Time-frequency resource mapping conflict avoidance method, apparatus, device, and medium

Abstract

The application relates to the technical field of optical communication and optical signal processing, and provides a time-frequency resource mapping conflict avoidance method, a mapping device, equipment and a medium, which comprises the following steps: acquiring physical parameters and system configuration parameters of multiple optical pulse signals, generating an initial mapping scheme, the mapping scheme comprising time slot positions and wavelength channels allocated to each signal; constructing an optimization objective function under an n-th round mapping scheme, the optimization objective function comprising an interference metric value and a conflict metric value; performing local disturbance on the n-th round mapping scheme to generate a candidate mapping scheme; comparing the optimization objective function values of the candidate mapping scheme and the n-th round mapping scheme, and determining an (n+1)-th round mapping scheme according to a comparison result; and judging whether a convergence condition is met, outputting a current mapping scheme if the convergence condition is met, or continuing iteration if the convergence condition is not met. The application realizes the cooperation of sufficient exploration in the early stage and forced convergence in the later stage through a dynamically increasing weight coefficient.

Description

Technical Field

[0001] This invention relates to the field of optical communication and optical signal processing technology, specifically to a time-frequency resource mapping conflict avoidance method, mapping device, equipment, and medium. Background Technology

[0002] In multiplexed transmission scenarios, optical pulse signals need to be allocated to different time slots and wavelength channels to form a two-dimensional time slot-wavelength resource mapping scheme. The quality of the mapping scheme directly determines the degree of mutual interference between the signals at the receiver, thus affecting the transmission performance and capacity of the entire optical communication system. As optical communication systems develop towards ultra-large-scale and highly dynamic services, how to achieve low-interference and low-collision signal mapping in the time slot-wavelength two-dimensional resource pool has become a continuously attracting technical focus in this field.

[0003] Existing technologies have developed various mapping scheduling schemes. In early optical communication systems, the number of signal paths was relatively small and the standard was relatively simple. The mapping problem was modeled as a discrete combinatorial optimization problem: using the number of wavelengths as discrete resources and time slot intervals as discrete windows, resource occupancy for each signal was determined through static allocation or first-fit algorithms. The maturity of wavelength division multiplexing (WDM) systems made wavelengths allocatable resources, and the wavelength routing problem in optical switching networks was further modeled as a graph coloring problem, thus forming a basic processing framework that understands conflicts as competition for discrete resources. With the increase in the number of signal paths and the enhancement of service dynamism, the limitations of static allocation and first-fit algorithms in terms of resource utilization have gradually become apparent. Those skilled in the art have introduced general heuristic optimization methods such as genetic algorithms and simulated annealing into mapping scheduling, improving the quality of mapping schemes through random search. The above schemes have achieved resource allocation and conflict avoidance for multiple optical pulse signals to a certain extent.

[0004] However, the above-mentioned schemes face a dilemma in practical situations where the number of signal paths increases and pulse physical effects become significant: On the one hand, if the discrete rules are strictly used to determine conflicts (two signals occupying the same wavelength and overlapping time slots are considered a conflict), the optimization process can only obtain binary information about the presence or absence of a conflict, lacking a continuous measurement of the severity of the conflict, and the optimization search lacks effective gradient guidance; on the other hand, if a conflict penalty is introduced in the optimization process, existing practices usually use fixed weights, and the conflict cost remains unchanged regardless of the iteration stage, making it impossible to achieve a balance between fully exploring the solution space in the early stage and forcing convergence in the later stage. The fundamental reason for the above dilemma is that existing schemes are all based on the premise that conflict is a discrete violation event, placing conflict avoidance at the end of the mapping optimization process for post-event inspection. The conflict itself does not participate in the continuous guidance of the objective function, while the interference degree between actual optical pulse signals is continuously affected by physical parameters such as pulse width, spectral width, peak power, and polarization state. There is a fundamental mismatch between discrete judgment and continuous physical reality.

[0005] Therefore, in order to simultaneously achieve effective conflict avoidance and signal quality maintenance in the time-frequency resource mapping of optical pulse signals, we must break away from the traditional framework of discrete resource competition and seek a new solution that can continuously quantify the intensity of physical interference and dynamically adjust the conflict cost as the optimization process progresses, thereby fundamentally breaking the deadlock of conflict avoidance and performance optimization being mutually exclusive. Summary of the Invention

[0006] This invention provides a time-frequency resource mapping conflict avoidance method, mapping device, equipment, and medium, which can solve the problems of conflict avoidance and performance optimization being disconnected in related technologies, and the inability to dynamically adjust conflict costs as the optimization process progresses. The technical solution is as follows: On the one hand, a time-frequency resource mapping conflict avoidance method is provided, the method comprising: Step S1: Obtain the physical parameters and system configuration parameters of the multi-channel optical pulse signals, and generate an initial mapping scheme based on the physical parameters. The mapping scheme includes the time slot position and wavelength channel allocated to each signal, and sets the initial iteration rounds. ; Step S2: Construct the first Wheel mapping scheme The optimization objective function is as follows: ,in: For the first Wheel mapping scheme The interference metric value is used to characterize the degree of mutual interference between signals in the time and frequency domain under the current mapping scheme; For the first Wheel mapping scheme Conflict metrics This is used to characterize the number of collisions between signals under the current mapping scheme; For the first Wheel mapping scheme The corresponding dynamic weighting coefficient, the dynamic weighting coefficient With each iteration round The increase is monotonically increasing; Step S3: For the first Wheel mapping scheme Perform a local perturbation to generate the first... Round candidate mapping scheme ; Step S4: Calculate the candidate mapping scheme for the (n+1)th round. Values ​​under the stated objective function: The candidate mapping scheme The optimization objective function value and the first Wheel mapping scheme Optimization objective function value Compare the results to obtain the mapping scheme for the (n+1)th round. ; Step S5: Map the (n+1)th round's mapping scheme Updated to the current nth round mapping scheme Let n = n + 1, and determine the mapping scheme for the current nth round. Does it meet the preset convergence condition? If it does, output the current nth round mapping scheme. If the convergence condition is not met, return to step S2 and use the current mapping scheme. Continue the next iteration from this starting point.

[0007] On the other hand, a time-frequency resource mapping apparatus is provided, the apparatus comprising: The parameter acquisition unit is used to acquire the physical parameters and system configuration parameters of the multi-channel optical pulse signals; The mapping initialization unit is used to generate an initial mapping scheme based on the physical parameters. The mapping scheme includes the time slot position and wavelength channel allocated to each signal, and sets the initial iteration round. ; Objective function building unit, used to construct the nth round mapping scheme. The optimization objective function is as follows: in, For the nth round of mapping scheme The interference metric value is used to characterize the degree of mutual interference between signals in the time and frequency domain under the current mapping scheme; For the nth round of mapping scheme The collision metric value is used to characterize the number of collisions between signals under the current mapping scheme; For the nth round of mapping scheme The corresponding dynamic weighting coefficient, the dynamic weighting coefficient It monotonically increases with the increase of the iteration number n; Local perturbation units are used to modify the mapping scheme in the nth round. Perform a local perturbation to generate candidate mapping schemes for the (n+1)th round. ; The comparison and determination unit is used to calculate the candidate mapping scheme in the (n+1)th round. Value under the optimization objective function and the candidate mapping scheme The optimized objective function value and the mapping scheme in the nth round Optimization objective function value Compare the results to obtain the mapping scheme for the (n+1)th round. ; The iterative control unit is used to process the mapping scheme of the (n+1)th round. Updated to the current nth round mapping scheme Let n = n + 1, and determine the mapping scheme for the current nth round. Does it meet the preset convergence condition? If it does, output the current nth round mapping scheme. As the final mapping scheme; if the convergence condition is not met, the objective function construction unit, the local perturbation unit, and the comparison and determination unit are triggered to use the current mapping scheme. Continue the next iteration from this starting point.

[0008] On the other hand, a computer device is provided, the computer device including a memory and a processor, the memory for storing computer programs, and the processor for executing the computer programs stored in the memory to implement the steps of the time-frequency resource mapping conflict avoidance method described above.

[0009] On the other hand, a computer-readable storage medium is provided, wherein a computer program is stored therein, and when the computer program is executed by a processor, it implements the steps of the time-frequency resource mapping conflict avoidance method described above.

[0010] The technical solution provided by this invention can bring at least the following beneficial effects: First, by measuring the interference value Conflict metrics They are all incorporated into the same optimization objective function This allows the mapping scheme adjustment in each iteration to simultaneously take into account both signal quality preservation and conflict avoidance, breaking the fragmented processing flow of existing technologies that first map, then check for conflicts, and then retry if a conflict occurs, thus achieving a unity of conflict avoidance and performance optimization.

[0011] Second, by using dynamic weighting coefficients for conflict metrics. It monotonically increases with the iteration number n, allowing the relative cost of the conflict metric in the objective function to adaptively increase with the iteration process: in the early stage of iteration With smaller values, the search process can fully explore the solution space containing some conflicts; in the later stages of iteration... As the value increases, the search process is gradually guided to a low-conflict or even zero-conflict mapping scheme, thus automatically completing the transition from exploration to convergence without human intervention.

[0012] Third, by combining the convergence determination mechanism in step S5 with the characteristic that the dynamic weight coefficients in step S2 monotonically increase with each iteration, the iteration process has a definite termination condition—it terminates and outputs when the mapping scheme satisfies the conflict metric value of zero or the interference metric value meets the standard. As the number of iterations increases, a stronger repulsive force is formed against any conflicting schemes, ensuring that the iterative process can approach and satisfy the convergence condition within a finite number of iterations, thus avoiding the problems of relying on probabilistic termination or infinite oscillation. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a flowchart of a time-frequency resource mapping conflict avoidance method according to an embodiment of this application; Figure 2 This is a schematic diagram of the time slot-wavelength two-dimensional resource pool mapping according to an embodiment of this application; Figure 3(a) is a schematic diagram of the time-domain overlap integration of two Gaussian pulses in an embodiment of this application, showing the overlap of the time-domain envelopes of the two signals when the time slot interval is small; Figure 3(b) is a schematic diagram of the time-domain overlap integration of two Gaussian pulses in an embodiment of this application, showing the overlap of the time-domain envelopes of the two signals when the time slot interval is large; Figure 4(a) is a schematic diagram of the overlap integral of the two-signal spectra of an embodiment of this application, showing the overlap of the spectral envelopes of the two signals when the center wavelength interval is small; Figure 4(b) is a schematic diagram of the overlap integral of the two-signal spectra of an embodiment of this application, showing the overlap of the spectral envelopes of the two signals when the center wavelength interval is large; Figure 5 This is a curve showing the increase of the penalty coefficient with the number of iterations in an embodiment of this application; Figure 6 This is a structural block diagram of a time-frequency resource mapping device according to an embodiment of this application; Figure 7 This is a hardware architecture diagram of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0016] As mentioned above, existing time-frequency mapping schemes for optical pulse signals suffer from the following technical problems: collision detection uses binary judgment, which cannot reflect the continuous differences in interference intensity; the collision penalty weight is fixed, which cannot take into account both early exploration and later convergence; and the pulse physical parameters are not included in the mapping decision, making it difficult to guarantee signal quality. To address this, this application provides a collision avoidance method in time-frequency resource mapping of optical pulse signals. The core idea is to use the product of the time-domain overlap integral, the spectral overlap integral, and the peak power as the time-frequency cross-energy, and simultaneously construct interference and collision metrics based on this, unifying them under the same optimization objective function; the weight coefficient of the collision metric is monotonically increased with each iteration, allowing the optimization process to fully explore the solution space in the early stages and force convergence to a zero-collision solution in the later stages; simultaneously, an incremental update strategy is used to reduce the computational complexity of a single round from O(N²) to O(N). These schemes collectively achieve synergistic optimization of collision avoidance and signal quality preservation. The technical solution of this application will be described in detail below with reference to the accompanying drawings and embodiments.

