A C+L+S band multi-device configuration optimization method and device

Abstract

The application provides a C+L+S band multi-device configuration optimization method and device, comprising: using a local optimal approximate global optimal algorithm to calculate an approximate optimal launch power at a fiber input; taking the approximate optimal launch power as a received power of each span, recalculating a launch signal power at a start of the span by inversely solving a power evolution equation of stimulated Raman scattering effect; setting a gain and a gain slope of a fiber amplifier, so that a signal power at a start of each span approximates the launch signal power; constructing a loss function based on a generalized signal-to-noise ratio of each band; taking the set gain and gain slope after the inverse of the stimulated Raman scattering effect as initial values, using a gradient ascent algorithm, taking the loss function as a target, and iteratively optimizing the gain and gain slope of the fiber amplifier and a channel loss value of a wavelength selective switch. The method provided by the application only needs a small amount of iteration, has a high running speed, and can realize multi-device second-level configuration optimization of a wide-band optical transmission system.

Description

Technical Field

[0001] This invention relates to the field of optical fiber communication technology, and in particular to a method and apparatus for optimizing the configuration of multiple devices in the C+L+S band. Background Technology

[0002] Currently, with the continuous growth in traffic demand, the industry has proposed various solutions to match the capacity growth trend by upgrading optical networks. Multi-band transmission exhibits significant economic advantages due to its ability to effectively utilize existing fiber optic infrastructure and is more promising than space-division multiplexing technology. At present, C-band and L-band device technologies have matured and are gradually being deployed and promoted in existing networks. At the same time, the industry is shifting its R&D focus to the S-band, mainly based on two considerations: firstly, the signal attenuation characteristics of the S-band are better than those of the E-band and U-band; secondly, the relatively mature thulium-doped fiber amplifier (TDFA) can be used for amplification in this band. Given the adjacent spectrum between the S-band and L-band and the technological maturity of TDFA, the S-band naturally becomes an ideal choice for extending C+L transmission systems. However, it is worth noting that while the introduction of the S-band expands the transmission bandwidth, it also exacerbates the stimulated Raman scattering (SRS) effect, causing power transfer from high-frequency channels to low-frequency channels. This not only causes performance imbalances between bands but also seriously degrades the transmission quality of the S-band. Against this backdrop, achieving performance balance and quality optimization in C+L+S multi-band systems has become a pressing technical challenge for the industry.

[0003] Generalized signal-to-noise ratio (GSNR), as a core indicator for measuring transmission quality (QoT), is currently a hot research topic. Optimizing GSNR in C+L+S band transmission systems, which are characterized by significant and complex nonlinear effects, is a key research focus. Currently, academic and industrial research on GSNR optimization for multi-band transmission systems mainly focuses on the following directions: First, maximizing link capacity while ensuring transmission stability, which requires precise configuration of power allocation for each band; second, improving the system's maximum GSNR value and optimizing its spectral flatness, which is particularly important for long-distance transmission; third, improving the GSNR performance of the worst-performing channel in the system, which directly determines the lower limit of the overall transmission quality; and finally, comprehensively considering the dynamic balance between amplified spontaneous emission (ASE) noise and nonlinear interference (NLI), which is crucial for achieving optimal transmission performance.

[0004] To address the above optimization objectives, many methods have been proposed in recent years, primarily through controlling transmit power or configuring amplifiers based on various algorithms. These include Local-to-Global Optimization (LOGO) algorithms, Genetic Algorithms (GA), Particle Swarm Optimization (PSO) algorithms, greedy search, and gradient descent algorithms. The LOGO algorithm is a hierarchical optimization strategy that decomposes the optimization problem into local and global levels for collaborative solution. In the local optimization phase, the algorithm configures devices for a single segment or band. In the global optimization phase, it comprehensively considers the interactions between segments throughout the entire transmission link to achieve optimal overall performance. Genetic Algorithms are heuristic search algorithms that progressively optimize solutions based on natural selection and genetic mechanisms. They simulate an evolutionary process, including selection, crossover, and mutation operations, to find the optimal solution in the search space. During optimization, the algorithm evaluates the fitness of each configuration (e.g., average GSNR or flatness), retaining high-quality individuals and eliminating low-quality ones, converging to an approximate optimal solution after multiple generations of evolution. The PSO algorithm is a swarm intelligence-based optimization algorithm that finds the optimal solution by simulating the cooperative search behavior of a swarm of particles in the solution space. Each particle represents a possible solution, and the particles adjust their positions based on their own experience and the experience of the swarm, gradually approaching the global optimum. The greedy search algorithm employs a stepwise optimization strategy, selecting the best local improvement at each step. Starting from the initial configuration, this algorithm compares the performance metrics of neighboring solutions and always chooses the direction that maximizes the improvement of the objective function. This algorithm is computationally efficient and simple to implement, but may get stuck in local optima. It is suitable for quickly obtaining approximate optimal configurations of various components in scenarios with small parameter spaces. The gradient descent algorithm can adjust the configuration of various components to achieve optimal system performance under certain constraints (such as signal quality or noise limitations). The fast convergence characteristic of the gradient direction makes it suitable for handling optimization problems with relatively smooth objective functions.

