JP2016014943A - Information processing device and parameter calculation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To shorten processing time from the start of optimization processing of a device to the completion of optimization of the device.SOLUTION: A random parameter generation unit 32 randomly generates parameters for controlling the operation of an optical transmission module. A parameter extraction unit 33 extracts the parameters when the operation of the optical transmission module which operates on the basis of the parameters generated by the random parameter generation unit 32 is within a predetermined range with respect to a target value. A regression model generation unit 34 generates a regression model of the operation of the optical transmission module using the parameters as explanatory variables on the basis of the parameter extracted by the parameter extraction unit 33. An initial value calculation unit 35 calculates initial value parameters of the optical transmission module by using the regression model generated by the regression model generation unit 34.

Description

  The present invention relates to an information processing apparatus and a parameter calculation method.

  In Patent Document 1, a candidate solution of a problem is expressed by an individual that is a gene sequence, and a genetic operation is performed on those individuals according to a genetic algorithm to search for an optimal solution of the problem by updating the generation. The calculation means for calculating the value indicating the diversity of the individual group, and adjusting the parameters of the genetic algorithm according to the value calculated by the calculation means during execution of the genetic algorithm by the search means There is disclosed an optimum solution search apparatus comprising: a determination unit that determines whether or not there is an adjustment unit; and an adjustment unit that dynamically adjusts parameters of a genetic algorithm according to a determination result of the determination unit. Yes.

JP 2000-348010 A

  However, in Patent Document 1, when the value indicating the diversity of the population of the genetic algorithm is calculated and the value indicates that the diversity is large, the parameters of the genetic algorithm are adjusted so that the diversity is reduced. When the value indicating diversity is small, the genetic algorithm parameters are adjusted so that the diversity is large. Therefore, in Patent Document 1, there is a problem that the processing time until the device is optimized after the device optimization process is started cannot be shortened.

  Therefore, an object of the present invention is to provide a technique capable of shortening a processing time until the device is optimized after the device optimization process is started.

  The present application includes a plurality of means for solving at least a part of the above-described problems. Examples of the means are as follows. In order to solve the above problems, an information processing apparatus according to the present invention includes a random parameter generation unit that randomly generates parameters, and an operation of the apparatus that operates based on the parameters generated by the random parameter generation unit. A parameter extraction unit that extracts the parameter when the value is within a predetermined range; and a regression model of the operation of the device using the parameter as an explanatory variable based on the parameter extracted by the parameter extraction unit. A regression model generation unit to generate, and an initial value calculation unit to calculate an initial value parameter of the apparatus using the regression model generated by the regression model generation unit.

  According to the present invention, it is possible to shorten the processing time until the device is optimized after the device optimization process is started.

  Problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.

It is the figure which showed the structural example of the optical transmission module to which the optimization assistance apparatus which concerns on 1st Embodiment is applied. It is the figure which showed the block structural example of the optimization assistance apparatus of FIG. It is the figure which showed the block structural example of the optimization part of FIG. It is a figure explaining particle | grains and a particle group. It is a figure explaining an optimization process. It is the figure which showed the block structural example of the initial value parameter calculation part of FIG. It is a figure explaining extraction of a parameter. It is the flowchart which showed the operation example of the optimization assistance apparatus. It is the figure which showed the block structural example of the optimization assistance apparatus which concerns on 2nd Embodiment. It is the figure which showed the block structural example of the initial value parameter calculation part of FIG. It is the flowchart which showed the operation example of the optimization assistance apparatus. It is the figure which showed the example of the optical transmission system to which the optimization assistance apparatus which concerns on 3rd Embodiment is applied. It is the figure which showed the example of application of the optimization assistance apparatus which concerns on 4th Embodiment.

  Hereinafter, the case where the information processing apparatus of the present invention is applied to an optimization support apparatus will be described.

[First Embodiment]
FIG. 1 is a diagram illustrating a configuration example of an optical transmission module to which the optimization support apparatus according to the first embodiment is applied. The optical transmission module 1 shown in FIG. 1 converts an input signal, which is an electrical signal, into an optical signal, and transmits the optical signal to the measuring device 2 connected by, for example, an optical fiber to observe the signal waveform of the optical transmission module. . As shown in FIG. 1, the optical transmission module 1 includes an optimization support device 3, a driver circuit 4, and a laser diode 5.

  A part of the optical signal output from the laser diode 5 is input to the measuring device 2. The measuring device 2 measures the input optical signal and outputs measurement data of the measured optical signal to the optimization support device 3. The measurement data measured by the measurement device 2 is a characteristic of the optical signal output from the laser diode 5, and is, for example, the optical transmission power, jitter, extinction ratio, or amplitude of the optical signal.

  The optimization support device 3 outputs control data for outputting a desired optical signal from the laser diode 5 to the driver circuit 4 based on the measurement data of the optical signal measured by the measurement device 2. For example, the optimization support apparatus 3 outputs control data for outputting an optical signal having desired optical transmission power, jitter, extinction ratio, and amplitude from the laser diode 5 to the driver circuit 4.

  As will be described in detail below, the optimization support device 3 starts the processing for outputting a desired optical signal from the laser diode 5 (optimization processing of the optical transmission module 1). After the optimization process is started, an initial value parameter for reducing the processing time until the optical transmission module 1 is optimized is calculated.

  For example, the optimization support apparatus 3 performs optimization processing when the power of the optical transmission module 1 is turned on. Alternatively, the optimization support apparatus 3 periodically performs optimization processing when the optical transmission module 1 is operating. Alternatively, the optimization support apparatus 3 performs optimization processing when the communication speed of the optical transmission module 1 is switched by the user while the optical transmission module 1 is operating.