[0017] Example 1 The following describes the specific implementation of the above concept.

[0018] like Figure 1 As shown, this embodiment provides a time-frequency resource mapping conflict avoidance method. This method is used to determine the allocated time slot position and wavelength channel for each signal through iterative optimization during the mapping of multiple optical pulse signals to a time slot-wavelength two-dimensional resource pool, thereby achieving a balance between conflict avoidance and signal quality preservation in the final mapping scheme.

[0019] Step S1: Obtain the physical parameters and system configuration parameters of the multi-channel optical pulse signals, and generate an initial mapping scheme based on the physical parameters. The mapping scheme includes the time slot location and wavelength channel allocated to each signal, and sets the initial iteration rounds. .

[0020] Specifically, physical parameters refer to parameters related to the physical characteristics of each optical pulse signal, such as pulse width, spectral width, peak power, polarization state, and priority weight. System configuration parameters refer to parameters related to the mapping resource pool and receiver operating conditions, such as the available wavelength set, frame period, collision detection threshold, and receiver noise parameters.

[0021] like Figure 2 As shown, the time-slot-wavelength two-dimensional resource pool is spanned by both the time-slot dimension and the wavelength dimension, where the horizontal axis represents the time-slot position and the vertical axis represents the wavelength channel. The figure illustrates this using 5 wavelength channels (λ1 to λ5) and 6 time-slot positions (t1 to t6) as an example. Each signal is mapped to a specific coordinate point in this two-dimensional plane, where the horizontal coordinate of the coordinate point is the time-slot position allocated to that signal, and the vertical coordinate is the wavelength channel allocated to that signal.

[0022] exist Figure 2 In the specific mapping states shown, optical pulse signals numbered 1 to 7 are assigned to the following resource locations: signal number Allocate time slots wavelength channel allocation Coordinate representation Signal 1 <![CDATA[t2]]> <![CDATA[λ3]]> <![CDATA[(t2,λ3)]]> Signal 2 <![CDATA[t4]]> <![CDATA[λ2]]> <![CDATA[(t4,λ2)]]> Signal 3 <![CDATA[t1]]> <![CDATA[λ5]]> <![CDATA[(t1,λ5)]]> Signal 4 <![CDATA[t5]]> <![CDATA[λ4]]> <![CDATA[(t5,λ4) <!-- 4 -->]]> Signal 5 <![CDATA[t3]]> <![CDATA[λ1]]> <![CDATA[(t3,λ1)]]> Signal 6 <![CDATA[t6]]> <![CDATA[λ3]]> <![CDATA[(t6,λ3)]]> Signal 7 <![CDATA[t4]]> <![CDATA[λ5]]> <![CDATA[(t4,λ5)]]> The seven coordinate points mentioned above are the constituent elements of the mapping scheme M for each signal in the two-dimensional resource pool. In this example, each signal independently occupies a time slot-wavelength combination. However, whether the mapping scheme conflicts (i.e., whether there are signal pairs with the same wavelength and overlapping time slots) needs to be determined by calculating the time-frequency crossover energy based on the physical parameters such as the time-domain pulse width, spectral width, and peak power of each signal and comparing it with a preset threshold. For example, if signals 1 and 6 are both allocated to channel λ3, and their time slot positions t2 and t6 do not overlap in the time domain, then there is no conflict; if the time slots overlap and the time-frequency crossover energy exceeds the preset threshold, then it is determined to be a conflict. The initial mapping scheme M0 can be generated randomly or based on a certain preset rule; this embodiment does not limit this. M0 provides a search starting point for subsequent iterative optimization, and its quality does not directly determine the quality of the final scheme, because the subsequent iteration process will continuously improve the mapping scheme. The initial iteration round n=0 is set, and the iteration round number increases by one after each subsequent iteration.

[0023] Step S2, construct the mapping scheme for the nth round. The optimization objective function is as follows: ,in For the nth round of mapping scheme The interference metric value is used to characterize the degree of mutual interference between signals in the time and frequency domain under the current mapping scheme; For the nth round of mapping scheme The collision metric value is used to characterize the number of collisions between signals under the current mapping scheme; For the nth round of mapping scheme The corresponding dynamic weighting coefficients, dynamic weighting coefficients It monotonically increases with the increase of the iteration number n.

[0024] Specifically, interference metric It is a continuous quantity, and its magnitude reflects the degree of mutual interference between the signals in the time-frequency two-dimensional domain under the current mapping scheme. The stronger the interference, The larger the value, the greater the conflict metric. This reflects the number of conflicting signal pairs under the current mapping scheme; the more conflicts, the higher the probability of collisions. The larger the value of , the better. Although both interference and collision metrics originate from the time-frequency overlap between various signals, they describe information at different levels: interference metrics describe the continuity of signal quality loss, while collision metrics describe the number of collision events. Both are unified in the same objective function for collaborative optimization, enabling the mapping scheme search process to simultaneously consider signal quality preservation and collision avoidance.

[0025] Dynamic weighting coefficients The value monotonically increases with the iteration number n, a design that allows the relative cost of the conflict metric in the objective function to gradually increase with the iteration progress. In the early stages of iteration, When the value is small, the contribution of the conflict metric to the objective function is relatively low. The optimization process can fully explore the solution space containing some conflict to find mapping structures with lower interference metrics. In the later stages of iteration, As the value increases, the contribution of the conflict metric to the objective function increases significantly. Any scheme containing conflict will be at a disadvantage in terms of the objective function value, and the optimization process is forced to guide the mapping scheme with low or even zero conflict.

[0026] Step S3, for the nth round of mapping scheme Perform a local perturbation to generate candidate mapping schemes for the (n+1)th round. .

[0027] Specifically, local perturbation refers to making minor modifications to the current mapping scheme, changing only the time slot position and / or wavelength channel of one or a few signals, without reconstructing the entire mapping scheme. There are various ways to perform local perturbation, such as adjusting the time slot position of a signal, reallocating the wavelength channel of a signal, or swapping the time slot positions of two signals. This embodiment does not limit the specific actions of local perturbation, as long as it can generate a candidate mapping scheme that has local relevance to the current mapping scheme. Since the candidate mapping scheme differs only locally from the current mapping scheme, this locality can be used to reduce computational overhead when calculating the objective function value of the candidate mapping scheme.

[0028] Step S4: Calculate candidate mapping schemes The value under the optimization objective function: Candidate mapping schemes The optimized objective function value and the mapping scheme in the nth round Optimization objective function value Compare the results to obtain the mapping scheme for the (n+1)th round. .

[0029] Specifically, the objective function value of the candidate mapping scheme is calculated in the same way as that of the current mapping scheme. Both use the optimized objective function constructed in step S2, and compare the two objective function values ​​to obtain the mapping scheme for the (n+1)th round. The specific comparison method will not be elaborated in this embodiment, but will be explained in detail in subsequent embodiments.

[0030] Step S5, map the (n+1)th round's mapping scheme Updated to the current nth round mapping scheme Let n = n + 1, and determine the mapping scheme for the current nth round. Does it meet the preset convergence condition? If it does, output the current nth round mapping scheme. If the convergence condition is not met, return to step S2 and use the current mapping scheme. Continue the next iteration from this starting point.

[0031] Specifically, after completing one iteration and determining the mapping scheme for the (n+1)th round, the iteration round number is incremented by one, and then it is determined whether the current mapping scheme meets the preset convergence condition. The convergence condition is used to determine whether the iteration process should be terminated; its specific details are not elaborated in this embodiment but will be provided in detail in subsequent embodiments. When the convergence condition is met, the current mapping scheme is output as the final mapping scheme, which is the time slot-wavelength resource allocation result determined after iterative optimization. When the convergence condition is not met, the current mapping scheme is used as a new starting point, and the process returns to step S2 to continue constructing the next round of optimization objective function and executing subsequent steps. This process is repeated until the convergence condition is met.

[0032] Through iterative execution of steps S1 to S5, this embodiment incorporates interference and conflict metrics into the same optimization objective function and dynamically increases the weight coefficient of the conflict metric with each iteration. This allows the mapping optimization process to fully explore the solution space in the early stage with a low conflict cost and to force convergence to a low-conflict scheme in the later stage with a higher conflict cost, thereby achieving synergistic optimization of conflict avoidance and signal quality maintenance.

[0033] In one specific embodiment, the above-described candidate mapping scheme The optimized objective function value and the mapping scheme in the nth round Optimization objective function value Compare the results to obtain the mapping scheme for the (n+1)th round. To elaborate further.

[0034] Specifically, if , indicating the candidate mapping scheme It is superior to the current [number] in terms of the overall cost of interference and conflict. Wheel mapping scheme The candidate mapping scheme As the mapping scheme for the (n+1)th round , that is to say ; like , indicating the candidate mapping scheme In terms of the combined costs of interference and conflict, it is not superior to the current [number] [unit / system / etc.]. Wheel mapping scheme However, to avoid the search process getting stuck in local optima, a probability is set. Accept the candidate mapping scheme As the mapping scheme for the (n+1)th round , that is to say If not accepted, then the current version will continue. Wheel mapping scheme As the mapping scheme for the (n+1)th round , that is to say ; By comparing the values ​​of the two objective functions, if the objective function value of the candidate mapping scheme is smaller, it indicates that the candidate mapping scheme outperforms the current mapping scheme in terms of the combined performance of interference and conflict metrics. In this case, the candidate mapping scheme is directly accepted as the starting point for the next iteration. If the objective function value of the candidate mapping scheme is not less than that of the current mapping scheme, it indicates that the candidate mapping scheme is not better than the current mapping scheme in terms of overall performance. In this case, it is not directly rejected, but accepted with a set probability P. This mechanism of accepting a poorer candidate scheme with a certain probability enables the iteration process to escape local optima and avoids the search process from prematurely stagnating on a local optimum mapping scheme. The specific calculation method of the set probability P is not elaborated in this embodiment, but will be explained in detail in subsequent embodiments.

