auxiliary variables are advertised to neighboring devices. The variable calculation unit 2 also includes the potential variable and two auxiliary variables of each neighboring device held by the neighboring device variable management unit 1, the potential variable of the own device held by the own device variable management unit 3, 2 Based on the two auxiliary variables and the density control variable, the values of the potential variable of the own device and the two auxiliary variables are newly determined at a timing specified by a predetermined schedule.

  In this embodiment, the variable advertised by each communication device increases, and the amount of data advertised and the amount of calculation of the variable increase. However, when ΔT, Δx, and A are the same, the time for convergence to the steady state is shorter than in the first embodiment. Below, the update method of the potential variable and auxiliary variable in 2nd Embodiment is demonstrated. Other processes are the same as those in the first embodiment.

First, in an initial state, each communication device sets a predetermined value, for example, 0 for two auxiliary variables and a potential variable. It is assumed that the variable u has been calculated at a certain time t. u _i ^t is the value of the variable u of the node calculated by the node i at time t. u _n ^t is the value of the variable u received by the node i from the adjacent node between the time t−Ic and the time t + Ic. v _i ^{t + Ic} is the value of the variable v of the node calculated by node i at time t + Ic. v _n ^t is the value of the variable v received by the node i from the adjacent node between time t and time t + 2Ic. w _i ^t is the value of the variable w of the node calculated by the node i at time t. w _n ^t is the value of the variable w received by the node i from the adjacent node between time t and time t + 2Ic. c _i ^t is the value of the density control variable C at time t of the node i. The variable calculation unit 2 also includes the potential variable u _i ^t held by the own device variable management unit 3, the two auxiliary variables v _i ^t and w _i ^t , the density control variable c _i ^t , and the adjacent device variable management unit 1. The updated potential variables u _i ^{t + 2I} and w _i ^{t + 2I} of the own apparatus are ^calculated from the two auxiliary variables v _n ^t and w _n ^{t of the} adjacent apparatus held by the following expressions (7) and (8), respectively. Ask for. Thereafter, the own apparatus variable management unit 3 ^updates the held potential variable u _i ^t to u _i ^{t + 2Ic} , and further ^updates the held auxiliary variable w _i ^t to w _i ^{t + 2Ic} . Subsequently, the variable calculation unit 2 determines the potential variable held by the own device variable management unit 3 from u _i ^t and the auxiliary variable of the adjacent device held by the adjacent device variable management unit 1 from u _n ^t . The auxiliary variable v _i ^{t + Ic} after the update of the device is obtained by the following equation (9).

Note that A, ΔT, Δx, g (u _i ^t ), δ (t), h and k, and the meaning of Σ is the same as in the first embodiment. Note that, similarly to the first embodiment, the auxiliary variable v is first updated by the equation (9), and then the potential variable and the auxiliary variable w are updated by the equations (7) and (8). good. However, the update order according to equations (7) and (8) and the update order according to equation (9) are the same at each timing. That is, the potential variable u and the auxiliary variable w may be updated at time T + x · Ic (x is an integer), and the auxiliary variable v may be updated at time T + x · Ic + Ic / 2.

  Expressions (7), (8), and (9) are obtained by converting the following expression (10) in the n-dimensional Euclidean space into a simultaneous differential equation of a partial differential equation of time 1st floor and space 2nd floor and a differential equation of space 2nd floor. Can be considered as equations that are discretized in time and space. In the system according to the present invention, unlike the n-dimensional Euclidean space, the number of adjacent communication devices is not constant. Therefore, in equations (7), (8), and (9), the variables of each adjacent device and its own device are used. The difference is integrated.

なお、Ａは、式（１）と同じく正の定数であり、ラプラシアンに関しては第１実施形態と同様である。

Note that A is a positive constant as in equation (1), and Laplacian is the same as in the first embodiment.

  In the two embodiments described above, a predetermined value is set in advance as the density control variable C in each communication device. However, each communication device determines the value of the density control variable C, for example, the load factor of the CPU, the free space of the communication buffer, the environmental value such as the temperature measured by the sensor, and the battery capacity when the communication device is battery driven. It may be a form that dynamically changes depending on the remaining amount, the amount of power generated when the communication device is driven by a solar cell, and the like.

  For example, when the communication device is driven by a solar cell and the density control variable C is determined by the amount of power generation, the communication device that performs a predetermined function at a high density in a region that is in the sun and has a large amount of power generation Is placed in a shaded area where the amount of power generation is small, a communication device that performs a predetermined function at a low density is arranged, and the density is autonomously changed according to the change in the amount of power generation. . Similarly, if the density control variable C is determined by the load factor of the CPU, it is possible to control the density of the communication device that performs a predetermined function according to the load.

  The communication apparatus according to the present invention can be realized by a program that causes a computer to function as each unit in FIG. These computer programs can be stored in a computer-readable storage medium or distributed via a network. Furthermore, the present invention can be realized by a combination of hardware and software.