  In addition, although only four characteristics of transmission power, jitter, extinction ratio, and amplitude are shown as a desired optical signal, it is not restricted to this. For example, the desired optical signal may include characteristics such as the frequency and waveform of the optical signal.

  The driver circuit 4 amplifies an input signal, which is an electrical signal, in accordance with the control data output from the optimization support apparatus 3 and outputs the amplified signal to the laser diode 5. For example, the driver circuit 4 amplifies and outputs the input signal so that a desired optical signal is output from the laser diode 5 in accordance with the control data output from the optimization support apparatus 3.

  The laser diode 5 converts the electrical signal output from the driver circuit 4 into an optical signal and outputs the optical signal.

  FIG. 2 is a block diagram showing an example of the block configuration of the optimization support apparatus shown in FIG. As illustrated in FIG. 2, the optimization support apparatus 3 includes a reception unit 11, a holding unit 12, an optimization unit 13, a control data output unit 14, and an initial value parameter calculation unit 15.

  The receiving unit 11 receives measurement data output from the measurement device 2. For example, the receiving unit 11 receives the optical transmission power, jitter, extinction ratio, and amplitude of the optical signal measured by the measuring device 2. The receiving unit 11 outputs the received measurement data to the holding unit 12.

  The holding unit 12 stores measurement data output from the receiving unit 11. The measurement data stored in the holding unit 12 is read by the optimization unit 13 and the initial value parameter calculation unit 15.

  The optimization unit 13 calculates a plurality of parameters based on the measurement data stored in the holding unit 12 so that the operation of the optical transmission module 1 operates at the target value. For example, when the optical transmission power, jitter, and amplitude are reduced due to deterioration of the laser diode 5, the optimization unit 13 calculates a plurality of parameters so that the optical transmission power, jitter, and amplitude satisfy target values. .

  There are a plurality of elements that define the operation of the optical transmission module 1. For example, the elements that define the operation of the optical transmission module 1 include the optical transmission power, jitter, extinction ratio, and amplitude of the optical signal output from the laser diode 5 as described above. For example, the optimization unit 13 calculates one index obtained from the plurality of elements (measurement data), and calculates a plurality of parameters so that the calculated index becomes a target value.

  The optimization unit 13 calculates a plurality of parameters by, for example, a particle group optimization algorithm. The optimization unit 13 outputs the calculated plurality of parameters to the control data output unit 14. The optimization unit 13 may calculate a plurality of parameters by an algorithm other than particle group optimization. For example, a plurality of parameters may be calculated using a genetic algorithm or the like. In the following description, the optimization unit 13 is described as calculating a plurality of parameters by the particle swarm optimization algorithm.

  The control data output unit 14 generates control data based on the plurality of parameters calculated by the optimization unit 13. For example, the control data output unit 14 converts a plurality of parameters into control data that allows the driver circuit 4 to control an input signal. When the driver circuit 4 can directly receive a plurality of parameters calculated by the optimization unit 13 and control the input signal, the control data output unit 14 is not necessary.

  The initial value parameter calculation unit 15 calculates initial values of parameters when the optimization unit 13 starts optimization processing. The optimization unit 13 starts an optimization process using the parameter calculated by the initial value parameter calculation unit 15 as an initial value parameter, and calculates a plurality of parameters for the operation of the optical transmission module 1 to be a target value.

  The functions shown in FIG. 2 can be realized by, for example, an application specific integrated circuit (ASIC) or a field-programmable gate array (FPGA). 2 includes, for example, a CPU (Central Processing Unit) that executes a program for realizing the functions of the optimization support apparatus 3, a ROM (Read Only Memory) that stores the program, and a part of the program. This function can be realized by a RAM (Random Access Memory) or the like that temporarily stores data or temporarily stores calculation data.

  FIG. 3 is a diagram illustrating a block configuration example of the optimization unit in FIG. 2. As illustrated in FIG. 3, the optimization unit 13 includes random number generation units 21 and 22, calculation units 23 to 25, and a movement amount calculation unit 26. As described above, the optimization unit 13 calculates a plurality of parameters by, for example, a particle group optimization algorithm.

  The particle swarm optimization algorithm is a kind of swarm intelligence, and is modeled by a particle swarm having position and velocity in a multidimensional space. The movement amount v ′ of the particle swarm optimization algorithm is expressed by the following equation (1).

  Here, w is an inertia constant, v is a current movement amount of particles, c1 and c2 are fixed values, xp is a particle best value, x is a current position of particles, xg is a group best value, and r1 and r2 are random numbers.

  The optimization unit 13 calculates the movement amount v ′ of each of the plurality of parameters according to the equation (1), and updates the values of the plurality of parameters. Then, the optimization unit 13 calculates a plurality of parameters for the optical transmission module 1 to operate at the target value.

  The random number generator 21 generates a random number r1. The random number generator 22 generates a random number r2. The random number r1 generated by the random number generation unit 21 is output to the calculation unit 24, and the random number r2 generated by the random number generation unit 22 is output to the calculation unit 25.

  The calculation unit 23 receives the inertia constant w and the current movement amount v. The calculation unit 23 multiplies the input inertia constant w and the current movement amount v, and outputs the result to the movement amount calculation unit 26.

  The calculation unit 24 receives the fixed value c1, the current position x, the particle best value xp, and the random number r1. The calculation unit 24 multiplies the value obtained by subtracting the current position x from the particle best value xp by the fixed value c1 and the random number r1, and outputs the result to the movement amount calculation unit 26.