[0035] In one specific embodiment, a conflict metric is further provided. The method for determining the nth round of mapping schemes. For any two signals i and j, calculate their time-domain overlap integral and spectral overlap integral. Multiply the time-domain overlap integral, spectral overlap integral, and peak power of the two signals to obtain the time-frequency crossover energy. Compare the time-frequency crossover energy with a preset threshold. If it exceeds the preset threshold, it is determined that the two signals have collided. Perform the above determination on all signal pairs and accumulate the signal pairs determined to collide to obtain the nth round mapping scheme. Conflict metrics .

[0036] As shown in Figure 3, the time-domain overlap integral is used to measure the degree of overlap of the envelopes of two signals in the time domain. Figure 3(a) shows that the overlap of the time-domain envelopes of the two signals is greater when the time slot interval is small, and Figure 3(b) shows that the overlap of the time-domain envelopes of the two signals is smaller when the time slot interval is large. Let the time-domain envelope shape function of the i-th signal be Pi(·), the time-domain envelope shape function of the j-th signal be Pj(·), and the time slot position of the i-th signal under the current mapping scheme be... If the time slot position of the j-th signal under the current mapping scheme is τj, then the formula for calculating the time-domain overlap integral is:

[0037] in, Indicates the i-th signal with The actual time-domain intensity distribution at the time-domain center. Let represent the actual time-domain intensity distribution of the j-th signal centered at τj. Integration is performed over the entire time axis, and the result is the cumulative value of the product of the time-domain envelopes of the two signals. This integral value reaches its maximum when the time slots of the two signals completely overlap; as the time slot interval between the two signals increases, this integral value monotonically decreases. The specific form of the time-domain envelope shape function is not limited; it can be a Gaussian function, a super-Gaussian function, a hyperbolic secant function, or any other function that can describe the actual pulse time-domain intensity distribution, as long as it satisfies the normalization condition.

[0038] As shown in Figure 4, the spectral overlap integral is used to measure the degree of spectral overlap between two signals in the frequency domain. Figure 4(a) shows a larger overlap of the spectral envelopes of the two signals when the center wavelength interval is small, while Figure 4(b) shows a smaller overlap when the center wavelength interval is large. Let the shape function of the spectral envelope of the i-th signal be... The shape function of the spectral envelope of the j-th signal is: The wavelength channel allocated to the i-th signal under the current mapping scheme is: The wavelength channel allocated to the j-th signal under the current mapping scheme is: The formula for calculating the spectral overlap integral is:

[0039] in, Indicates the i-th signal with The actual spectral intensity distribution at the spectral center. Indicates that the j-th signal is... The actual spectral intensity distribution at the spectral center. Integration is performed over the entire wavelength axis, and the result is the cumulative value of the product of the spectral envelopes of the two signals. This integral value reaches its maximum when the wavelength channels of the two signals are exactly the same; it monotonically decreases as the wavelength interval between the two signals increases. The specific form of the spectral envelope shape function is also not limited; it can be a Gaussian function, a Lorentz function, or other function forms that can describe the spectral intensity distribution of the actual light source.

[0040] After obtaining the time-domain overlap integral and the spectral overlap integral, they are multiplied by the peak powers of the two signals to obtain the time-frequency crossover energy. Let the peak power of the i-th signal be... The peak power of the j-th signal is Then time-frequency crossover energy The calculation formula is:

[0041] The physical meaning of time-frequency overlap energy can be understood as follows: when a receiver performs matched filtering demodulation on signal i, the portion of the pulse of signal j that falls within the receiving window of signal i in both the time and frequency domains. The time-domain overlap integral characterizes the proportion of signal j's energy falling within the receiving window of signal i in the time domain, while the spectral overlap integral characterizes the proportion of signal j's energy falling within the receiving filter of signal i in the frequency domain. Both, along with the peak power of signal j, represent the proportion of signal j's energy falling within the receiving window of signal i in the frequency domain. After multiplying, multiply by the peak power of signal i. This yields the cross-energy metric of the two signals in the time-frequency two-dimensional domain. The larger the cross-energy, the stronger the actual physical interference of signal j on signal i.

[0042] Time-frequency crossover energy With preset threshold If a comparison is made, Exceed If the signal pair (i,j) has a valid collision under the current mapping scheme, then it is determined that the signal pair (i,j) has a valid collision; if No more than If a signal is considered to be in conflict, then the two signals are considered to coexist without conflict. This process is repeated for all signal pairs, and the number of signal pairs deemed to be in conflict is accumulated to obtain the mapping scheme for the nth round. Conflict metrics .

[0043] It is important to emphasize that the conflict determination method described above is fundamentally different from the traditional discrete time slot-wavelength occupancy rule. Under the traditional rule, as long as two signals occupy the same wavelength and their time slots overlap, they are considered to be in conflict, regardless of the degree of overlap, and the severity of the conflict is completely ignored. In this embodiment, however, conflict determination is based on the product of the time-domain overlap integral, the spectral overlap integral, and the peak power, which is a calculation process based on continuous physical quantities. The time-frequency cross-energy of two signals that only overlap at the edge and those that completely overlap may differ by several orders of magnitude. This difference in the magnitude of interference can be effectively distinguished by the conflict determination method of this embodiment. This means that even if both signals are determined to be in conflict under the discrete rule, this embodiment can still reflect the difference in the severity of interference between different signal pairs through the specific value of the time-frequency cross-energy, providing richer gradient information for the subsequent optimization process.

[0044] Preset threshold The value of is determined by the receiver sensitivity and bit error rate requirements; this embodiment does not limit its specific value. In actual deployment, it can be comprehensively determined based on the characteristics of the photodetector at the receiver front end, the upper limit of bit error rate tolerance, and the system's tolerance to interference. For example, for systems with high error rate requirements, it is possible to... Setting it to a lower value means that even weaker time-frequency crossover energy will be considered a conflict, thus more strictly avoiding potential interference during the mapping optimization process; for systems with relatively relaxed bit error rate requirements, it can be... Setting it to a higher level allows for a certain degree of weak interference in exchange for a larger search space for mapping schemes.

[0045] The conflict metric value determined by the above method The interference metric described in the foregoing embodiments Although both originate from the time-frequency overlap between various signals, they describe information at different levels. The interference metric describes the continuity of signal quality loss, while the collision metric describes the number of collision events. Both are unified in the same objective function for collaborative optimization, enabling the mapping scheme search process to simultaneously consider signal quality preservation and collision avoidance.

[0046] In one specific embodiment, an interference metric is further provided. The method for determining the nth round of mapping schemes. For any target signal i, multiply the peak power of the other signals by the time-domain overlap integral, spectral overlap integral, and polarization matching factor between the target signal and the target signal, respectively, and then sum them up to obtain the total interference power of the target signal i. :

[0047] in, Let be the peak power of the j-th signal. The polarization matching factor between the j-th signal and the i-th signal is... and These are the time-domain overlap integral and the spectral overlap integral defined in the aforementioned embodiments, respectively.

[0048] Specifically, total interference power This characterizes the sum of interference power caused by all signals other than the target signal i to the target signal i under the current mapping scheme. For each interference signal j, its interference contribution to the target signal i is determined by four factors: the peak power of the interference signal j... The polarization matching factor between the interference signal j and the target signal i Temporal overlap integral and spectral overlap integral Peak power The energy intensity of the interference signal itself is determined; the time-domain overlap integral. Characterizing the proportion of the energy of the interference signal j that falls within the receiving window of the target signal i in the time domain; spectral overlap integral The polarization matching factor represents the proportion of the energy of the interference signal j that falls into the receiving filter of the target signal i in the frequency domain. This characterizes the degree of matching between the polarization states of the two signals. Multiplying these four factors together yields the actual interference power contribution of the interfering signal j to the target signal i.

[0049] Polarization matching factor The physical meaning can be understood as follows: Two optical signals each have a specific polarization state during transmission. When the polarization states of the two signals are perfectly aligned, the electric field vector of the interfering signal j is completely consistent with the received polarization direction of the target signal i. The interference energy can be completely coupled to the receiver of the target signal i. At this time... The maximum value is 1; when the polarization states of the two signals are completely orthogonal, the electric field vector of the interference signal j is perpendicular to the receiving polarization direction of the target signal i, and the interference energy cannot be coupled by the receiver of the target signal i. Take the minimum value of 0; when the polarization states of the two signals are between perfectly aligned and perfectly orthogonal, Values ​​range from 0 to 1. Polarization matching factor The value range is [0,1]. Its specific value can be obtained by actual measurement with a polarization measuring instrument, or it can be calculated based on the relative position of the polarization state vectors of the two signals on the Poincaré sphere.

[0050] To obtain the total interference power Then, the signal-to-interference-plus-noise ratio (SIR) of the target signal i at the receiver is calculated based on the total interference power. :

[0051] in, Let be the effective signal power of the i-th signal. For noise power spectral density, This is the receiver's equivalent noise bandwidth.

[0052] Specifically, signal-to-interference-plus-noise ratio It is the actual usable signal quality indicator at the receiving end, and its numerator is the effective signal power. The denominator is the total interference power. With receiver noise power The sum of the total interference power. When the value is zero, the signal-to-interference-plus-noise ratio (SIR) reaches its maximum value, which is the ideal SIR under interference-free conditions. With the total interference power As the noise level increases, the signal-to-interference-plus-noise ratio (SIR) decreases monotonically, and the signal quality gradually deteriorates.

[0053] Effective signal power It is determined by the peak power and the time-domain envelope autocorrelation integral and the spectral envelope autocorrelation integral. Specifically, the effective signal power... It can be expressed as peak power Multiply by the time-domain envelope autocorrelation integral Multiply by the spectral envelope autocorrelation integral The time-domain envelope autocorrelation integral reflects the contribution of the pulse's time-domain width to the received energy, while the spectral envelope autocorrelation integral reflects the contribution of the spectral width to the received energy. The narrower the pulse and the narrower the spectrum, the higher the effective signal power.

[0054] After obtaining the signal-to-interference-plus-noise ratio (SIR) of each signal, the deterioration of the actual SIR of each signal relative to the ideal SIR under interference-free conditions is weighted and summed to obtain the nth round mapping scheme. Interference metric value :

[0055] in, Let be the priority weight of the i-th signal.

[0056] Specifically, for each signal i, the expression within square brackets represents the degradation of the actual signal-to-interference-plus-noise ratio (SIR) of that signal under the current mapping scheme relative to the ideal SIR without interference. This is true when the signal is unaffected by any interference. at this time, This equals the ideal signal-to-interference-plus-noise ratio (SNR), with zero degradation; as the interference power increases, The degradation rate decreases, and the degradation amount monotonically increases. The priority weight wi is used to reflect the relative importance of different signals in the system; for example, the weight of high-priority service signals can be set larger, and the weight of low-priority service signals can be set smaller. The interference metric of the entire mapping scheme is obtained by summing the degradation amounts of each signal according to its priority weight. .