[0005] In C+L+S band transmission systems, the introduction of the S-band leads to a larger optimization space, more complex calculations, higher time costs, and the need for more configured equipment, such as transceivers, amplifiers, and wavelength selective switches (WSS). Existing technologies often configure only a single type of device, and each type can only control a portion of the transmission performance. For example, an in-line amplifier can only improve the average performance of a single frequency band but lacks the ability to control the performance of a single channel. Therefore, single-objective optimization may produce suboptimal results. Furthermore, as the number of objects to be configured increases, the computational complexity of the aforementioned algorithms also increases, resulting in longer computation times for complex iterative calculations during the optimization process. Summary of the Invention

[0006] In view of this, embodiments of the present invention provide a method and apparatus for optimizing the configuration of multiple devices in the C+L+S band, so as to eliminate or improve one or more defects existing in the prior art.

[0007] On one hand, the present invention provides a method for optimizing the configuration of multiple devices in the C+L+S band, the method comprising the following steps:

[0008] A local optimum approximating global optimum algorithm is used to calculate the approximate optimal transmit power at the fiber optic entry point;

[0009] The approximate optimal transmit power is used as the receive power of each segment. The transmit signal power at the beginning of the segment is recalculated by solving the power evolution equation of stimulated Raman scattering in reverse. The signal power input at the beginning of each segment is made to approximate the transmit signal power by setting the gain and gain slope of the fiber amplifier.

[0010] The generalized signal-to-noise ratios of the C, L, and S bands are calculated respectively to construct the loss function. The gain and gain slope of the fiber amplifier, which are set after the stimulated Raman scattering effect inversion, are used as initial values. The gradient ascent algorithm is used to iteratively optimize the gain and gain slope of the fiber amplifier with the goal of improving the loss function, so as to obtain the optimal transmission power.

[0011] Based on the initial loss value preset at the factory of the wavelength selective switch, the gradient ascent algorithm is used to iteratively fine-tune the channel loss value of the wavelength selective switch with the goal of improving the loss function.

[0012] In some embodiments of the present invention, a local optimal approximation global optimal algorithm is used to calculate the approximate optimal transmit power at the fiber optic entry point, including:

[0013] The local optimal approximation global optimal algorithm ignores the stimulated Raman scattering effect, and each segment in the link can be optimized individually to achieve the optimal optical performance of each segment. The optimal optical performance of the entire multi-segment link is approximately calculated by directly adding or superimposing the optimal optical performance of each segment.

[0014] In some embodiments of the present invention, the calculation process includes:

[0015] According to the local optimum approximation global optimum algorithm, the optical signal-to-noise ratio of any segment is expressed as:

[0016]

[0017] Based on the aforementioned local optimum approximation global optimum algorithm, the optimal optical signal-to-noise ratio nonlinear noise power spectrum can only yield a suboptimal transmit power, expressed as:

[0018]

[0019] Among them, osnr n G represents the optical signal-to-noise ratio of the nth segment; WDM,n The nonlinear noise power spectrum representing the optical signal-to-noise ratio; R s Indicates symbol rate; a n B represents the optical transmission attenuation in the nth segment; N The channel bandwidth is represented by h; Planck's constant is represented by v; the channel center frequency is represented by F. n ρ represents the noise figure of the fiber amplifier in the nth span; NLI,n P represents the nonlinear crosstalk coefficient; in This represents the suboptimal transmit power; This represents the optimal nonlinear noise power spectrum for optical signal-to-noise ratio.