  A fixed value c2, a current position x, a group best value xg, and a random number r2 are input to the calculation unit 25. The calculation unit 25 multiplies the value obtained by subtracting the current position x from the group best value xg by the fixed value c2 and the random number r2, and outputs the result to the movement amount calculation unit 26.

  The movement amount calculation unit 26 adds the values output from the calculation units 23 to 25. As a result, the movement amount calculation unit 26 obtains the movement amount v ′ shown in Expression (1).

  FIG. 4 is a diagram for explaining particles and particle groups. FIG. 4 shows an example in which there are three parameters.

  A to C shown in FIG. 4 indicate types of parameters. For example, each of the optical transmission power, jitter, and extinction ratio of the optical signal output from the laser diode 5 is controlled by the parameters A to C.

  The maximum and minimum shown in FIG. 4 indicate the maximum and minimum values that the parameters A to C can take. The maximum value and the minimum value are specified in advance by the user so as not to take parameter values that cause the measuring apparatus to operate abnormally. For example, the parameter A indicates that the maximum value is “6” and the minimum value is “0”.

  S1 to S4 shown in FIG. 4 indicate particles. FIG. 4 shows an example of four particles. Each of the four particles S1 to S4 has parameters A to C as elements. For example, in the example of FIG. 4, the particle S <b> 1 has a parameter A value “0”, a parameter B value “0”, and a parameter C value “1”.

  A dotted frame A1 shown in FIG. 4 indicates a particle group. In the example of FIG. 4, the particle group is composed of four particles S1 to S4.

  For example, the optimization unit 13 updates each value of the particles S1 to S4 illustrated in FIG. 4 so that the measurement data output from the measurement device 2 achieves the target value based on the equation (1), and the parameter The optimum values of A to C are calculated.

  The functions shown in FIG. 3 can be realized by, for example, an ASIC or FPGA. Alternatively, each unit illustrated in FIG. 3 includes, for example, a CPU that executes a program for realizing the function of the optimization unit 13, a ROM that stores the program, a part of the program, and temporarily stores operation data. The function can be realized by a RAM or the like for storing.

  FIG. 5 is a diagram for explaining the optimization process. The horizontal axis shown in FIG. 5 indicates the number of measurements of the measuring device 2, and the vertical axis indicates the measurement value (measurement data index) of the measuring device 2.

  FIG. 5 shows particles S1 to S4. The optimization unit 13 calculates the movement amount v ′ of the particles S1 to S4 and updates the particles S1 to S4 so that the measurement data output from the measurement device 2 approaches the target value. In the example of FIG. 5, the particles S <b> 1 to S <b> 4 and the particle group thereof approach the target value as the number of measurements increases (some particles S <b> 1 to S <b> 4 may be updated so as not to approach the target value. is there).

  When one particle among the particles S1 to S4 reaches the target value, the optimum parameters A to C of the optical transmission module 1 are obtained. For example, if the particle indicated by the arrow A <b> 11 in FIG. 5 is obtained by the optimization unit 13, the parameters A to C included in the particle are the optimum parameters of the optical transmission module 1. The white circle particles S1 'to S4' shown in FIG. 5 will be described later.

  Hereinafter, in order to simplify the description, the plurality of parameters will be described as three parameters A to C. Parameters A to C take integer values and take the maximum and minimum ranges shown in FIG. The plurality of elements that define the operation of the optical transmission module 1 are optical transmission power, jitter, and extinction ratio. In addition, it is assumed that each of the plurality of elements is controlled based on a plurality of parameters A to C. For example, each of optical transmission power, jitter, and extinction ratio is controlled depending on (related to) a plurality of parameters A to C.

  FIG. 6 is a diagram illustrating a block configuration example of the initial value parameter calculation unit of FIG. As shown in FIG. 6, the initial value parameter calculation unit 15 includes a random number generation unit 31, a random parameter generation unit 32, a parameter extraction unit 33, a regression model generation unit 34, and an initial value calculation unit 35. ing.

  The random number generator 31 generates a random number and outputs the generated random number to the random parameter generator 32.

  The random parameter generation unit 32 generates random value parameters A to C based on the random number generated by the random number generation unit 31. The random parameter generation unit 32 outputs the generated random value parameters A to C to the control data output unit 14.

  The control data output unit 14 receives each of the parameters A to C having random values and converts them into control data that can be controlled by the driver circuit 4 to control the optical transmission power, jitter, and extinction ratio. The control data of the control data output unit 14 is output to the driver circuit 4, whereby the laser diode 5 outputs optical signals corresponding to random parameters A to C. The optical transmission power, jitter, and extinction ratio of the optical signal output from the laser diode 5 are measured by the measuring device 2. Measurement data measured by the measuring device 2 is stored in the holding unit 12.

  Based on the measurement data for the random parameters A to C stored in the holding unit 12, the parameter extraction unit 33 uses the parameters A to when the operation of the optical transmission module 1 is within a predetermined range with respect to the target value. Extract C.

  FIG. 7 is a diagram for explaining parameter extraction. In FIG. 7, the horizontal axis indicates the number of times the optical signal is measured by the measuring device 2, and the vertical axis indicates the measurement value (measurement data index) obtained by the measuring device 2 measuring the optical signal. Black circles shown in FIG. 7 indicate measured values of the optical signal with respect to random parameters A to C.

  FIG. 7 shows an example of measured values when the measuring apparatus 2 performs the optical signal measurement for random parameters A to C “n times”. In this case, the random parameter generation unit 32 generates “n” random parameters A to C. That is, the measuring apparatus 2 measures “n” optical signal measurement data for “n” random parameters A to C.