  Moreover, all the embodiments described above are illustrative of the present invention and are not intended to limit the present invention, and the present invention can be implemented in other various modifications and changes. Therefore, the scope of the present invention is defined only by the claims and their equivalents.

1 Adjacent device variable management unit 2 Variable calculation unit 3 Self device variable management unit 4 Execution determination unit

Claims (9)

The potential variable and auxiliary variable of the neighboring device connected by the communication link are saved, and each time the potential variable and auxiliary variable are received from the neighboring device, the stored potential variable and auxiliary variable of the neighboring device are updated. One means,
A second means for storing the potential variable, auxiliary variable, and density control variable of the own device, and transmitting the potential variable and auxiliary variable of the own device to an adjacent device;
A third means for updating the potential variable and auxiliary variable of the own device based on the potential variable and auxiliary variable of each adjacent device, and the potential variable, auxiliary variable, and density control variable of the own device by a predetermined calculation formula;
A fourth means for determining whether or not to execute a predetermined function based on the potential variable of the own device and the potential variable of each neighboring device;
With
The fourth means is that one of the conditions for executing the predetermined function is that the potential variable of its own device is larger or smaller than the potential variable of all adjacent devices,
The predetermined calculation formula includes a term that is 0 when the potential variable is equal to or greater than a certain threshold when the potential variable of the device is updated, and a random term.
Communication device. The predetermined calculation formula is to distribute the potential variable of each communication device in a network composed of a plurality of the communication devices so that the maximum value and the minimum value repeat,
The repetition density of the maximum value and the minimum value varies depending on the density control variable.
The communication apparatus according to claim 1. If A, ΔT, Δx, h and k are positive constants,
From the auxiliary variable v _n ^t of the neighboring device n, the potential variable u _i ^{t of} the own device, the auxiliary variable v _i ^t of the own device, and the density control variable c _i ^t of the own device, the potential variable u _i ^{t + 2Ic} after updating of the own device The calculation formula for

And
From the potential variable u _n ^t of the adjacent device n and the potential variable u _i ^t of the own device, a calculation formula for ^obtaining the updated auxiliary variable v _i ^{t + Ic} of the own device is as follows:

And

The communication apparatus according to claim 1, wherein δ (t) is a random number satisfying −1 ≦ δ (t) ≦ 1. If A, ΔT, Δx, h and k are positive constants,
From the two auxiliary variables v _n ^t and w _n ^{t of} the neighboring device n, the potential variable u _i ^{t of} the own device, the two auxiliary variables v _i ^t and w _i ^{t of} the own device, and the density control variable c _i ^{t of the} own device ^Formulas for calculating the potential variable u _i ^{t + 2Ic} and auxiliary variable w _i ^{t + 2Ic} after updating the device are respectively

And
From the potential variable u _n ^t of the adjacent device n and the potential variable u _i ^t of the own device, a calculation formula for ^obtaining the updated auxiliary variable v _i ^{t + Ic} of the own device is as follows:

And

The communication apparatus according to claim 1, wherein δ (t) is a random number satisfying −1 ≦ δ (t) ≦ 1. The fourth means determines the magnitude of the change in the value due to updating the potential variable of the own device based on a predetermined threshold, and sets the small change as one of the conditions for executing the predetermined function.
The communication apparatus according to any one of claims 1 to 4. The density control variable is determined by the value of the monitoring target inside or outside the device.
The communication apparatus according to any one of claims 1 to 5. In a system including a plurality of communication devices each managing a potential variable, an auxiliary variable, and a density control variable, a method for selecting a communication device that performs a predetermined function,
A first step in which each communication device advertises potential variables and auxiliary variables to neighboring devices connected by a communication link;
When each communication device receives the potential variable and auxiliary variable from the adjacent device, and already stores the potential variable and auxiliary variable of the adjacent device, the potential variable and auxiliary variable of the stored adjacent device are stored. If the variable is updated and not saved, the second step of saving the potential variable and auxiliary variable of the neighboring device;
Each communication device updates the potential variable and auxiliary variable of its own device with a predetermined calculation formula based on the potential variable and auxiliary variable of each neighboring device and the potential variable, auxiliary variable, and density control variable of its own device according to a predetermined calculation formula. Steps,
Repeat
Each communication device has a fourth step of determining whether or not to execute a predetermined function based on the potential variable of the own device and the potential variable of each adjacent device,
In the fourth step, one of the conditions for determining that the predetermined function is to be executed is that the potential variable is larger or smaller than the potential variables of all adjacent devices.
The predetermined calculation formula includes a term that is 0 when the potential variable is equal to or greater than a certain threshold when the potential variable of the device is updated, and a random term.
Method. The predetermined calculation formula is such that the potential variable distribution of each communication device has a shape in which peaks and valleys repeat, assuming that the value of the potential variable is high.
The density at which the peaks and valleys occur varies according to the density control variable.
The method of claim 7.   A program for causing a computer to function as the communication device according to claim 1.

Images