[0057] It is important to emphasize that the interference metric value It reflects signal quality loss rather than a simple collision count, and is consistent with collision metrics. They complement each other. Conflict metric. The answer is "how many pairs of signals collided", and its value is a discrete integer; interference metric. The answer is "how severe is the quality loss of each signal?", and its value is a continuous real number. Although both originate from the time-frequency overlap between signals, they describe information at different levels. In optimizing the objective function... In the middle, interference metric It provides continuous gradient information, guiding the search process towards directions with minimal signal quality loss; conflict metric. Provides discrete constraint information, in dynamic weight coefficients under the action of , the search process is gradually forced to converge towards zero conflict. The synergistic effect of the two enables the mapping optimization process to take into account both signal quality maintenance and conflict avoidance.

[0058] The interference metric value determined in the above manner quantifies the signal quality loss in the form of a continuous value, and is synergistically optimized with the conflict metric value in the same objective function. Compared with the conventional scheme that only uses the presence or absence of conflict as the optimization basis, the interference metric value of this embodiment can distinguish subtle differences between different mapping schemes at the signal quality level, and provide richer gradient guidance information for iterative search.

[0059] In a specific embodiment, the specific form of the dynamic weight coefficient is further provided . The dynamic weight coefficient increases superlinearly with the increase of the iteration number n, as shown in Figure 5 , and its expression is:

[0060] wherein, is the initial weight coefficient, and p≥1.

[0061] Specifically, the meaning of superlinear growth can be understood from two levels. First, increases monotonically with the increase of the iteration number n, that is, the weight coefficient of the subsequent round is always greater than the weight coefficient of the previous round. Second, the increasing rate of increases with the increase of the iteration number n, that is, the difference between the weight coefficients of two adjacent rounds is itself a quantity that increases as n increases. Taking as an example, when , ; when , ; when , . It can be seen that from the 0th round to the 10th round, the weight coefficient increases by about 120 times; while from the 10th round to the 100th round, the weight coefficient increases by about 84 times. This continuous amplification of the growth rate causes the relative cost of the conflict metric value in the objective function to rise sharply at a speed far exceeding the linear speed in the later stage of iteration.

[0062] It should be noted that the value of the exponent p is not limited to . When , , the weight coefficient increases linearly with the number of iterations, and at this time the increasing rate remains constant; when The rate of increase increases with the number of iterations; the larger p is, the more drastic the increase in the weight coefficient in later stages. In practical deployments, the value of p can be determined comprehensively based on the number of signal paths, the number of available wavelength channels, and the system's requirements for convergence speed. For example, in scenarios with a large number of signal paths and a large solution space, a larger p value can be chosen to accelerate later convergence; in scenarios with a small number of signal paths and a relatively compact solution space, a smaller p value can be chosen. A value slightly greater than 1 is sufficient to meet the convergence requirement.

[0063] Initial weighting coefficients The value is the average of the ideal signal-to-interference-plus-noise ratio (SINR) of each signal channel:

[0064] in, Let be the effective signal power of the i-th signal. For noise power spectral density, This is the receiver's equivalent noise bandwidth.

[0065] Specifically, This represents the ideal signal-to-interference-plus-noise ratio (SNR) of the i-th signal under interference-free conditions. Physically, it is the maximum SNR the receiver can achieve when this signal is not interfered with by any other signal. The initial weighting coefficients are obtained by taking the arithmetic mean of the ideal SNRs of all signals. .

[0066] The physical basis for this value lies in: in the 0th iteration, i.e. hour, At this point, the penalty term for the conflict metric in the objective function is... .when When a conflict exists in the mapping scheme, the penalty term takes the value of... This is exactly equal to the average of the ideal signal-to-interference-plus-noise ratio (SNR) of each signal. Meanwhile, the interference metric... The magnitude of the interference metric is determined by the weighted sum of the signal-to-interference-plus-noise ratio (SNR) degradation of each signal, and the contribution of the interference metric of a single signal is comparable in magnitude to the ideal SNR. Therefore, in the 0th iteration, the penalty for a single collision is on the same order of magnitude as the interference metric of a single signal, and the two have comparable relative weights in the objective function. This design avoids the penalty term excessively dominating the search direction in the early stages of optimization: if If the value is set too high, the objective function will be dominated by the conflict penalty term from round 0 onwards. The search process will be prematurely limited to the zero-conflict subspace, making it impossible to transition to a globally better mapping structure through intermediate solutions that contain conflicts but have low interference metrics. If the value is set too low, the proportion of the conflict penalty term in the objective function will be too small. The search process may still accept schemes containing conflicts in the later stages of iteration, and will not be able to effectively converge to a zero-conflict mapping scheme.

[0067] By combining the superlinear growth form with the initial weight coefficient values, the overall design effectiveness of the dynamic weighting mechanism can be further explained. In the early stages of iteration, due to... The values ​​are on the same order of magnitude as the interference metric, and the interference and conflict metrics contribute comparablely to the objective function. The search process can fully explore the solution space containing some conflict, finding mapping structures with lower interference metrics. As the iterations progress, It grows at a superlinear rate, and the relative weight of the conflict metric in the objective function continues to increase. As the iteration progresses to later stages... The value of is already far greater than the typical magnitude of the interference metric. Any scheme containing conflicts will have a much higher objective function value than a zero-conflict scheme. Even if the conflict-containing scheme is slightly better than the zero-conflict scheme in terms of interference metric, its objective function value will be at a disadvantage due to the contribution of the conflict penalty term. This dynamic evolution of the weight structure allows the search process to have sufficient exploratory capabilities in the early stages, and is forced to the zero-conflict mapping scheme in the later stages, thus achieving an adaptive balance between exploration and convergence.

[0068] It is important to emphasize that the dynamic weighting coefficient The specific form described above is merely one implementation method given in this embodiment. In other embodiments, the dynamic weight coefficient can also adopt other function forms that monotonically increase with the iteration rounds and whose rate of increase increases with the iteration rounds, such as exponential function forms, polynomial function forms, or other function forms that satisfy the conditions of monotonically increasing and increasing rate of increase, as long as they can form a strong repulsion effect against conflicting schemes in the later stages of iteration. Initial weight coefficient The value of is not limited to the average of the ideal signal-to-interference-plus-noise ratio of each signal; it can also be appropriately scaled according to the system's requirements for the stringency of collision avoidance. For example, in scenarios where earlier convergence to a zero-collision scheme is required, can be... Adjust the value appropriately; in scenarios where a more thorough exploration of the solution space is required, it can be increased. Adjust it appropriately.

[0069] In one specific embodiment, three specific methods of local perturbation are further provided. The local perturbation method is any one of the following three methods: Method 1, randomly select one signal from all signals and randomly shift its time slot position within a preset range; Method 2, randomly select one signal from all signals and reallocate its wavelength channel to other available wavelength channels; Method 3, randomly select two signals from all signals and swap their time slot positions.

[0070] Specifically, Method 1 corresponds to the time slot drift operation. Let the current mapping scheme be... There are N signals in total, and the time slot positions of each signal are as follows: In execution mode one, one signal is randomly selected from these N signals with equal probability, denoted as signal i, and then a random offset is generated. The time slot position of signal i is changed from Adjusted to Random offset The generation range is determined by the preset upper limit of the time slot drift step size, for example... From Uniform sampling within the interval, where This is the preset maximum time slot drift step size. The offset time slot position needs to be maintained within the system frame period. Within the defined effective range, if the offset exceeds the range, it can be processed by truncation, wrapping, or resampling. This embodiment does not limit this. Method 1 only changes the time slot position of signal i, while the time slot positions of the other signals and the wavelength channels of all signals remain unchanged.

[0071] Method 2 corresponds to wavelength reallocation. When executing Method 2, one signal, denoted as signal i, is randomly selected from the N signals with equal probability. Then, a new wavelength channel is selected for signal i from the available wavelength set Λ. During reselection, the wavelength channel currently occupied by signal i can be excluded from the available wavelength set Λ, and one of the remaining wavelength channels can be randomly selected with equal probability as the new wavelength channel; alternatively, a new wavelength channel can be randomly selected directly from the available wavelength set Λ. If the selected channel is the same as the current wavelength channel, a new channel is selected. Method 2 only changes the wavelength channel of signal i; the wavelength channels of all other signals and the time slot positions of all signals remain unchanged.

[0072] Method 3 corresponds to a time slot swapping operation. When executing Method 3, two different signals, denoted as signal i and signal j, are randomly selected from N signals with equal probability. Then, their time slot positions are swapped; that is, signal i acquires the original time slot position of signal j, and signal j acquires the original time slot position of signal i. Method 3 changes the time slot positions of both signals simultaneously, while the wavelength channels of all signals remain unchanged.

[0073] The selection mechanism for the three methods can be flexibly configured. In one implementation, the three methods are selected randomly with equal probability, meaning that each time a local perturbation is performed, the probability of selecting method one, method two, and method three is 1 / 3. In another implementation, the selection probabilities of the three methods can be adjusted according to the actual scenario. For example, when there are many available wavelength channels but time slot resources are relatively scarce, the selection probability of method two can be appropriately increased, making the search process more inclined to explore in the wavelength dimension; when the time slot positions of each signal are relatively evenly distributed but there are obvious conflicts in wavelength channel allocation, method two can also be selected first. Furthermore, when the conflicts in the mapping scheme are mainly concentrated in the time slot dimension, the selection probabilities of method one and method three can be appropriately increased, making the search process more inclined to adjust in the time slot dimension. The specific values ​​of the selection probabilities can be comprehensively determined based on the scale of the number of signal paths, the number of available wavelength channels, the frame period length, and the system's requirements for convergence speed; this embodiment does not impose any limitations.

[0074] From the perspective of the impact range on the mapping scheme, methods one and two only change the parameters of one signal, which are single-point disturbances; method three changes the time slot positions of two signals simultaneously, which is a two-point disturbance. The impact range of a single-point disturbance is small; the difference between the candidate mapping scheme and the current mapping scheme only involves one signal. When calculating the objective function value of the candidate mapping scheme, it is only necessary to recalculate the various indicators of the signal pair related to that signal. The indicators of the other signal pairs can reuse the values ​​already calculated and stored in the current mapping scheme. The impact range of a two-point disturbance is slightly larger; the difference between the candidate mapping scheme and the current mapping scheme involves two signals. Subsequent calculations require recalculating the various indicators of the signal pairs related to these two signals. Regardless of the method used, the modification magnitude of the mapping scheme by a local disturbance is much smaller than that of the overall reconstruction, and the candidate mapping scheme and the current mapping scheme always maintain a local correlation.