  As described above, the parameter extraction unit 33 extracts random parameters A to C when the operation of the optical transmission module 1 is within a predetermined range with respect to the target value. For example, the parameter extraction unit 33 extracts random parameters A to C when the operation of the optical transmission module 1 (measured value of the optical signal) is within the range of “Y” with respect to the target value. More specifically, the parameter extraction unit 33 extracts parameters A to C corresponding to the measurement values existing in the frame of the one-dot chain line A21 in FIG.

  Returning to the description of FIG. The regression model generation unit 34 generates a regression model for the operation of the optical transmission module 1 using the parameters A to C as explanatory variables based on the parameters A to C extracted by the parameter extraction unit 33. That is, the regression model generation unit 34 generates a regression model using the optical transmission power, jitter, and extinction ratio of the optical signal output from the laser diode 5 as objective variables and parameters A to C as explanatory variables.

  The general formula of the regression model is represented by the following formula (2) assuming a linear formula, for example.

Here, EO in Expression (2) indicates an evaluation value, x _i indicates a parameter to be optimized, α _i indicates a correlation coefficient, and ε indicates a correction coefficient.

  From Equation (2), for example, regression models in the optical transmission power, jitter, and extinction ratio of the optical transmission module 1 are expressed by the following Equations (3) to (5).

Here, “P” in Expression (3) indicates optical transmission power, “A, B, C” indicates parameter values of parameters A to C, “a _{1 to} a ₃ ” indicate correlation coefficients, “Ε ₁ ” indicates a correction coefficient. In equation (4), “J” represents jitter, “A, B, C” represents parameter values of parameters A to C, “b _{1 to} b ₃ ” represents correlation coefficients, and “ε ₂ ” represents A correction coefficient is shown. In equation (5), “E” represents the extinction ratio, “A, B, C” represents the parameter values of the parameters A to C, “c _{1 to} c ₃ ” represents the correlation coefficient, and “ε ₃ ”. Indicates a correction coefficient.

  For example, the regression model generation unit 34 performs multiple regression based on the parameters A to C extracted by the parameter extraction unit 33 and measured values (optical transmission power, jitter, and extinction ratio) corresponding to the parameters A to C. Analysis is performed to calculate the correlation coefficients of Expressions (3) to (5), and a regression model of optical transmission power, jitter, and extinction ratio is calculated.

  The initial value calculation unit 35 calculates the initial value parameter of the optical transmission module 1 using the regression model generated by the regression model generation unit 34. That is, the initial value calculation unit 35 calculates an initial value parameter when the optimization unit 13 starts the optimization process.

  The initial value calculating unit 35 calculates initial value parameters A to C by substituting the target value of the operation of the optical transmission module 1 into the objective variable of the regression model. That is, the initial value calculation unit 35 calculates initial value parameters by substituting the target values of the transmission power, jitter, and extinction ratio of the optical signal output from the laser diode 5 into the objective variables of the regression model.

For example, the target values of the transmission power, jitter, and extinction ratio of the optical signal output from the laser diode 5 are “P _o ”, “J _o ”, and “E _o ”, respectively. In this case, the initial value calculation unit 35 sets the target values “P _o ” and “J _o ” to the objective variables “P”, “J”, and “E” of the regression models of the equations (3) to (5). ”And“ E _o ”are substituted, and the three simultaneous equations with the parameters A to C as variables are solved.

  The initial value calculation unit 35 outputs the solutions of the ternary simultaneous equations to the optimization unit 13 as initial value parameters A to C. The optimization unit 13 generates, for example, initial values of the particles S1 to S4 based on the initial value parameters A to C calculated by the initial value calculation unit 35, and performs an optimization process.

  Here, as described above, the parameter extraction unit 33 extracts the parameters A to C when the operation of the optical transmission module 1 is close to the target value from the random parameters A to C, and the regression model generation unit 34. Generates a regression model based on parameters A to C when close to the target value. That is, the regression model generation unit 34 generates a regression model when the operation of the optical transmission module 1 is close to the target value. Then, the initial value calculation unit 35 calculates, as an initial value parameter, a parameter for the operation of the optical transmission module 1 to be a target value from the regression model. Therefore, the initial value parameter calculated by the initial value calculator 35 is a value close to the target value for the operation of the optical transmission module 1.

  For example, the particles S <b> 1 ′ to S <b> 2 ′ illustrated in FIG. 5 indicate the initial values of the particles based on the initial value parameters A to C calculated by the initial value calculation unit 35. As shown in FIG. 5, since the particles S1 ′ to S2 ′ are closer to the target value than the particles S1 to S4, the optimization process of the optimization unit 13 can calculate particles that reach the target value in a short time. .

  The functions of each unit shown in FIG. 6 can be realized by, for example, an ASIC or FPGA. Alternatively, each unit illustrated in FIG. 6 includes, for example, a CPU that executes a program for realizing the function of the initial value parameter calculation unit 15, a ROM that stores the program, a part of the program, and temporarily stores operation data. The function can be realized by a RAM or the like that is temporarily stored.

  FIG. 8 is a flowchart showing an operation example of the optimization support apparatus. The flowchart shown in FIG. 8 is periodically executed, for example, when the power of the optical transmission module 1 is turned on or when the optical transmission module 1 is operating. Alternatively, the flowchart illustrated in FIG. 8 is executed, for example, when the communication speed is switched by the user when the optical transmission module 1 is operating. For example, the flowchart shown in FIG. 8 is executed when the communication speed is switched from 25G to 10G by the user, or when the communication speed is switched from 10G to 25G.

  First, for example, the initial value parameter calculation unit 15 uses the random parameters A to C to measure the number of measurements of the optical signal, and target values “Po”, “Jo”, optical transmission power, jitter, and extinction ratio target values. And setting conditions, such as "Eo", are received from a user via the input device which is not illustrated (step S1).