[0020] In some embodiments of the present invention, the optical signal-to-noise ratio formula is as follows:

[0021] To achieve the optimal optical signal-to-noise ratio nonlinear noise power spectrum, the calculation formula is as follows:

[0022]

[0023] In the formula, ρ NLI,n The nonlinear crosstalk coefficient is expressed as follows:

[0024]

[0025] in, The optical signal-to-noise ratio optimal nonlinear noise power spectrum is represented by h; Planck's constant is represented by v; the channel center frequency is represented by F. n a represents the noise figure of the fiber amplifier in the nth span; n The light transmission attenuation in the nth segment is represented by α; γ represents the attenuation; β represents the nonlinear coefficient; and L represents the group velocity dispersion. eff Indicates the effective length of the optical fiber; B wDM This represents the entire transmission bandwidth.

[0026] In some embodiments of the present invention, the generalized signal-to-noise ratios of the C, L, and S bands are calculated respectively to construct a loss function, including:

[0027] The loss function comprises the negative of the average value and standard deviation of the generalized signal-to-noise ratio, and is calculated as follows:

[0028] J = mean(GSNR) - τstd(GSNR);

[0029] Where J represents the loss function; meas(GSNR) represents the average value of the generalized signal-to-noise ratio; std(GSNR) represents the standard deviation of the generalized signal-to-noise ratio; and τ represents the weighting coefficient.

[0030] In some embodiments of the present invention, a gradient ascent algorithm is used to iteratively optimize the gain and gain slope of the fiber amplifier with the goal of improving the loss function, including:

[0031] The gradient ascent algorithm is used to optimize each fiber amplifier segment independently or to jointly optimize multiple fiber amplifier segments.

[0032] In some embodiments of the present invention, the method further includes:

[0033] In the independent optimization scenario, it is assumed that the effects between each segment are not coupled; the gain and gain slope of the fiber amplifier, set after the stimulated Raman scattering effect inversion, are used as initial values; for each segment, the current generalized signal-to-noise ratio is calculated, and the loss function is calculated based on the generalized signal-to-noise ratio; the gradient of the loss function with respect to the gain and gain slope of the fiber amplifier is calculated; the gain and gain slope of each segment are updated using the gradient ascent algorithm; when the loss function converges or reaches the preset number of iterations, the optimization is stopped, and the optimal gain and optimal gain slope of each segment are obtained;

[0034] In the joint optimization scenario, it is assumed that the effects between each segment are coupled; the gain and gain slope of the fiber amplifier set after the stimulated Raman scattering effect inversion are used as initial values; for all segments, the generalized signal-to-noise ratio after joint optimization is calculated, and the loss function is calculated based on the generalized signal-to-noise ratio; the gradient of the loss function with respect to the gain and gain slope of the fiber amplifier is calculated; the gain and gain slope of all segments are updated using the gradient ascent algorithm; when the loss function converges or reaches the preset number of iterations, the optimization is stopped, and the optimal gain and optimal gain slope of all segments are obtained.

[0035] In some embodiments of the present invention, based on the initial loss value preset at the time of manufacture of the wavelength selective switch, the gradient ascent algorithm is used to iteratively fine-tune the channel loss value of the wavelength selective switch with the goal of improving the loss function, including:

[0036] The gradient ascent algorithm is used to fine-tune the loss value of each channel of the wavelength selection switch, or the superchannel optimization method is used to adjust the loss values ​​of multiple adjacent channels simultaneously.

[0037] In some embodiments of the present invention, the method further includes:

[0038] In the channel-by-channel optimization scenario, the gradient of the loss function with respect to the loss value of each channel of the wavelength selective switch is calculated; the gradient ascent algorithm is used to update the loss value of each channel of the wavelength selective switch; when the loss function converges or reaches the preset number of iterations, the optimization is stopped, and the optimal loss value of each channel of the wavelength selective switch is obtained.

[0039] In the superchannel optimization scenario, a preset number of adjacent channels in the wavelength selection switch are merged into a superchannel, and the gradient of the loss function with respect to the loss value of the superchannel is calculated; the gradient ascent algorithm is used to update the loss value of the superchannel; when the loss function converges or reaches a preset number of iterations, the optimization is stopped, and the optimal loss value of the superchannel is obtained.

[0040] On the other hand, the present invention also provides a C+L+S band multi-device configuration optimization apparatus, the apparatus being used to implement the steps of any of the C+L+S band multi-device configuration optimization methods mentioned above.