[0075] The purpose of local perturbation is to make only small modifications to the mapping scheme each time, so that the candidate scheme and the current scheme have local correlation, which is beneficial to the stable progress of the iterative search. Specifically, since there are only local differences between the candidate mapping scheme and the current mapping scheme, the change in the objective function value mainly comes from the local signal pairs involved in the modified signal. The search process can move gradually in the solution space with small step sizes, avoiding missing high-quality mapping structures with low interference and conflict metrics due to large jumps. At the same time, local correlation also enables the incremental update strategy to be effectively implemented, and the computational cost of each iteration can be controlled at a low level. It should be noted that the specific implementation of local perturbation is not limited to the three mentioned above. In other embodiments, other types of local perturbation operations can also be used, such as simultaneously adjusting the time slot position and wavelength channel of one signal, or slightly shifting the time slot position of one signal while reallocating the wavelength channel of another signal, as long as each perturbation only makes small modifications to the mapping scheme and maintains the local correlation between the candidate scheme and the current scheme.

[0076] In one specific embodiment, an incremental update method for optimizing the objective function value of the candidate mapping scheme is further provided. The candidate mapping scheme is calculated in step S4. When optimizing the objective function value, an incremental update method is used: only candidate mapping schemes are updated. For signal pairs involving signals whose parameters have changed, recalculate their time-domain overlap integral, spectral overlap integral, time-frequency cross energy, and conflict determination results, as well as recalculate the interference power associated with the signal whose parameters have changed; for the remaining signal pairs, directly reuse the nth round mapping scheme for their time-domain overlap integral, spectral overlap integral, time-frequency cross energy, conflict determination results, and interference power. The corresponding value that has already been calculated and stored.

[0077] Specifically, the core idea of ​​the incremental update method is to avoid redundant calculations by utilizing the local correlation between the candidate mapping scheme and the current mapping scheme. As described in the previous embodiments, the local perturbation only slightly modifies the mapping scheme each time, and there are only parameter differences between one or two signals between the candidate mapping scheme and the current mapping scheme. Since the time-domain overlap integral, spectral overlap integral, time-frequency cross energy, and conflict determination results between the signals depend only on the time slot position and wavelength channel of the two signals involved, when only a few signals change their parameters, the calculated values ​​of most signal pairs will not change accordingly, and there is no need to recalculate. The incremental update method utilizes this property to reduce the calculation scope of each iteration from all signal pairs to those signal pairs that only involve signals whose parameters have changed.

[0078] The following explanation uses Method 1 (time slot drift) from the above embodiments as an example. Assume that in the nth round of the mapping scheme... In the process, the local disturbance selects signal i and changes its time slot position from... Adjusted to The time slot positions and wavelength channels of all other signals remain unchanged. At this point, the candidate mapping scheme... Compared with the current mapping scheme The difference lies only in the time slot position of signal i. For any pair of signals (a, b) that does not contain signal i, where a ≠ i and b ≠ i, since the time slot positions and wavelength channels of signals a and b remain unchanged, the time-domain overlap integral between them is... Spectral overlap integral Time-frequency cross energy And the conflict determination results are all consistent with the mapping scheme in the nth round. The values ​​in the original text are completely identical, and the stored corresponding values ​​can be directly reused. What truly needs recalculation are the N-1 pairs of signals containing signal i, namely signal pairs (i,1), (i,2), ..., (i,i-1), (i,i+1), ..., (i,N). For each of these N-1 pairs of signals (i,j), the time-domain overlap integral needs to be recalculated. and spectral overlap integral Therefore, the time-frequency crossover energy is recalculated. and will new With preset threshold The comparison is used to update the conflict determination results.

[0079] Regarding the interference power, only the part related to signal i needs to be recalculated. For the target signal i itself, the total interference power it experiences... The total interference power needs to be recalculated based on the new time-domain overlap integral and spectral overlap integral between the remaining signals j and i. In the calculation, only the interference contribution from signal i changes with the time slot position of signal i; the interference contributions from other signals remain unchanged. Therefore, in the recalculation... At that time, it is only necessary to update the interference contribution term of signal i to signal j, that is For this item, the other items directly reuse the stored values.

[0080] Taking method two (wavelength redistribution) as an example, assume that the local disturbance selects signal i and changes its wavelength channel from Reassigned to The wavelength channels of all other signals and the time slot positions of all signals remain unchanged. At this point, for any pair of signals (a, b) that does not contain signal i, its spectral overlap integral... Overlapping integrals in the time domain Everything remains unchanged, and all calculated values ​​can be directly reused. Only the spectral overlap integral of N-1 pairs of signals containing signal i needs to be recalculated. and the resulting time-frequency crossover energy Changes in conflict determination results. Temporal overlap integral. Since the time slot positions of signals i and j remain unchanged, they can be directly multiplexed. Regarding interference power, the total interference power experienced by signal i is... The total interference power of other signals j needs to be recalculated based on the new spectral overlap integral. The interference contribution term from signal i needs to be updated.

[0081] Taking Method 3 (time slot swapping) as an example, assume that a local disturbance selects signal i and signal j and swaps their time slot positions. In this case, for any signal pair (a, b) that contains neither signal i nor signal j, all calculated values ​​remain unchanged and can be directly reused. What needs to be recalculated are signal pairs that contain either signal i or signal j. Specifically, there are N-1 signal pairs containing signal i and N-1 signal pairs containing signal j. Since signal pair (i, j) contains both signal i and signal j, it is counted twice. Therefore, the actual number of signal pairs that need to be recalculated is 2(N-1)-1=2N-3 pairs. For these signal pairs, the time-domain overlap integral needs to be recalculated, and the time-frequency crossover energy and conflict determination results updated accordingly. The spectral overlap integral can be directly reused since the wavelength channels of all signals remain unchanged. Regarding interference power, the total interference power received by signals i and j needs to be recalculated, and the interference contribution terms from signals i and j in the total interference power received by other signals k (k≠i and k≠j) need to be updated.

[0082] From a computational complexity perspective, incremental updates offer a significant advantage over global recalculation. In global recalculation, each iteration requires checking all N(N-1) / 2 pairs of signals, recalculating the time-domain overlap integral, spectral overlap integral, time-frequency cross-energy, and conflict determination results for each pair, resulting in a computational complexity of O(N²). In incremental updates, taking methods one and two as examples, each iteration only requires checking N-1 pairs of signals, with a computational complexity of O(N). For method three, each iteration requires checking 2N-3 pairs of signals, still with a computational complexity of O(N). When the number of signal paths N is large, the difference in computational complexity becomes very significant. For instance, when N=100, global recalculation requires checking 4950 pairs of signals, while incremental updates only require checking 99 pairs, reducing computation by approximately 50 times; when N=1000, global recalculation requires checking 499500 pairs of signals, while incremental updates only require checking 999 pairs, reducing computation by approximately 500 times. This means that in optical pulse signal mapping scenarios with thousands of signal paths, the incremental update method can compress the computation time of a single iteration from seconds or even minutes under the global recalculation method to milliseconds, so that the entire iterative optimization process can be completed within an acceptable time range.

[0083] It should be noted that the implementation of the incremental update method relies on the storage unit saving the calculated values ​​from each round. In each iteration, the processing unit calculates the current mapping scheme... After calculating the time-domain overlap integral, spectral overlap integral, time-frequency cross energy, collision determination results, and interference power of each signal pair, these calculated values ​​need to be written to the storage unit. In the next iteration, when the processing unit needs to calculate candidate mapping schemes... When calculating the objective function value, the previously saved calculated values ​​are read from the storage unit. Only signal pairs whose parameters have changed are recalculated; the calculated values ​​for other signal pairs are directly read from the storage unit and reused. The historical mapping scheme data stored in the storage unit includes the temporal overlap integral value, spectral overlap integral value, conflict metric value, and interference metric value under each round of mapping scheme. This data forms the basis for data reuse in incremental update calculations.

[0084] The incremental update method is applicable not only to the three local perturbation methods given in the above embodiments, but also to other types of local perturbation operations. As long as the difference between the candidate mapping scheme and the current mapping scheme involves only parameter changes in a few signals, the incremental update method can reduce the computational scope to signal pairs related to the signals whose parameters have changed, and the calculated values ​​of the remaining signal pairs can be directly reused. The fewer the number of signals whose parameters have changed, the more obvious the advantage of the incremental update method. In extreme cases, when the local perturbation only changes one parameter of one signal, the computational complexity of the incremental update method is minimized, requiring only the recalculation of the relevant indices of N-1 pairs of signals.

[0085] By using the incremental update method described above, this embodiment reduces the computational complexity of each iteration from O(N²) of global recalculation to O(N) involving only the modified signal. This significantly reduces the computational overhead of a single iteration when the number of signal paths is large, making the conflict avoidance method of this application applicable to large-scale optical pulse signal mapping scenarios.

[0086] In one specific embodiment, the calculation method for the set probability P is further given. In step S4, when the candidate mapping scheme The optimization objective function value is not less than the mapping scheme in the nth round. When optimizing the objective function value, the candidate mapping scheme is not directly rejected, but accepted with a set probability P. The set probability P is calculated as follows:

[0087] in, The temperature parameter corresponding to the nth round. It decreases monotonically as the iteration number n increases.

[0088] Specifically, the physical meaning of the above acceptance criteria can be derived from the difference. We can understand this in two cases: positive and negative. When the objective function value of the candidate mapping scheme is better than the current mapping scheme, that is... When the difference is negative, the exponent is positive, and P is greater than 1. Since the upper limit of the probability is 1, the candidate mapping scheme is unconditionally accepted. This is consistent with the description in step S4 of the previous embodiment of "if..." , indicating the candidate mapping scheme It is superior to the current [number] in terms of the overall cost of interference and conflict. Wheel mapping scheme The candidate mapping scheme As the mapping scheme for the (n+1)th round , that is to say The judgment logic is consistent with ";". When the objective function value of the candidate mapping scheme is not better than the current mapping scheme, that is... ， When the difference is non-negative, the exponent is negative or zero, and the value of P falls between 0 and 1. The larger the difference, the larger the absolute value of the exponent, the closer P is to zero, and the lower the probability of accepting a poor candidate mapping scheme; the smaller the difference, the closer P is to 1, and the higher the probability of accepting a poor candidate mapping scheme.

[0089] Temperature parameters In the calculation of the set probability P, it plays a role in adjusting the overall level of the acceptance probability. When the difference is the same, the temperature parameter... The higher the value, the smaller the absolute value of the exponent, the closer P is to 1, and the greater the probability of accepting a poor candidate mapping scheme; temperature parameter The lower the temperature, the larger the absolute value of the exponent, the closer P is to zero, and the lower the probability of accepting a poor candidate mapping scheme. Temperature parameter The design of the algorithm monotonically decreases as the iteration number n increases, which gradually weakens the tendency to accept poor candidate mapping schemes as the iteration progresses.