  Next, the random parameter generation unit 32 generates parameters A to C having random values (step S2). In step S1, the random parameter generation unit 32 generates random value parameters A to C by the number of times the initial value parameter calculation unit 15 has received from the user. For example, if the number of measurements is “n”, the random parameter generation unit 32 generates n parameters A to C. Thereby, the measuring apparatus 2 can measure n measurement values (measurement data) for n random parameters A to C, for example, as shown in FIG. The measuring device 2 stores the measured values measured in the holding unit 12.

  Next, the parameter extraction unit 33 is based on the measurement data for the random parameters A to C stored in the holding unit 12 when the operation of the optical transmission module 1 is within a predetermined range with respect to the target value. Parameters A to C are extracted (step S3). For example, the parameter extraction unit 33 extracts the parameters A to C of the measurement values within the frame of the alternate long and short dash line A21 illustrated in FIG.

  Next, the regression model generation unit 34 generates a regression model for the measurement value of the optical transmission module 1 using the parameters A to C as explanatory variables based on the parameters A to C extracted by the parameter extraction unit 33 ( Step S4). For example, the regression model generation unit 34 generates a regression model of optical transmission power, jitter, and extinction ratio shown in Expressions (3) to (5) by multiple regression analysis.

  Next, the initial value calculation unit 35 calculates the initial value parameter of the optical transmission module 1 using the regression model generated by the regression model generation unit 34 (step S5). For example, the initial value calculation unit 35 obtains the target values “Po”, “Jo”, and “Eo” of the transmission power, jitter, and extinction ratio received from the user by the initial value parameter calculation unit 15 in step S1. Substituting into the objective variables “P”, “J”, and “E” in the equations (3) to (5), the ternary simultaneous equations with the parameters A to C as variables are solved. Then, the initial value calculation unit 35 sets the obtained parameters A to C as initial value parameters A to C, and outputs them to the optimization unit 13.

  Next, the optimization unit 13 sets the initial value parameters A to C calculated by the initial value calculation unit 35 as initial values of parameters for controlling the optical transmission module 1 (step S6). For example, the optimization unit 13 sets initial values of the particles S1 to S4 based on the initial value parameters A to C calculated by the initial value calculation unit 35 when performing the optimization process using the particle group optimization algorithm. .

  Next, when the optimization unit 13 proceeds from step S6 to step S7, the optimization unit 13 starts optimization processing of the optical transmission module 1 based on the initial value set in step S6 (step S7). Further, the optimization unit 13 determines in step S8 that the measurement data output from the measurement device 2 has not achieved the target value and moves to the step (“No” in step S8). The optical transmission module 1 is optimized (step S7).

  Next, the optimization unit 13 determines whether or not the measurement data output from the measurement device 2 has achieved the target values “Po”, “Jo”, and “Eo” received from the user in step S1. (Step S8). For example, the optimization unit 13 determines whether or not the particles have reached a target value as indicated by an arrow A11 in FIG. That is, the optimization unit 13 determines whether or not the operation of the optical transmission module has reached the target value by the optimization process. If the optimization unit 13 determines that the measurement data output from the measurement device 2 has reached the target value (“Yes” in step S8), the optimization unit 13 ends the process of the flowchart. When the optimization unit 13 determines that the measurement data output from the measurement device 2 has not reached the target value (“No” in step S8), the optimization unit 13 proceeds to the process of step S7.

  As described above, the random parameter generation unit 32 of the optimization support apparatus 3 randomly generates the parameters A to C, and the parameter extraction unit 33 operates the optical transmission module 1 that operates based on the random parameters A to C. Extract parameters A to C when the value is within a predetermined range with respect to the target value. And the regression model production | generation part 34 produces | generates the regression model of operation | movement of the optical transmission module 1 which used the parameters AC as explanatory variables based on the parameters AC extracted by the parameter extraction part 33, and made initial value The calculation unit 35 calculates the initial value parameter of the optical transmission module 1 using the generated regression model. Accordingly, the operation of the optical transmission module 1 with respect to the initial value parameter is close to the target value, and the processing time until the optical transmission module 1 is optimized after the optimization processing of the optical transmission module 1 is started is reduced. It can be shortened.

Note that the regression model generation unit 34 may generate a regression model by removing parameters that are less relevant to the operation of the optical transmission module 1. For example, it is assumed that the parameter C has low relevance to the control of optical transmission power. In this case, the regression model generation unit 34 generates a regression model based on the parameters A and B excluding the parameter C. Specifically, the regression model generation unit 34 sets the correlation coefficient a ₃ in the equation (3) to “0”. The parameter with low relevance is set (instructed) by the user via the input device in step S1 of FIG. 8, for example.

  Moreover, although the regression model production | generation part 34 assumed the linear form as a regression model in the above, it is not restricted to this. For example, the regression model may be a quadratic expression or an exponential function.

  In the above description, three objective variables (“P”, “J”, and “E”) and three explanatory variables (parameters A to C) have been described as examples, but the present invention is not limited to this. For example, there may be n objective variables and n explanatory variables. In this case, the initial value calculation unit 35 calculates the initial value parameter by solving the n-ary simultaneous equations.

In the above description, the optimization unit 13 calculates one parameter obtained from a plurality of elements (measurement data), and calculates a plurality of parameters so that the calculated index becomes a target value. However, a plurality of parameters may be calculated such that a plurality of target values are obtained. For example, the optimization unit 13 has measured data of optical transmission power, jitter, extinction ratio, and amplitude,
A plurality of parameters may be calculated so as to be target values set for the optical transmission power, jitter, extinction ratio, and amplitude.