[0041] This invention provides a method and apparatus for optimizing the configuration of multiple devices in the C+L+S band, comprising: employing a local optimal approximation-global optimal algorithm to calculate the approximate optimal transmit power at the fiber optic entry point; using the approximate optimal transmit power as the received power for each span, and recalculating the transmit signal power at the start of each span by inversely solving the power evolution equation of stimulated Raman scattering; setting the gain and gain slope of the fiber amplifier to make the input signal power at the start of each span approximate the transmit signal power; constructing a loss function based on the generalized signal-to-noise ratio of each band; using the gain and gain slope set after stimulated Raman scattering inversion as initial values, and using a gradient ascent algorithm to iteratively optimize the gain and gain slope of the fiber amplifier and the channel loss value of the wavelength selection switch with the goal of improving the loss function. Furthermore, a superchannel optimization method is employed to simultaneously adjust the channel loss values ​​of multiple adjacent channels of the wavelength selection switch, reducing parameter dimensionality, improving efficiency, and reducing time costs. The method provided by this invention requires only a few iterations, operates quickly, and can achieve multi-device configuration optimization for broadband optical transmission systems. It also ensures a flat and stable output of the generalized signal-to-noise ratio, making the entire system run stably and resources be used efficiently.

[0042] Additional advantages, objects, and features of the invention will be set forth in part in the description which follows, and will also become apparent in part to those skilled in the art upon studying the description, or may be learned by practice of the invention. The objects and other advantages of the invention can be realized and obtained by means of the structures specifically pointed out in the description and drawings.

[0043] Those skilled in the art will understand that the objectives and advantages achievable with the present invention are not limited to those specifically described above, and that the above and other objectives achievable with the present invention will become clearer from the following detailed description. Attached Figure Description

[0044] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, are not intended to limit the scope of the invention. In the drawings:

[0045] Figure 1 This is a schematic diagram illustrating the steps of a C+L+S band multi-device configuration optimization method in one embodiment of the present invention.

[0046] Figure 2 This is a schematic diagram illustrating the principle of a multi-device configuration optimization method for the C+L+S band in one embodiment of the present invention.

[0047] Figure 3 This is a flowchart illustrating a multi-device configuration optimization method for the C+L+S band in one embodiment of the present invention. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the embodiments and accompanying drawings. Here, the illustrative embodiments and descriptions of this invention are used to explain the invention, but are not intended to limit the invention.

[0049] It should also be noted that, in order to avoid obscuring the invention with unnecessary details, only the structures and / or processing steps closely related to the solution according to the invention are shown in the accompanying drawings, while other details that are not closely related to the invention are omitted.

[0050] It should be emphasized that the term "including / comprises" as used herein refers to the presence of a feature, element, step, or component, but does not exclude the presence or addition of one or more other features, elements, steps, or components.

[0051] It should also be noted that, unless otherwise specified, the term "connection" in this article can refer not only to a direct connection, but also to an indirect connection involving an intermediary.

[0052] In the following description, embodiments of the invention will be illustrated with reference to the accompanying drawings. In the drawings, the same reference numerals represent the same or similar parts, or the same or similar steps.

[0053] It should be emphasized here that the step markers mentioned below are not a limitation on the order of the steps, but should be understood as meaning that the steps can be executed in the order mentioned in the embodiments, or in a different order than in the embodiments, or several steps can be executed simultaneously.

[0054] To address the shortcomings of existing optical transmission system device configuration optimization techniques, such as the need for numerous complex iterations, susceptibility to local optima, slow operating speed, inability to scale well with the number of spans, unsuitability for real-time reconfiguration and control of dynamic networks, and inadequate for rapid fault recovery, as well as the increased complexity and suboptimal performance of single device configurations with wider bands, this invention proposes a C+L+S band multi-device configuration optimization method. This method optimizes the final output performance by configuring various devices in the link, including transmitters, amplifiers, and wavelength selective switches. Firstly, based on a local optimum approximating global optimum algorithm, it further considers stimulated Raman scattering (SRS). Finally, using a gradient ascent algorithm, it directly aims to maximize the output generalized optical signal-to-noise ratio (SNR) by adjusting the gain and gain slope settings of the fiber amplifiers in each span and the channel loss configuration of the wavelength selective switches, while ensuring the output SNR is as flat as possible, achieving fast and accurate optimization. Specifically, as shown... Figure 1 As shown, the method includes the following steps S101 to S104:

[0055] Step S101: Use a local optimal approximation global optimal algorithm to calculate the approximate optimal transmission power at the fiber entry point.

[0056] Step S102: Use the approximate optimal transmit power as the receive power for each segment; recalculate the transmit signal power at the beginning of the segment by solving the power evolution equation of stimulated Raman scattering effect in reverse; and make the input signal power at the beginning of each segment approximate the transmit signal power by setting the gain and gain slope of the fiber amplifier.