[0090] From the overall behavior of iterative search, the temperature parameter The monotonically decreasing design has a clear optimization guiding role. In the early stages of iteration, the temperature parameter... A higher value means that even if the objective function value of a candidate mapping scheme is significantly worse than the current mapping scheme, it still has a non-negligible probability of being accepted. This tolerance for inferior schemes allows the search process to traverse regions where the objective function value locally increases, jumping away from the vicinity of a locally optimal mapping scheme and continuing to search for a better mapping structure in the solution space. Without this escape mechanism, once the search process enters a locally optimal mapping scheme, the search process cannot continue because the objective function values ​​of all locally perturbation candidate schemes for that locally optimal mapping scheme are not better than its own. The quality of the final output mapping scheme is limited by the initial mapping scheme and the early search path. In the later stages of iteration, the temperature parameter... As the value approaches zero, the probability of accepting a poor candidate mapping scheme approaches zero, and the search process gradually degenerates into a greedy descent algorithm, that is, only accepting candidate mapping schemes with better objective function values ​​and no longer accepting any poor schemes. At this point, the search process makes fine adjustments around the high-quality mapping structures discovered in the early exploration, gradually approaching the mapping scheme with the lowest objective function value.

[0091] Temperature parameters There are multiple implementations for the specific scheduling method. In one implementation, the temperature parameter... Exponential cooling method is used:

[0092] in, The initial temperature. The cooling coefficient is 0. It should be noted that the temperature parameter The scheduling method is not limited to the exponential cooling form described above. In other embodiments, the temperature parameter... Linear cooling methods can also be used, for example... ,in The maximum number of iterations can be preset; alternatively, a segmented cooling method can be used, maintaining a higher temperature in the early stages of iteration, rapidly cooling down in the middle stages, and maintaining a lower temperature in the later stages. Only the temperature parameter needs to be considered. The temperature parameter decreases monotonically as the iteration number n increases, and it can provide sufficient exploration capability in the early stage of iteration and sufficient convergence pressure in the later stage of iteration, thus satisfying the acceptance criterion requirements for temperature parameters in this embodiment.

[0093] It should also be noted that the calculation method for the set probability P is not limited to the exponential function form mentioned above. In other embodiments, other function forms that monotonically decrease with increasing difference and monotonically increase with increasing temperature parameter can also be used, such as linear function forms or piecewise function forms, as long as the judgment logic of "unconditionally accepting the better candidate solution, accepting the worse candidate solution with a certain probability, and the acceptance probability decreasing with increasing degree of deterioration" can be achieved.

[0094] Based on the above calculation method for probability P and temperature parameters In this embodiment, a temperature-controlled probability acceptance mechanism is introduced in the acceptance decision step S4, which is based on a monotonically decreasing design. This mechanism enables the iterative search process to escape local optima in the early stages and gradually degenerates into a greedy descent to approximate the mapping scheme with the lowest objective function value in the later stages, thereby achieving a dynamic balance between exploring the solution space and converging to a high-quality mapping scheme.

[0095] In one specific embodiment, the specific details of the convergence condition are further given. In step S5, let... Next, it is determined whether the preset convergence conditions are met. The convergence conditions include the following three conditions: Condition 1, the current mapping scheme. Conflict metrics Condition 2, the current mapping scheme Interference metric value Less than the preset threshold Condition 3: The current iteration number n reaches the preset maximum iteration number. When either condition one or condition two is met, the mapping scheme is considered to have converged effectively, and the current mapping scheme is output. If condition three is satisfied but conditions one and two are not, the process is considered a timeout and terminates, outputting the optimal mapping scheme recorded during the current iteration. Specifically, condition one corresponds to a zero-conflict mapping scheme. When the conflict metric... This means that under the current mapping scheme, no pair of signals has a time-frequency crossover energy exceeding a preset threshold. No effective collisions occurred between the various signals in the time-frequency two-dimensional domain. At this point, regardless of the interference metric... Regardless of the value of , the iteration process is considered to have converged effectively and the current mapping scheme is output. The design basis of condition one is that conflict avoidance is one of the core objectives of this application. Once the search process finds a zero-conflict mapping scheme, although continuing the iteration may further reduce the interference metric, the zero-conflict state itself has already met the basic requirements of the mapping scheme at the conflict level, so the scheme can be output.

[0096] Condition 2 corresponds to an interference metric that meets the standard. When the interference metric... Less than the preset threshold At this point, it means that the weighted sum of the signal-to-interference-plus-noise ratio (SNR) degradation of each signal under the current mapping scheme has been reduced to within the preset accuracy, and the signal quality loss is within an acceptable range. At this point, even if a small number of conflicts still exist in the mapping scheme, i.e. If the value is not zero, the iteration process is still considered to have converged effectively, and the current mapping scheme is output. Condition two is based on the interference metric. It reflects the continuity of signal quality loss. When this value is small enough, it means that the actual physical interference between the signals has been reduced to a level that can be ignored in terms of its impact on system performance. At this point, even if there are a few signal pairs that are judged to be in conflict, the actual impact of these conflicts on signal quality is already weak enough, and the marginal benefit of continuing to iterate is limited.

[0097] The difference between condition one and condition two can be understood from the relationship between conflict metrics and interference metrics. Conflict metrics The answer is the number of signal collisions, and its value is a discrete integer; interference metric. This answers the question of how severe the quality loss of each signal is, and its value is a continuous real number. Condition one corresponds to the state where collisions are completely eliminated, and condition two corresponds to the state where the signal quality loss meets the standard. In most cases, the interference metric of the zero-collision mapping scheme is usually also low, but the two are not entirely equivalent. For example, when a preset threshold is set... When the setting is low, even if the time-frequency crossover energy between each signal is very small, some signal pairs may still be judged as collisions. Not zero, but It may have already fallen below the preset threshold. Condition two is satisfied before condition one. For example, when a preset threshold is set... When the value is set too high, the search process may find a zero-collision mapping scheme earlier, but the interference metric of that scheme will be higher. It has not yet dropped to the preset threshold. In the following, condition one is satisfied before condition two. The two conditions are listed in parallel with an OR logic, and the convergence is determined to be valid if either condition is satisfied. This ensures that the iterative process can terminate in a timely manner when the conflict is completely eliminated or when the signal quality loss reaches the target, thus avoiding unnecessary iteration overhead.

[0098] Condition three corresponds to timeout termination. This occurs when the number of iterations n reaches the preset maximum number of iterations. However, if neither condition one nor condition two is met, it means that within the preset iteration budget, the search process has neither found a zero-conflict mapping scheme nor reduced the interference metric to below the preset threshold. In this case, the iteration process is determined to have timed out and terminates, outputting the optimal mapping scheme recorded during the current iteration. The purpose of condition three is to ensure that the algorithm will terminate under limited computational resources. Regardless of the complexity of the solution space or the quality of the initial mapping scheme, as long as the number of iterations reaches [a certain threshold], the algorithm will terminate. The algorithm will stop and will not loop indefinitely. Optimal mapping scheme. The maintenance method is as follows. Before the iteration begins, the initial mapping scheme is... Record as the optimal mapping scheme and its objective function value Record as the optimal objective function value In each iteration, when the mapping scheme for the (n+1)th round is determined... Next, the value of the mapping scheme under the corresponding optimization objective function is calculated. Compare it with the currently recorded optimal objective function value. Compare. If Less than Then Record as the new optimal mapping scheme and will Record as the new optimal objective function value Otherwise, keep and Unchanged. It should be noted that due to the dynamic weighting coefficients... The objective function value varies with the number of iterations n. and Strictly speaking, they are not on the same weighting scale. In one implementation, the dynamic weighting coefficient of the current round is used uniformly for comparison. To calculate the objective function value, i.e., to compare and ,in This represents the dynamic weighting coefficient in the nth round. Recalculate The objective function value. In another implementation, interference metrics can also be compared directly. and The mapping scheme with the lowest interference metric is selected as the optimal mapping scheme. Both maintenance methods can guarantee the output when timeout occurs. It is the mapping scheme with the best overall performance during the iteration process, and this embodiment does not limit it.

[0099] Preset threshold The value of can be determined based on the system's signal quality requirements. For example, in systems with high bit error rate requirements, can be... Setting it to a smaller value means that the iteration process needs to reduce the interference metric to an even lower level to satisfy condition two; in systems with relatively relaxed bit error rate requirements, it can be... Setting the maximum number of iterations makes condition two easier to satisfy, allowing the iteration process to terminate earlier. The value of can be determined based on the number of signal paths, the number of available wavelength channels, and the system's constraints on computation time. The more signal paths, the larger the solution space. It typically needs to be set larger to give the search process a sufficient iteration budget.

[0100] By combining the above three conditions, this embodiment implements the complete logic for convergence determination in step S5. Conditions one and two define effective convergence from the two dimensions of conflict elimination and signal quality compliance, respectively, while condition three, as a timeout protection, ensures that the algorithm will inevitably terminate under limited computing resources. Simultaneously, the optimal mapping scheme... The maintenance mechanism ensures that even if a timeout occurs, the output mapping scheme is the best performing scheme experienced during the iteration process, rather than the scheme obtained in the last iteration.

[0101] Example 2 like Figure 6 As shown, this embodiment provides a time-frequency resource mapping conflict avoidance device, which is used to execute the time-frequency resource mapping conflict avoidance method described in any of the embodiments. The device includes seven components: a parameter acquisition unit 61, a mapping initialization unit 62, an objective function construction unit 63, a local perturbation unit 64, a comparison and judgment unit 65, an iterative control unit 66, and an output unit 67. Each component works collaboratively according to a determined data flow direction to jointly complete the entire processing from physical parameter acquisition to the final mapping scheme output.

[0102] The parameter acquisition unit 61 is used to acquire the physical parameters and system configuration parameters of the multi-channel optical pulse signals.

[0103] Specifically, physical parameters refer to parameters related to the physical characteristics of each optical pulse signal, including but not limited to pulse width, spectral width, peak power, polarization state, and priority weight. System configuration parameters refer to parameters related to the mapping resource pool and receiver operating conditions, including but not limited to the available wavelength set, frame period, collision determination threshold, and receiver noise parameters. The parameter acquisition unit can receive these parameters through the communication interface with the upper control plane, or read them through a manual configuration interface or a local parameter file. The specific implementation form of the parameter acquisition unit is not limited; for example, it can be a management interface module, a configuration register group, or a data transceiver port that interfaces with the network management system in an optical communication network. After acquiring the physical parameters and system configuration parameters, the parameter acquisition unit transmits these parameters to the mapping initialization unit for processing.

[0104] Mapping initialization unit 62 is used to generate an initial mapping scheme based on the physical parameters. The mapping scheme includes the time slot position and wavelength channel allocated to each signal, and sets the initial iteration round n=0.

[0105] Specifically, after receiving physical parameters from the parameter acquisition unit, the mapping initialization unit generates an initial mapping scheme according to a preset initialization strategy. The initialization strategy can be random, i.e., uniformly and randomly selecting time slots for each signal within a frame period, and uniformly and randomly selecting wavelength channels from the available wavelength set; or it can be generated based on a certain preset rule. Initial mapping scheme It provides a search starting point for subsequent iterative optimizations, but its quality does not directly determine the quality of the final solution, as subsequent iterations continuously improve the mapping scheme. The mapping initialization unit is generated... Afterwards, The parameters and related data are transmitted to the objective function construction unit, and the initial mapping scheme data is stored in the storage unit (not shown in the figure) for reuse in subsequent iterations.