  In the above description, the optical transmission module 1 has the optimization support device 3. However, the optimization may be outside the optical transmission module 1.

  The optimization support apparatus 3 can also be applied to an optical reception module that receives an optical signal. For example, the measurement apparatus inputs a part of the optical signal received by the optical reception module and measures the input optical signal. Based on the optical signal measured by the measurement device, the optimization support device 3 controls the electrical signal that has been subjected to O / E conversion to be a desired electrical signal, and analyzes the variation factors of the received optical signal. To do.

  Moreover, although the example which applied the optimization assistance apparatus 3 to the optical transmission module 1 was demonstrated above, it is not restricted to this. For example, the optimization support device 3 can be applied to a device whose parameters are controlled. For example, the present invention can be applied to a module that outputs a signal such as an electric signal or an infrared signal that is parameter-controlled.

[Second Embodiment]
Next, a second embodiment will be described. In the first embodiment, the number of simultaneous equations (the number of elements defining the operation of the optical transmission module 1) and the number of parameters are the same. For example, the number of multiple elements of optical transmission power, jitter, and extinction ratio is “3”, and the number of parameters A to C is “3”. Therefore, the initial value calculation unit 35 was able to obtain a solution of simultaneous equations (initial value parameter).

  However, there are generally several tens of parameters of the optical transmission module 1, and there are cases where the number of a plurality of elements that define the operation of the optical transmission module 1 is smaller than the number of parameters. For example, for six elements of optical transmission power, jitter, extinction ratio, amplitude, frequency, and waveform, there may be several tens of parameters that control these elements. In this case, the initial value calculation unit 35 cannot obtain a solution of the simultaneous equations.

  Thus, in the second embodiment, the initial value parameter can be obtained even when the number of parameters is larger than the number of elements that define the operation of the optical transmission module 1. In the following, in order to simplify the explanation, it is assumed that there are three elements of optical transmission power, jitter, and extinction ratio, and there are six parameters A to F that control them. To do.

  FIG. 9 is a diagram illustrating a block configuration example of the optimization support apparatus according to the second embodiment. In FIG. 9, the same components as those in FIG. The optimization unit 41 and the initial value parameter calculation unit 42 illustrated in FIG. 9 have the same functions as the optimization unit 13 and the initial value parameter calculation unit 15 illustrated in FIG. 2 except for the following points.

  For example, when the power of the optical transmission module 1 is turned on, the optimizing unit 41 transmits the setting initial value parameters A to F set in advance by the user to the initial value parameter calculating unit 42 instead of the control data output unit 14. Output.

  The initial value parameter calculation unit 42 generates a regression model as in the first embodiment. The initial value parameter calculation unit 42 substitutes the set initial value parameters A to F output from the optimization unit 41 into the generated regression model, and calculates the optical transmission power, jitter, and extinction ratio of the optical signal. The initial value parameter calculation unit 42 outputs the optical transmission power, jitter, and extinction ratio of the optical signal calculated using the regression model to the optimization unit 41.

  The optimization unit 41 updates the parameters A to F so that the optical transmission power, jitter, and extinction ratio output from the initial value parameter calculation unit 42 reach the target values. For example, the optimization unit 41 updates the parameters A to F with a particle swarm optimization algorithm. The optimization unit 41 outputs the updated parameters A to F to the initial value parameter calculation unit 42.

  The initial value parameter calculation unit 42 substitutes the parameters A to F updated by the optimization unit 41 into the regression model, and calculates the optical transmission power, jitter, and extinction ratio for the updated parameters A to F. The initial value parameter calculation unit 42 outputs the calculated optical transmission power, jitter, and extinction ratio of the optical signal to the optimization unit 41.

  The optimization unit 41 and the initial value parameter calculation unit 42 repeat the above processing. That is, the optimization unit 41 performs optimization processing so that the measurement value calculated by the regression model of the initial value parameter calculation unit 42 reaches the target value, not the measurement data output from the measurement device 2. When the optical transmission power, jitter, and extinction ratio according to the regression model reach the target values of the optical transmission module 1, the optimization unit 41 sets the parameters A to F as initial value parameters. Then, the optimization unit 41 sets the obtained initial value parameter as the initial value of the parameter of the optimization process based on the actual measurement of the optical signal of the optical transmission module 1, and starts the optimization process based on the actual measurement of the optical signal.

  FIG. 10 is a diagram illustrating a block configuration example of the initial value parameter calculation unit of FIG. In FIG. 10, the same components as those in FIG. The regression model generation unit 42a illustrated in FIG. 10 has the same function as the regression model generation unit 42a illustrated in FIG. 6, except for the following points. In the second embodiment, the initial value calculation unit 35 does not calculate an initial value parameter.

  The regression model generation unit 42 a receives parameters A to F from the optimization unit 41. The regression model generation unit 42a substitutes the parameters A to F received from the optimization unit 41 into the regression model, and optimizes the regression model values (optical transmission power, jitter, and extinction ratio) in the substituted parameters A to F. To the conversion unit 41. That is, the regression model generation unit 42 a transmits simulated measurement data (simulated measurement data) of the optical transmission module 1 based on the regression model to the optimization unit 41.

  FIG. 11 is a flowchart illustrating an operation example of the optimization support apparatus. The flowchart shown in FIG. 11 is periodically executed, for example, when the power of the optical transmission module 1 is turned on or when the optical transmission module 1 is operating. Alternatively, the flowchart illustrated in FIG. 11 is executed, for example, when the communication speed is switched by the user when the optical transmission module 1 is operating. Specifically, the flowchart shown in FIG. 11 is executed when the communication speed is switched from 25G to 10G by the user, or when the communication speed is switched from 10G to 25G.