[0057] Step 103: Calculate the generalized signal-to-noise ratios of the C, L, and S bands respectively to construct the loss function; use the gain and gain slope of the fiber amplifier set after stimulated Raman scattering inversion as initial values, and use the gradient ascent algorithm to iteratively optimize the gain and gain slope of the fiber amplifier with the goal of improving the loss function.

[0058] Step S104: Based on the initial loss value preset when the wavelength selective switch leaves the factory, the channel loss value of the wavelength selective switch is iteratively fine-tuned using the gradient ascent algorithm with the goal of improving the loss function.

[0059] like Figure 2 The diagram shown illustrates the principle of a wideband multi-device configuration optimization algorithm adapted to dynamic scenarios. Figure 3 The diagram shows a flowchart of a wideband multi-device configuration optimization algorithm.

[0060] In step S101, the stimulated Raman scattering effect is ignored, and only the balance between the amplifier's spontaneous emission and nonlinear crosstalk is considered. The local optimum approximate global optimum (LOGO) algorithm is used to calculate the approximate optimal transmit power at the fiber optic amplifier's input point.

[0061] The LOGO (Lossless Optimization of Gain and Slope) algorithm is derived from a nonlinear Gaussian noise model in optical transmission. It ignores stimulated Raman scattering (SRS) effects, allowing each span in the link (e.g., the amplification section of each fiber amplifier) ​​to be optimized individually to achieve the optimal optical performance (e.g., signal-to-noise ratio) for each span. By directly adding or superimposing the optimal optical performance of each span, the optimal optical performance of the entire multi-span link is approximately calculated.

[0062] In some embodiments, the fiber amplifier employs an erbium-doped fiber application amplifier (EDFA) and a thulium-doped fiber amplifier (TDFA).

[0063] In some embodiments, according to the LOGO algorithm, the optical signal-to-noise ratio of the nth segment can be expressed as formula (1):

[0064]

[0065] Among them, osnr n G represents the optical signal-to-noise ratio of the nth segment; WDM,n The nonlinear noise power spectrum representing the optical signal-to-noise ratio; R s Indicates symbol rate; a n B represents the optical transmission attenuation in the nth segment; N The channel bandwidth is represented by h; Planck's constant is represented by v; the channel center frequency is represented by F. n ρ represents the noise figure of the nth span fiber amplifier; NLI,n This represents the nonlinear crosstalk coefficient.

[0066] In formula (1), to achieve the optical signal-to-noise ratio osnr n The optimal nonlinear noise power spectrum is calculated as shown in formula (2):

[0067]

[0068] in, Optical signal-to-noise ratio (OSNR) n The optimal nonlinear noise power spectrum and the explanation of the other parameters are given in Formula (1), and will not be repeated here.

[0069] In formula (2), ρ NLI,n The nonlinear crosstalk coefficient is represented by the formula shown in equation (3):

[0070]

[0071] Where α represents the attenuation when frequency (channel) correlation is not considered; γ represents the nonlinear coefficient; β represents the group velocity dispersion; L eff Indicates the effective length of the optical fiber; B WDM This represents the entire transmission bandwidth.

[0072] Since the derivation of the LOGO algorithm is completed without considering the stimulated Raman scattering effect, the optical signal-to-noise ratio (OSNR) is reduced in C+L band transmission. n Optimal nonlinear noise power spectrum Only a suboptimal transmission power can be obtained, as shown in formula (4):

[0073]

[0074] Among them, P in This indicates the suboptimal transmit power.

[0075] In step S102, the suboptimal transmit power P calculated based on step S101 is... in This serves as the received power for each span. Since stimulated Raman scattering (SRS) was neglected in step S101, in step S102, the transmitted signal power at the beginning of each span is recalculated by inversely solving the SRS power evolution equation. This process yields a more accurate signal power, thereby optimizing the gain and gain slope of the fiber amplifier. These gains and gain slopes are used to adjust the fiber amplifier so that the input signal power at the beginning of each span is closer to the ideal transmitted signal power obtained through SRS inversion calculation.

[0076] Gain refers to the degree to which an optical fiber amplifier amplifies the input signal, usually expressed as the ratio of output power to input power. It measures the change in signal strength after passing through the amplifier. Gain is usually expressed in decibels (dB). The higher the gain, the stronger the signal is amplified; conversely, a negative gain indicates signal attenuation. Gain slope refers to the variation of signal gain over a frequency range, or more specifically, the degree to which signal gain changes with wavelength (or frequency) in optical fiber communication. A slope indicates the change in signal gain across different frequency bands (or wavelengths). In some optical fiber communication systems, because different wavelengths of signals have different gains when passing through the amplifier, the gain slope describes this difference. Typically, systems aim for a gain as flat as possible across different wavelengths to ensure uniform amplification of all signals.