[0106] Objective function construction unit 63 is used to construct the mapping scheme for the nth round. The optimization objective function is as follows: ,in, For the nth round of mapping scheme The interference metric value is used to characterize the degree of mutual interference between signals in the time and frequency domain under the current mapping scheme; For the nth round of mapping scheme The collision metric value is used to characterize the number of collisions between signals under the current mapping scheme; For the nth round of mapping scheme The corresponding dynamic weighting coefficient, the dynamic weighting coefficient It monotonically increases with the increase of the iteration number n.

[0107] Specifically, the objective function construction unit is one of the core computing components of the device. It is responsible for reading physical parameters, system configuration parameters, and historical mapping scheme data from the storage unit, and calculating the interference metric value under the current mapping scheme according to the definition of the optimization objective function. Conflict metrics and dynamic weighting coefficients And construct a complete optimization objective function. The objective function building unit needs to perform the above calculations in each iteration, where the dynamic weight coefficients... The implementation of the objective function building unit is automatically adjusted with each iteration round n. The specific implementation of the objective function building unit is not limited; for example, it can be a functional module in a general-purpose processor that executes computation instructions, or a hardware accelerator specifically designed to optimize objective function computation.

[0108] Local perturbation unit 64 is used for the nth round mapping scheme Perform a local perturbation to generate candidate mapping schemes for the (n+1)th round. .

[0109] Specifically, the local perturbation unit receives the current mapping scheme from the objective function construction unit or the iterative control unit. Then, the mapping scheme is slightly modified, changing only the time slot position and / or wavelength channel of one or a few signals, without reconstructing the entire mapping scheme. Specific methods of local perturbation include, but are not limited to: randomly selecting one signal from all signals and randomly shifting its time slot position within a preset range; randomly selecting one signal from all signals and reallocating its wavelength channel to other available wavelength channels; randomly selecting two signals from all signals and swapping their time slot positions. Each time the local perturbation unit executes, it randomly selects one of the above methods to generate a candidate mapping scheme that has a local correlation with the current mapping scheme. Since the candidate solution differs from the current solution only locally, subsequent incremental update calculations can leverage this locality to reduce computational overhead.

[0110] The comparison and determination unit 65 is used to calculate the candidate mapping scheme in the (n+1)th round. Value under the optimization objective function and the candidate mapping scheme The optimized objective function value and the mapping scheme M in the nth round n Optimization objective function value Compare the results to obtain the mapping scheme for the (n+1)th round. .

[0111] Specifically, the comparison and determination unit receives the candidate mapping schemes generated by the local perturbation unit. Obtain the current optimization objective function from the objective function construction unit, and calculate the optimization objective function value of the candidate solution. Compare it with the current mapping scheme M n The optimization objective function value F n (M n Compare them. If This indicates that the candidate solution is superior to the current solution in terms of the overall cost of interference and conflict. The comparison and judgment unit will select the candidate solution. As the mapping scheme for the (n+1)th round ;like This indicates that the overall cost of the candidate solution is not better than the current solution, but to avoid the search process getting stuck in a local optimum, the comparison and decision unit accepts the candidate solution with a set probability P. If not accepted, the current plan will continue. As The comparison and determination unit determines... Then, it is output to the iterative control unit.

[0112] The iterative control unit 66 is used to process the mapping scheme of the (n+1)th round. Updated to the current nth round mapping scheme Let n = n + 1, and determine the mapping scheme for the current nth round. Does it meet the preset convergence condition? If it does, output the current nth round mapping scheme. As the final mapping scheme to the output unit; if the convergence condition is not met, the objective function construction unit, the local perturbation unit, and the comparison and determination unit are triggered to continue the next round of iteration.

[0113] Specifically, the iterative control unit is the core component in the device that controls the iterative process. The iterative control unit receives the mapping scheme for the (n+1)th round from the comparison and decision unit. After that, first of all Update to the current mapping scheme (Variable mapping), then increment the iteration round n to n+1, and then determine the current mapping scheme. Does it meet the preset convergence conditions? The convergence conditions include: Condition 1, conflict metric value. Condition 2, Interference metric The value is less than the preset threshold ε; Condition 3: The number of iterations n reaches the preset maximum number of iterations. When either condition one or condition two is met, the iterative control unit determines that the convergence is effective and updates the current mapping scheme. The final mapping scheme is output to the output unit; if condition three is satisfied but conditions one and two are not satisfied, it is determined to be a timeout termination, and the optimal mapping scheme recorded during the iteration process is output. The process continues to the output unit. When none of the three conditions are met, the iterative control unit triggers the objective function construction unit, the local disturbance unit, and the comparison and judgment unit to continue the next round of iteration.

[0114] The output unit 67 is used to output the mapping scheme finally determined by the iterative control unit. The mapping scheme includes the time slot position and wavelength channel allocated to each signal.

[0115] Specifically, after receiving the final mapping scheme determined by the iterative control unit, the output unit outputs the mapping scheme to downstream devices or the upper control plane in a specified format. The mapping scheme includes the time slot position and wavelength channel allocated to each signal; that is, for each signal i in N signals, the mapping scheme gives its time slot position τ. i and wavelength channel λ i The specific value of the output unit is not limited. For example, it can be a communication interface with the optical switching node controller, a data reporting port with the network management system, or a control signal output port for driving the local optical switch array. The mapping scheme output by the output unit can be directly used to configure the time slot switching matrix and wavelength selection switch in the optical switching node, so that each optical pulse signal is transmitted according to the time slot position and wavelength channel determined by the mapping scheme.

[0116] From the perspective of data flow between units, parameter acquisition unit 61 transmits the acquired physical parameters and system configuration parameters to mapping initialization unit 62, and mapping initialization unit 62 generates an initial mapping scheme. The results are then sequentially passed to the objective function construction unit 63, the local perturbation unit 64, the comparison and judgment unit 65, and the iteration control unit 66. The iteration control unit determines whether to output the final mapping scheme to the output unit 67 or trigger a new round of iteration based on the convergence judgment result. During the iterative optimization process, historical data is saved and reused between units through a storage unit (not shown in the figure). When performing calculations, the objective function construction unit 63, the comparison and judgment unit 65, and the iteration control unit 66 read the previously saved calculation values ​​from the storage unit. They only recalculate the signal pairs involved in the signal whose parameters have changed, while the calculation values ​​of other signal pairs are directly read and reused from the storage unit. This reduces the computational complexity of each iteration from O(N²) for global recalculation to O(N) for only involving the modified signals.

[0117] Through the collaborative work of the seven components described above, this embodiment implements the conflict avoidance method described in Embodiment 1 at the hardware level. The parameter acquisition unit ensures complete input of physical parameters and system configuration parameters; the mapping initialization unit provides a search starting point for iterative optimization; the objective function construction unit unifies interference and conflict metrics under the same optimization objective function; the local perturbation unit continuously generates candidate schemes through small adjustments; the comparison and judgment unit determines the merits and rejection of schemes based on the objective function value; the iterative control unit manages the iteration process and performs convergence judgment; and the output unit outputs the final mapping scheme to downstream devices. The data flow between each unit is clear and unambiguous, enabling the device to achieve collaborative optimization of conflict avoidance and signal quality maintenance with low computational overhead in large-scale optical pulse signal mapping scenarios.

[0118] Example 3 This embodiment provides a time-frequency resource mapping device, such as... Figure 7 As shown, the device includes a memory and a processor. The memory is used to store computer programs, and the processor is used to execute the computer programs stored in the memory to implement the steps of the conflict avoidance method in the time-frequency resource mapping of optical pulse signals described in any of the embodiments.

[0119] Specifically, the processor is the core component of the device responsible for computation and control functions. It reads program instructions from the storage device and executes them one by one to complete the various processing operations in the iterative optimization process described in Example 1. The processor can take various specific implementation forms. For example, the processor can be a general-purpose processor, such as a central processing unit based on a reduced instruction set architecture or a complex instruction set architecture. This type of processor implements the various steps of the conflict avoidance method by executing program instructions stored in the storage device, offering advantages such as flexible implementation and short development cycles. The processor can also be a digital signal processor (DSP). This type of processor is optimized at the hardware level for signal processing tasks, exhibiting high execution efficiency in multiplication-accumulation operations and data transfer, making it suitable for processing requirements involving a large number of overlap integral calculations and interference power accumulation operations in optical pulse signal mapping scenarios. The processor can also be an application-specific integrated circuit (ASIC), i.e., an integrated circuit customized for the specific algorithm flow of the conflict avoidance method. This type of processor has significant advantages in power consumption and processing speed, making it suitable for deployment scenarios in optical communication networks where mapping scheduling latency has strict requirements. The processor can also be a field-programmable gate array (FPGA). Such processors can implement processing steps in conflict avoidance methods, such as objective function construction, local perturbation, incremental updates, and convergence determination, by configuring logic resources. This achieves near-dedicated hardware processing performance while maintaining a certain degree of flexibility. It should be understood that the processor types described above are merely illustrative examples. In other embodiments, the processor can also adopt a combination of the above types, such as a heterogeneous architecture where a general-purpose processor and a digital signal processor work together. The general-purpose processor is responsible for iterative process control and convergence determination, while the digital signal processor is responsible for batch calculation of overlap integrals and interference power, as long as it can execute the steps of any of the methods described in Embodiment 1.

[0120] The storage device is used to store one or more programs containing computer-readable instructions that implement the methods described in any of Embodiment 1. The storage device can take various specific implementation forms. For example, the storage device can be a read-only memory (ROM), which retains the stored program content even after the device is powered off, making it suitable for storing basic programs that need to be loaded when the device starts. The storage device can also be a random access memory (RAM), which has high read and write speeds, making it suitable for storing intermediate calculation results and temporary variables during program execution. The storage device can also be an external storage medium such as a magnetic disk or optical disk, which has the advantages of large capacity and low cost, making it suitable for storing complete program images and configuration data. In one implementation, the storage device can simultaneously contain a combination of the above-mentioned storage media, for example, storing the bootloader in a read-only memory, storing temporary data during program execution in a random access memory, and storing the complete conflict avoidance method program image on a magnetic disk. When the program stored in the storage device is executed by one or more processors, the processor performs iterative optimization according to the flow of steps S1 to S5 described in Embodiment 1, completing the entire processing from the generation of the initial mapping scheme to the output of the final mapping scheme. It should be noted that this embodiment does not repeat the specific details of the method steps, but only limits them by reference. The detailed description of the method steps is subject to the description in Embodiment 1.

[0121] This embodiment also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of a time-frequency resource mapping conflict avoidance method described in any of the embodiments.