  The processing in steps S11 to S14 is the same as the processing in steps S1 to S4 in the flowchart shown in FIG.

  The optimization unit 41 determines in step S17 that the simulated measurement data based on the regression model has not reached the target value, and proceeds to the process of step S15 (in the case of “No” in step S17). ), Parameters A to F are updated by optimization processing based on simulated measurement data of the regression model, and output to the regression model generation unit 42a (step S15). The optimization unit 41 outputs the set initial value parameters A to F set in advance by the user to the regression model generation unit 42a when the process proceeds from the process of step S14 to the process of step S15. (Step S15).

  Next, the regression model generation unit 42a substitutes the parameters A to F output from the optimization unit 41 for the regression model generated in step S14, and calculates simulated measurement data (step S16). The regression model generation unit 42 a outputs the calculated simulated measurement data to the optimization unit 41.

  Next, the optimization unit 41 determines whether or not the simulated measurement data calculated by the regression model generation unit 42a in step S16 has reached a target value for the operation of the optical transmission module 1 (step S17). When the simulated measurement data calculated by the regression model generation unit 42a reaches the target value of the operation of the optical transmission module 1 (“Yes” in step S17), the optimization unit 41 proceeds to the process of step S18. If the simulated measurement data calculated by the regression model generation unit 42a has not reached the target value of the operation of the optical transmission module 1 ("No" in step S17), the optimization unit 41 proceeds to the process of step S15. .

  When the optimization unit 41 determines in step S17 that the simulated measurement data calculated by the regression model generation unit 42a has reached the target value of the operation of the optical transmission module 1 (“Yes” in step S17), The parameters A to F that have reached the target value by the regression model calculated in step S15 are calculated as initial value parameters (step S18).

  The processing from step S19 to step S21 is the same as the processing from step S6 to S8 in the flowchart shown in FIG. Note that “optimization processing by actual measurement” in step S20 indicates that the optimization processing is performed based on the measurement data of the optical signal measured by the measurement apparatus 2.

  Thus, when the initial value calculation unit 35 cannot calculate the initial value parameter, the optimization unit 41 sets the parameters A to F so that the value of the regression model generated by the regression model generation unit 42a reaches the target value. And parameters A to F when the value of the regression model reaches the target value are used as initial value parameters. As a result, the optimization support apparatus 3 can calculate the initial value parameter even when the number of parameters is larger than the number of elements that define the operation of the optical transmission module 1.

  Further, the optimization unit 41 calculates an initial value parameter by calculation based on the regression model generated by the regression model generation unit 42a, not the actual measurement value (measurement data output from the measurement device 2). Thereby, the optimization part 41 can calculate an initial value parameter, without requiring time.

[Third Embodiment]
Next, a third embodiment will be described. In the first embodiment, the measurement device 2 is provided outside the optical transmission module 1, but in the third embodiment, the optical reception module has a measurement device. The optimization support apparatus controls the optical signal transmitted from the optical transmission module so that the optical signal received by the optical reception module becomes a desired optical signal.

  FIG. 12 is a diagram illustrating an example of an optical transmission system to which the optimization support apparatus according to the third embodiment is applied. In FIG. 12, the same components as those in FIG.

  As illustrated in FIG. 12, the optical transmission system includes an optical transmission module 50 and an optical reception module 60. The optical transmission module 50 includes an optimization support device 51.

  The optimization support device 51 receives measurement data output from the measurement device 63 of the optical reception module 60. Based on the received measurement data, the optimization support device 51 controls the driver circuit 4 so that the optical signal received by the optical reception module 60 becomes a desired optical signal. The optimization support apparatus 51 is different from the optimization support apparatus 3 shown in FIG. 1 in that the optical signal received by the optical reception module 60 is controlled to be a desired optical signal. Is the same as the optimization support apparatus 3.

  The optical receiving module 60 includes a photodiode 61, a receiving circuit 62, and a measuring device 63.

  The photodiode 61 receives the optical signal transmitted from the optical transmission module 50 and converts the received optical signal into an electrical signal. The photodiode 61 outputs the converted electric signal to the receiving circuit 62 and the measuring device 63.

  The receiving circuit 62 amplifies the electrical signal output from the photodiode 61, performs a predetermined process, and outputs the amplified signal to a subsequent circuit (not shown).

  The measuring device 63 measures the electrical signal output from the photodiode 61 and outputs measurement data of the measured electrical signal to the optimization support device 51 of the optical transmission module 50. The measurement data measured by the measurement device 63 is, for example, the characteristics of the optical signal received by the photodiode 61, and includes the optical reception power, jitter, extinction ratio, or amplitude of the optical signal.

  As described above, the optimization support apparatus 51 generates a regression model based on the optical signal on the receiving side, and shortens the processing time from the start of the optimization process of the optical transmission system to the optimization of the optical transmission system. can do.

  In the above description, the optical transmission module 50 includes the optimization support device 51. However, the optical reception module 60 may include the optimization support device 51.

  In the above description, the measurement device 63 measures the electrical signal output from the photodiode 61, but may measure the optical signal received by the photodiode 61.

[Fourth Embodiment]
Next, a fourth embodiment will be described. In the fourth embodiment, a case where the optimization target device, the measurement device, and the optimization support device are independent from each other will be described.

  FIG. 13 is a diagram illustrating an application example of the optimization support apparatus according to the fourth embodiment. FIG. 13 shows an apparatus 71, a measuring apparatus 72, and an optimization support apparatus 73.