[0077] In step S103, considering the average signal-to-noise ratio of the channel, the present invention aims to achieve a high and relatively flat generalized signal-to-noise ratio (GSNR) over a wide band. Therefore, a loss function is constructed based on the generalized signal-to-noise ratio.

[0078] In some embodiments, the loss function comprises the negative of the mean and standard deviation of the generalized signal-to-noise ratio, and is calculated as shown in Equation (5):

[0079] J=mean(GSNR)-τstd(GSNR); (5)

[0080] Where J represents the loss function; mean(GSNR) represents the average value of the generalized signal-to-noise ratio; std(GSNR) represents the standard deviation of the generalized signal-to-noise ratio; and τ represents the weighting coefficient.

[0081] Using the gain and gain slope of the fiber amplifier set after the stimulated Raman scattering effect inversion in step S102 as initial values, the gradient ascent algorithm is further used to iteratively optimize the gain and gain slope values ​​with the goal of improving the loss function J, so as to obtain a theoretically optimal transmit power.

[0082] In some embodiments, the gradient ascent algorithm can optimize each fiber amplifier segment independently or jointly across multiple fiber amplifier segments. Specifically:

[0083] In the independent optimization scenario, it is assumed that the effects between the segments are not coupled. The gain and gain slope of the fiber amplifier, set after stimulated Raman scattering inversion, are used as initial values. For each segment, the current generalized signal-to-noise ratio (SNR) is calculated, and the loss function is calculated based on the SNR. The gradient of the loss function with respect to the gain and gain slope of the fiber amplifier is calculated, and the gain and gain slope of each segment are updated using the gradient ascent algorithm. When the loss function converges or reaches the preset number of iterations, the optimization stops, and the optimal gain and optimal gain slope of each segment are obtained.

[0084] In this scenario, assuming that the effects between each segment are not coupled, the optimization is performed segment by segment. For example, the optimization space of each segment consists of two amplifiers, each with two optimizable parameters: gain and gain slope. Thus, the optimization space of each segment has only four parameter dimensions, which can achieve faster search optimization. However, since the optimization is performed segment by segment, it may not be possible to guarantee reaching the global optimum.

[0085] In the joint optimization scenario, it is assumed that the effects between the segments are coupled. The gain and gain slope of the fiber amplifier, set after stimulated Raman scattering inversion, are used as initial values. For all segments, the combined generalized signal-to-noise ratio (SNR) is calculated, and the loss function is calculated based on the SNR. The gradient of the loss function with respect to the gain and gain slope of the fiber amplifier is calculated, and the gain and gain slope of all segments are updated using the gradient ascent algorithm. When the loss function converges or reaches the preset number of iterations, the optimization stops, and the optimal gain and optimal gain slope of all segments are obtained.

[0086] The effects between the segments are coupled, so optimization is carried out jointly by all segments. For example, the optimization space for 10 segments has 40 parameter dimensions, which is quite large and the running speed will be relatively slow, but it can guarantee global optimality.

[0087] In step S104, similar to step S103, based on the initial loss value preset when the wavelength selective switch leaves the factory, with the goal of achieving a high and flat generalized signal-to-noise ratio, i.e., improving the loss function, the gradient ascent algorithm is used to iteratively fine-tune the channel loss value of the wavelength selective switch.

[0088] In some embodiments, the wavelength selection switch is manufactured with each channel having the same preset initial loss value. For example, the range of the initial loss value is 15 to 20 dB.

[0089] In some embodiments, a gradient ascent algorithm is used to fine-tune the loss value of each channel of the wavelength selective switch, or a superchannel optimization method is used to adjust the loss values ​​of multiple adjacent channels simultaneously.

[0090] Specifically, in the channel-by-channel optimization scenario, the gradient of the loss function relative to the loss value of each channel of the wavelength selective switch is calculated; the gradient ascent algorithm is used to update the loss value of each channel of the wavelength selective switch; when the loss function converges or reaches the preset number of iterations, the optimization is stopped, and the optimal loss value of each channel of the wavelength selective switch is obtained.

[0091] In the superchannel optimization scenario, a predetermined number of adjacent channels in the wavelength selection switch are merged into a superchannel. For example, four adjacent channels are merged into a superchannel, and the channels within the superchannel share the same loss value. The gradient of the loss function with respect to the loss value of the superchannel is calculated; the gradient ascent algorithm is used to update the loss value of the superchannel; when the loss function converges or reaches a predetermined number of iterations, the optimization stops, and the optimal loss value of the superchannel is obtained. This achieves the technical effect of optimizing transmission quality while reducing algorithm complexity and optimization time.