[0122] Specifically, a computer-readable storage medium refers to a physical carrier capable of carrying computer program instructions in a non-transient manner. A processor can read and execute the computer program from this medium to implement all steps of a conflict avoidance method. Computer-readable storage media can take various specific forms. For example, a computer-readable storage medium can be a read-only memory (ROM), including masked ROM, programmable ROM, erasable programmable ROM, and electrically erasable programmable ROM, etc. These media are suitable for embedding the conflict avoidance method program before leaving the factory or before deployment. A computer-readable storage medium can also be a random access memory (RAM), including static random access memory (SRAM) and dynamic random access memory (DRAM), etc. These media are suitable for temporarily loading program code for high-speed processor access during device operation. A computer-readable storage medium can also be a magnetic disk, including hard disks and floppy disks, etc. These media have the advantages of large capacity and repeated read / write capabilities, making them suitable for long-term storage of conflict avoidance method programs on time-frequency resource mapping devices in optical communication networks. A computer-readable storage medium can also be an optical disk, including read-only optical disks, recordable optical disks, and rewritable optical disks, etc. These media are suitable as carriers for program distribution and backup. It should be understood that the above-mentioned storage medium types are merely exemplary examples. In other embodiments, the computer-readable storage medium may also be other physical carriers capable of carrying computer program instructions, such as solid-state drives, flash memory cards, universal serial bus flash drives, etc., as long as the processor can read and execute the computer program from the medium.

[0123] From the perspective of the relationship between the device and the medium, the storage device in the time-frequency resource mapping device can itself contain the aforementioned computer-readable storage medium. That is, the storage device in the device carries the computer program of the conflict avoidance method in the form of a computer-readable storage medium, and the processor reads and executes the program from the storage device. Alternatively, the computer-readable storage medium can exist independently of the device, for example, carrying the computer program of the conflict avoidance method in the form of an optical disc or flash drive. When needed, the program is installed or loaded into the storage device of the optical pulse signal time-frequency resource mapping device and executed by the device's processor. Both carrying methods fall within the protection scope of this embodiment.

[0124] Through the aforementioned device and computer-readable storage medium, the conflict avoidance method described in Embodiment 1 can be loaded and executed in a practical optical communication network time-frequency resource mapping device, making the method feasible and deployable at the hardware level. The diverse selection of processor types allows the device to balance flexibility, processing speed, and power consumption according to different deployment scenarios, while the diverse selection of storage medium types allows the distribution, installation, and storage methods of the program to adapt to different engineering needs.

[0125] For ease of description, the above systems or devices are described separately as various modules or units based on their functions. Of course, in implementing this application, the functions of each unit can be implemented in one or more software and / or hardware components.

[0126] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of this application.

[0127] Finally, it should be noted that in this document, relational terms such as first, second, third, and fourth are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0128] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A time-frequency resource mapping conflict avoidance method, characterized in that, Includes the following steps: Step S1: Obtain the physical parameters and system configuration parameters of the multi-channel optical pulse signals, and generate an initial mapping scheme based on the physical parameters. The mapping scheme includes the time slot position and wavelength channel allocated to each signal, and sets the initial iteration rounds. ; Step S2: Construct the first Wheel mapping scheme The optimization objective function is as follows: ,in: For the first Wheel mapping scheme The interference metric value is used to characterize the degree of mutual interference between signals in the time and frequency domain under the current mapping scheme; For the first Wheel mapping scheme The collision metric value is used to characterize the number of collisions between signals under the current mapping scheme; For the first Wheel mapping scheme The corresponding dynamic weighting coefficient, the dynamic weighting coefficient With each iteration round The increase is monotonically increasing; Step S3: For the first Wheel mapping scheme Perform a local perturbation to generate the first... Round candidate mapping scheme ; Step S4: Calculate the candidate mapping scheme for the (n+1)th round. Values ​​under the stated objective function: The candidate mapping scheme The optimization objective function value and the first Wheel mapping scheme Optimization objective function value Compare the results to obtain the mapping scheme for the (n+1)th round. ; Step S5: Map the (n+1)th round's mapping scheme Updated to the current nth round mapping scheme Let n = n + 1, and determine the mapping scheme for the current nth round. Does it meet the preset convergence condition? If it does, output the current nth round mapping scheme. If the convergence condition is not met, return to step S2 and use the current mapping scheme as the final mapping scheme. Continue the next iteration from this starting point.

2. The method according to claim 1, characterized in that, The mapping scheme for obtaining the (n+1)th round ,include: like , indicating the candidate mapping scheme It is superior to the current [number] in terms of the overall cost of interference and conflict. Wheel mapping scheme The candidate mapping scheme As the mapping scheme for the (n+1)th round , that is to say ; like , indicating the candidate mapping scheme In terms of the combined costs of interference and conflict, it is not superior to the current [number] [unit / system / etc.]. Wheel mapping scheme However, to avoid the search process getting stuck in local optima, a probability is set. Accept the candidate mapping scheme As the mapping scheme for the (n+1)th round , that is to say If not accepted, then the current version will continue. Wheel mapping scheme As the mapping scheme for the (n+1)th round , that is to say .

3. The method according to claim 1, characterized in that, The conflict metric The method for determining is as follows: for the first Wheel mapping scheme Any two signals below and The time-domain overlap integral and spectral overlap integral are calculated. The time-domain overlap integral, the spectral overlap integral, and the peak power of the two signals are multiplied together to obtain the time-frequency crossover energy. This time-frequency crossover energy is compared to a preset threshold; if it exceeds the preset threshold, a collision is determined between the two signals. This determination is performed on all signal pairs, and the determined collision pairs are accumulated to obtain the first collision. Wheel mapping scheme Conflict metrics .

4. The method according to claim 3, characterized in that, The formula for calculating the time-domain overlap integral is as follows: ,in and The first Road and the The time-domain envelope shape function of the road signal, and The first Road and the The time slot position allocated to the path signal under the current mapping scheme; the formula for calculating the spectral overlap integral is: ,in and The first Road and the The shape function of the spectral envelope of the path signal. and The first Road and the The wavelength channel allocated to the signal under the current mapping scheme.

5. The method according to claim 1, characterized in that, The interference metric The method for determining is as follows: for the first Wheel mapping scheme Any target signal below The peak power of each of the remaining signals is multiplied by the time-domain overlap integral, spectral overlap integral, and polarization matching factor between that signal and the target signal, respectively, and then summed to obtain the target signal. Total interference power received : ; The target signal is calculated based on the total interference power. Signal-to-interference-plus-noise ratio at the receiving end : ,in For the first The effective signal power of the road signal, For noise power spectral density, The receiver's equivalent noise bandwidth is given; the actual signal-to-interference-plus-noise ratio (SIR) of each signal is weighted and summed relative to the ideal SIR under interference-free conditions to obtain the ... Wheel mapping scheme Interference metric value : ,in For the first Priority weights for road signals.

6. The method according to claim 1, characterized in that, The dynamic weighting coefficient With each iteration round The increase of exhibits superlinear growth, and its expression is: ,in These are the initial weighting coefficients. .

7. The method according to claim 6, characterized in that, The initial weighting coefficient The value is the average of the ideal signal-to-interference-plus-noise ratio (SINR) of each signal channel: ,in For the first The effective signal power of the road signal, For noise power spectral density, This is the receiver's equivalent noise bandwidth.

8. The method according to claim 1, characterized in that, The local perturbation method in step S3 can be any one of the following three methods: Method 1: Randomly select one signal from all signals and randomly offset its time slot position within a preset range; Method 2: Randomly select one signal from all signals and reallocate its wavelength channel to other available wavelength channels; Method 3: Randomly select two signals from all signals and swap their time slot positions.

9. The method according to claim 1, characterized in that, In step S4, candidate mapping schemes are calculated. When optimizing the objective function value, an incremental update method is used: only the candidate mapping schemes are updated. For signal pairs involving signals whose parameters have changed, recalculate their time-domain overlap integral, spectral overlap integral, time-frequency cross-energy, and conflict determination results, as well as recalculate the interference power associated with the signal whose parameters have changed; for other signal pairs, directly reuse the time-domain overlap integral, spectral overlap integral, time-frequency cross-energy, and conflict determination results, as well as the interference power of the remaining signals. Wheel mapping scheme The corresponding value that has already been calculated and stored.

10. The method according to claim 1, characterized in that, The set probability in step S4 The calculation method is as follows: ,in For the first The temperature parameters corresponding to the wheel, the temperature parameters With each iteration round It increases and then decreases monotonically.

11. The method according to claim 1, characterized in that, The convergence conditions in step S5 include the following three conditions: Condition 1: Current mapping scheme Conflict metrics Condition 2: Current mapping scheme Interference metric value Less than the preset threshold Condition 3: Current iteration round Reaching the preset maximum number of iterations When either condition one or condition two is met, the convergence is deemed valid, and the current mapping scheme is output. If condition three is met but conditions one and two are not, the process is considered a timeout and terminated, and the optimal mapping scheme recorded during the current iteration is output. .

12. A time-frequency resource mapping conflict avoidance device, characterized in that, include: The parameter acquisition unit is used to acquire the physical parameters and system configuration parameters of the multi-channel optical pulse signals; The mapping initialization unit is used to generate an initial mapping scheme based on the physical parameters. The mapping scheme includes the time slot position and wavelength channel allocated to each signal, and sets the initial iteration round. ; Objective function building unit, used to construct the nth round mapping scheme. The optimization objective function is as follows: in, For the nth round of mapping scheme The interference metric value is used to characterize the degree of mutual interference between signals in the time and frequency domain under the current mapping scheme; For the nth round of mapping scheme The collision metric value is used to characterize the number of collisions between signals under the current mapping scheme; For the nth round of mapping scheme The corresponding dynamic weighting coefficient, the dynamic weighting coefficient It monotonically increases with the increase of the iteration number n; Local perturbation units are used to modify the mapping scheme in the nth round. Perform a local perturbation to generate candidate mapping schemes for the (n+1)th round. ; The comparison and determination unit is used to calculate the candidate mapping scheme in the (n+1)th round. Value under the optimization objective function and the candidate mapping scheme The optimized objective function value and the mapping scheme in the nth round Optimization objective function value Compare the results to obtain the mapping scheme for the (n+1)th round. ; The iterative control unit is used to process the mapping scheme of the (n+1)th round. Updated to the current nth round mapping scheme Let n = n + 1, and determine the mapping scheme for the current nth round. Does it meet the preset convergence condition? If it does, output the current nth round mapping scheme. As the final mapping scheme; if the convergence condition is not met, the objective function construction unit, the local perturbation unit, and the comparison and determination unit are triggered to use the current mapping scheme. Continue the next iteration from this starting point.

13. A time-frequency resource mapping device, characterized in that, include: A memory and a processor, wherein the memory is used to store a computer program and the processor is used to execute the computer program stored in the memory to implement the steps of the time-frequency resource mapping conflict avoidance method as described in any one of claims 1 to 11.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the time-frequency resource mapping conflict avoidance method as described in any one of claims 1 to 11.

Images