  The device 71 is an optimization target device. The operation of the device 71 is controlled by a plurality of parameters. The apparatus 71 is, for example, an optical transmission module that does not include the optimization support apparatus 3 in FIG.

  A signal output from the device 71 is input to the measuring device 72. The measurement device 72 has the same function as the measurement device 2 shown in FIG. 1, measures the input signal, and outputs the measurement data to the optimization support device 73.

  The optimization support device 73 has the same function as the optimization support device 3 shown in FIG. 1 and outputs control data for controlling the device 71 based on the measurement data output from the measurement device 72. . Further, the optimization support apparatus 73 calculates initial value parameters when the optimization process of the apparatus 71 is started.

  The optimization support device 73 is realized by a terminal device such as a personal computer, for example. For example, the optimization support apparatus 73 can realize the same function as the optimization support apparatus 3 by causing a terminal device to execute a program that realizes the function of the optimization support apparatus 3 illustrated in FIG.

  As described above, even when the optimization target device, the measurement device, and the optimization support device are independent from each other, it is possible to calculate the initial value parameter of the parameter that controls the optimization target device.

  As mentioned above, although this invention was demonstrated using embodiment, the technical scope of this invention is not limited to the range as described in the said embodiment. It will be apparent to those skilled in the art that various modifications or improvements can be made to the above embodiment. In addition, it is apparent from the scope of the claims that the embodiments added with such changes or improvements can be included in the technical scope of the present invention. Moreover, it is also possible to combine each embodiment.

  In each of the above embodiments, the functional configuration of each unit is classified according to the main processing contents in order to facilitate understanding of the configuration. The present invention is not limited by the way of classification and names of the constituent elements. The configuration of each unit can be classified into more components according to the processing content. Moreover, it can also classify | categorize so that one component may perform more processes. Further, the processing of each component may be executed by one hardware or may be executed by a plurality of hardware.

  Further, the processing unit of the flowchart described above is divided according to the main processing contents in order to facilitate understanding of the processing of the optimization support device and the initial value parameter calculation unit. The present invention is not limited by the way of dividing the processing unit or the name. The processing of the optimization support device and the initial value parameter calculation unit can be divided into more processing units depending on the processing content. Moreover, it can also divide | segment so that one process unit may contain many processes. Further, the processing order of the above flowchart is not limited to the illustrated example.

  In each of the above-described embodiments, the control lines and information lines indicate what is considered necessary for the description, and not all the control lines and information lines on the product are necessarily shown. In practice, it may be considered that almost all configurations are connected to each other.

  1, 50, 60: Optical transmission module, 2, 63, 72: Measuring device, 3, 51, 73: Optimization support device, 4: Driver circuit, 5: Laser diode, 11: Receiving unit, 12: Holding unit, 13, 41: optimization unit, 14: control data output unit, 15, 42: initial value parameter calculation unit, 21, 22, 31: random number generation unit, 23-25: calculation unit, 26: movement amount calculation unit, 32 : Random parameter generator, 33: parameter extractor, 34, 42a: regression model generator, 35: initial value calculator, 61: photodiode, 62: receiver circuit, 71: device.

Claims (9)

A random parameter generator that randomly generates parameters;
A parameter extraction unit that extracts the parameter when the operation of the device that operates based on the parameter generated by the random parameter generation unit is within a predetermined range with respect to a target value;
Based on the parameters extracted by the parameter extraction unit, a regression model generation unit that generates a regression model of the operation of the device using the parameters as explanatory variables;
Using the regression model generated by the regression model generation unit, an initial value calculation unit for calculating an initial value parameter of the device;
An information processing apparatus comprising: The information processing apparatus according to claim 1,
The information processing apparatus, wherein the initial value calculation unit calculates an initial value parameter by substituting a target value of the operation of the apparatus into an explanatory variable of the regression model. The information processing apparatus according to claim 1,
An optimization unit for calculating a control parameter for the device to operate at a target value;
The information processing apparatus, wherein the optimization unit sets the initial value parameter calculated by the initial value calculation unit in the apparatus as an initial value of the control parameter. The information processing apparatus according to claim 3,
The regression model generation unit receives the control parameter from the optimization unit when the initial value calculation unit cannot calculate the initial value parameter, and sets the value of the regression model for the control parameter to the optimization unit. Send
The information processing apparatus, wherein the optimization unit uses the control parameter when the value of the regression model reaches a target value of the operation of the apparatus as the initial value parameter. The information processing apparatus according to claim 1,
The information processing apparatus, wherein the regression model generation unit generates the regression model by removing the parameter having low relevance to the operation of the apparatus. The information processing apparatus according to claim 1,
The information processing apparatus, wherein the regression model generation unit generates the regression model by multiple regression analysis. The information processing apparatus according to claim 3,
The information processing apparatus, wherein the optimization unit calculates the control parameter so that a signal output from the apparatus becomes a desired signal. The information processing apparatus according to claim 3,
The apparatus is an optical module,
The information processing apparatus, wherein the optimization unit calculates the control parameter so that an optical signal output from the optical module becomes a desired optical signal. A parameter calculation method for an information processing device, comprising:
A random parameter generation step for randomly generating parameters;
A parameter extraction step of extracting the parameter when the operation of the device operating based on the parameter generated by the random parameter generation step is within a predetermined range with respect to a target value;
Based on the parameters extracted by the parameter extraction step, a regression model generation step of generating a regression model of the operation of the device using the parameters as explanatory variables;
Using the regression model generated by the regression model generation step, an initial value calculation step of calculating an initial value parameter of the device;
The parameter calculation method characterized by having.

Images