[0092] From a technical implementation perspective, this step-by-step multi-device configuration optimization algorithm has low complexity. It configures different devices separately in each step without requiring numerous complex iterations. Compared to methods that configure a single type of device, it can achieve better transmission performance in a shorter time. Furthermore, the use of a super-channel optimization method to simultaneously fine-tune the loss values ​​of adjacent channels significantly shortens the algorithm's runtime.

[0093] In summary, the C+L+S band multi-device configuration optimization method provided by this invention first rapidly determines the approximate optimal transmit power in steps S101-S102 using numerical methods, approaching the optimal amplifier configuration. In step S103, a small number of gradient ascent iterations are used to finally approximate the optimal value. Traditional search algorithms, however, need to find the optimal value in a vast search space. For an eight-span link transmitting in the C+L+S band, each span contains three amplifiers, and each amplifier has two optimization parameters: gain and gain slope. This results in a total of 48 optimizable parameters, significantly increasing computational complexity. The step-by-step multi-device configuration optimization algorithm proposed in this invention first considers linear noise and Kerr nonlinear effects in a decoupled manner, then considers SRS, and finally compensates for the coupling relationships between various effects through a small number of iterations, greatly improving the operating speed.

[0094] The generalized signal-to-noise ratio obtained after optimization through steps S101 to S103 of the present invention was compared with that of the CMA-GA algorithm. The results were basically similar. Moreover, the optimization time of the present invention was only 47.43s, while the optimization time of the CMA-GA algorithm was 95.55s, which is about twice the optimization time required by the present invention.

[0095] In addition, this invention also considers the configuration of multiple devices. In the C+L+S transmission band, traditional methods only configure a single device, which often makes it difficult to obtain the optimal solution for the generalized signal-to-noise ratio in terms of both high mean and high flatness. The proposed step-by-step multi-device configuration optimization algorithm further configures a wavelength selection switch in step S104, which can further improve the flatness of the generalized signal-to-noise ratio and achieve higher transmission quality compared with traditional methods.

[0096] The present invention also provides a C+L+S band multi-device configuration optimization device, which is used to implement the steps of the C+L+S band multi-device configuration optimization method.

[0097] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the C+L+S band multi-device configuration optimization method.

[0098] Corresponding to the above method, the present invention also provides an apparatus comprising a computer device, the computer device including a processor and a memory, the memory storing computer instructions, the processor executing the computer instructions stored in the memory, and when the computer instructions are executed by the processor, the apparatus performs the steps of the method as described above.

[0099] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the aforementioned edge computing server deployment method. The computer-readable storage medium can be a tangible storage medium, such as random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, floppy disks, hard disks, removable storage disks, CD-ROMs, or any other form of storage medium known in the art.

[0100] Those skilled in the art will understand that the exemplary components, systems, and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software, or a combination of both. Whether implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this invention. When implemented in hardware, it can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this invention are programs or code segments used to perform the desired tasks. The programs or code segments can be stored in a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried in a carrier wave.

[0101] It should be clarified that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of the present invention.

[0102] In this invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with or in place of features of other embodiments.

[0103] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, various modifications and variations of the embodiments of the present invention are possible. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for optimizing the configuration of multiple devices in the C+L+S bands, characterized in that, The method includes the following steps: A local optimal approximation global optimal algorithm is used to calculate the approximate optimal transmit power at the fiber optic entry point. The local optimal approximation global optimal algorithm ignores stimulated Raman scattering, and each span in the link can be optimized individually to achieve the optimal optical performance of each span. The optimal optical performance of the entire multi-span link is approximately calculated by directly adding or superimposing the optimal optical performance of each span. According to the local optimum approximation global optimum algorithm, the optical signal-to-noise ratio of any segment is expressed as: ； Based on the aforementioned local optimum approximation global optimum algorithm, the optimal optical signal-to-noise ratio nonlinear noise power spectrum can only yield a suboptimal transmit power, expressed as: ； in, Indicates the first The optical signal-to-noise ratio of each segment; The nonlinear noise power spectrum represents the optical signal-to-noise ratio; Indicates the symbol rate; Indicates the first Optical transmission attenuation across a single segment; Indicates channel bandwidth; Denotes Planck's constant; Indicates the center frequency of the channel; Indicates the first Noise figure of a single-segment fiber amplifier; Represents the nonlinear crosstalk coefficient; This represents the suboptimal transmit power; This represents the optimal nonlinear noise power spectrum for optical signal-to-noise ratio; In the formula for optical signal-to-noise ratio (SNR), the calculation formula for the optimal nonlinear noise power spectrum is as follows: ； in, The nonlinear crosstalk coefficient is expressed as follows: ； in, Indicates attenuation; Indicates the nonlinear coefficient; Indicates group velocity dispersion; Indicates the effective length of the optical fiber; Indicates the total transmission bandwidth; The approximate optimal transmit power is used as the receive power for each segment; the transmit signal power at the beginning of the segment is recalculated by solving the power evolution equation of stimulated Raman scattering in reverse; and the signal power input at the beginning of each segment is made to approximate the transmit signal power by setting the gain and gain slope of the fiber amplifier. The generalized signal-to-noise ratios (SNRs) for the C, L, and S bands are calculated separately to construct a loss function. This loss function includes the negative of the average and standard deviation of the generalized SNR, and is calculated as follows: ； in, Represents the loss function; This represents the average value of the generalized signal-to-noise ratio; This represents the standard deviation of the generalized signal-to-noise ratio; Indicates the weighting coefficient; Using the gain and gain slope of the fiber amplifier set after stimulated Raman scattering inversion as initial values, the gradient ascent algorithm is used to iteratively optimize the gain and gain slope of the fiber amplifier with the goal of improving the loss function. Based on the initial loss value preset at the factory of the wavelength selective switch, the gradient ascent algorithm is used to iteratively fine-tune the channel loss value of the wavelength selective switch with the goal of improving the loss function.

2. The C+L+S band multi-device configuration optimization method according to claim 1, characterized in that, Using the gradient ascent algorithm, with the goal of improving the loss function, the gain and gain slope of the fiber amplifier are iteratively optimized, including: The gradient ascent algorithm is used to optimize each fiber amplifier segment independently or to jointly optimize multiple fiber amplifier segments.

3. The C+L+S band multi-device configuration optimization method according to claim 2, characterized in that, The method further includes: In the independent optimization scenario, it is assumed that the effects between each segment are not coupled; the gain and gain slope of the fiber amplifier set after the stimulated Raman scattering effect inversion are used as initial values; the gradient of the loss function with respect to the gain and gain slope of the fiber amplifier is calculated; the gain and gain slope of each segment are updated using the gradient ascent algorithm; when the loss function converges or reaches the preset number of iterations, the optimization is stopped, and the optimal gain and optimal gain slope of each segment are obtained. In the joint optimization scenario, it is assumed that the effects between each segment are coupled; the gain and gain slope of the fiber amplifier set after the stimulated Raman scattering effect inversion are used as initial values; the gradient of the loss function with respect to the gain and gain slope of the fiber amplifier is calculated; the gain and gain slope of all segments are updated using the gradient ascent algorithm; when the loss function converges or reaches the preset number of iterations, the optimization is stopped, and the optimal gain and optimal gain slope of all segments are obtained.

4. The C+L+S band multi-device configuration optimization method according to claim 1, characterized in that, Based on the initial loss value preset at the factory of the wavelength selective switch, the gradient ascent algorithm is used to iteratively fine-tune the channel loss value of the wavelength selective switch with the goal of improving the loss function, including: The gradient ascent algorithm is used to fine-tune the loss value of each channel of the wavelength selection switch, or the superchannel optimization method is used to adjust the loss values ​​of multiple adjacent channels simultaneously.

5. The C+L+S band multi-device configuration optimization method according to claim 4, characterized in that, The method further includes: In the channel-by-channel optimization scenario, the gradient of the loss function with respect to the loss value of each channel of the wavelength selective switch is calculated; the gradient ascent algorithm is used to update the loss value of each channel of the wavelength selective switch; when the loss function converges or reaches the preset number of iterations, the optimization is stopped, and the optimal loss value of each channel of the wavelength selective switch is obtained. In the superchannel optimization scenario, a preset number of adjacent channels in the wavelength selection switch are merged into a superchannel, and the gradient of the loss function with respect to the loss value of the superchannel is calculated; the gradient ascent algorithm is used to update the loss value of the superchannel; when the loss function converges or reaches a preset number of iterations, the optimization is stopped, and the optimal loss value of the superchannel is obtained.

6. A C+L+S band multi-device configuration optimization device, characterized in that, The apparatus is used to implement the steps of the method as described in any one of claims 1 to 5.

Images