Inversion method of atmospheric ozone vertical profile based on improved particle swarm optimization algorithm
By introducing Hammersley's low-differential sequence and centroid opposition learning strategy, T-distribution design mobile weights, adaptive learning rate and mutation strategy in the inversion method of ozone vertical profile in the atmosphere, the inadequate initialization of the inversion population and improper learning rate design are solved. The problem improves the accuracy and robustness of inversion, avoids the local optimal trap, and achieves a more efficient inversion process.
Patent Information
- Application Number
- CN202411303354.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-19
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-09-19
AI Technical Summary
The existing inversion method of ozone vertical profile in the atmosphere has improper initialization of the inversion population, insufficient position diversity of individuals, too random initial solution quality, insufficient stability of movement weights, resulting in poor inversion effect, and improper learning rate design leads to inversion convergence. Slow speed, easy to fall into local optimal.
Hammersley's low-difference sequence and centroid opposition learning strategy are introduced to improve the initial position diversity of populations, the T-distribution design is adopted to design mobile weights, and the adaptive T-distribution correction learning rate and mutation strategy are introduced, and neighbor nodes are used to perturbate to improve the algorithm's global search ability.
It improves the accuracy and robustness of the inversion process, avoids falling into local optimal solutions, balances the algorithm's convergence speed and search accuracy, and enhances the inversion effect.
Smart Images

Figure CN118821628B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of profile inversion, and in particular to an atmospheric ozone vertical profile inversion method based on an improved particle swarm algorithm. Background Art
[0002] The inversion method of atmospheric ozone vertical profile is a technical method to infer the vertical distribution of ozone in the atmosphere by observing the spectral characteristics of ozone. In the inversion process, the metaheuristic optimization algorithm is used to optimize the inversion model to more accurately infer the vertical distribution of ozone in the atmosphere. However, the general inversion method of atmospheric ozone vertical profile has the problem of improper initialization of the inversion population, resulting in insufficient diversity of individual positions, excessive randomness of the initial solution quality, and insufficient stability of the moving weights, which in turn leads to poor inversion effect; the general inversion method of atmospheric ozone vertical profile has the problem of improper learning rate design, resulting in slow inversion convergence speed, poor population movement diversity, and easy to fall into the local optimum. Summary of the invention
[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides an atmospheric ozone vertical profile inversion method based on an improved particle swarm algorithm. In view of the problem that the general atmospheric ozone vertical profile inversion method has improper initialization of the inversion population, resulting in insufficient diversity of individual positions, excessive randomness of the initial solution quality, insufficient stability of the moving weights, and thus poor inversion effect, this scheme introduces Hammersley low-discrepancy sequence and centroid opposition learning strategy to improve the diversity of the initial position of the population; introduces T-distribution to design moving weights to improve the accuracy and robustness of the algorithm in the optimization process; in view of the problem that the general atmospheric ozone vertical profile inversion method has improper learning rate design, resulting in slow inversion convergence speed, poor population movement diversity, and easy to fall into local optimality, this scheme introduces adaptive T distribution to correct the learning rate, which is helpful to balance the convergence speed and search accuracy of the algorithm, introduces a mutation strategy, and uses neighbor nodes for perturbation, which is helpful to introduce diversity, avoid falling into local optimal solutions, improve the global search ability of the algorithm, and thus improve the accuracy of the inversion process.
[0004] The technical solution adopted by the present invention is as follows: The present invention provides an atmospheric ozone vertical profile inversion method based on an improved particle swarm algorithm, the method comprising the following steps:
[0005] Step S1: data collection;
[0006] Step S2: data cleaning;
[0007] Step S3: inversion modeling;
[0008] Step S4: Inversion based on improved particle swarm.
[0009] Furthermore, in step S1, the data collection is to collect atmospheric ozone vertical profile inversion data; the atmospheric ozone vertical profile inversion data includes incident radiation intensity data, absorption coefficient data, actual observed radiation intensity data and ozone vertical distribution data.
[0010] Furthermore, in step S2, the data cleaning is to process the invalid data and noise collected; and format the data to ensure that the data conforms to the inversion modeling.
[0011] Furthermore, in step S3, the inversion modeling specifically includes the following steps:
[0012] Step S31: The atmospheric radiation transfer equation describes the absorption and scattering effects of photons in the atmosphere during propagation. The specific form is:
[0013] ;
[0014] Where I(τ,μ) is the radiation intensity at optical thickness τ with direction cosine μ; I0(τ0,μ) is the incident radiation intensity at initial optical thickness τ0; k(·) is the absorption coefficient of ozone; S(·) is the source function; and are the first and second derivatives of τ respectively;
[0015] Step S32: Assuming that the source function is zero, that is, there is no internal source, the simplified atmospheric radiation transfer equation is:
[0016] ;
[0017] Step S33: The vertical distribution of ozone is the target function obtained by inversion, which is expressed as the change of ozone concentration with height. The formula used is as follows:
[0018] ;
[0019] In the formula, is the ozone concentration at the ground level; H is the scale height of ozone; z t is the height of the tropopause; n t is the ozone concentration at the tropopause; n(·) is the ozone concentration at the target altitude; z is the altitude;
[0020] Step S34: Calculate the optical thickness, which is the integral result of the vertical distribution of ozone. The formula used is as follows:
[0021] ;
[0022] In the formula, is the vertical height; is the wavelength.
[0023] Further, in step S4, the inversion based on the improved particle swarm specifically includes the following steps:
[0024] Step S41: Pre-setting, based on the ozone concentration at the ground , ozone scale height H, tropopause height z t and the ozone concentration at the top of the troposphere, n t Establish the search space; calculate the inversion value, the formula used is as follows:
[0025] ;
[0026] Where Y modeli,i1 is the inversion value of the ith particle for the ith data point; τ i and μ i are the optical thickness and direction cosines based on the position of the ith particle, respectively; I(·) is the radiation intensity;
[0027] Step S42: define a fitness value function, which is used to measure the error between the inversion value and the actual observation value. The formula used is as follows:
[0028] ;
[0029] Where f(·) is the fitness value of the particle position; x i is the position of the ith particle; Y obs,i1 is the actual observed value, i.e., the actual radiation intensity; n is the number of data points;
[0030] Step S43: Population initialization, introducing the Hammersley low-discrepancy sequence and centroid adversarial learning strategy to initialize the population, the formula used is as follows:
[0031] ;
[0032] ;
[0033] ;
[0034] ;
[0035] In the formula, is the i-th point in the Van der Corput sequence generated using base b, and is a vector of length M; a l (i) is the lth digit after conversion to base b; N is the total length of the sequence; and are the i-th point in the Van der Corput sequence generated using bases b1 and bn-1 respectively; Hi is the i-th point of the Hammersley low-discrepancy sequence; lb and ub are the lower and upper bounds of the search space, respectively; , and are the 1st, i-th and N-th quasi-individual positions respectively;
[0036] Step S44: Design the moving weight and introduce T-distribution to enhance the search capability. The formula used is as follows:
[0037] ;
[0038] In the formula, and are the moving weights at the tth iteration and the t+1th iteration respectively, and q is the degree of freedom parameter; is the gamma function;
[0039] Step S45: Design the learning rate and introduce the adaptive T distribution to modify the learning rate to balance the convergence speed and accuracy. The formula used is as follows:
[0040] ;
[0041] In the formula, and are the learning rate and initial learning rate at the t+1th iteration, respectively; T(·) is the T distribution; d is the scaling factor; C p is the critical value;
[0042] Step S46: Design a movement strategy. The higher the individual fitness value, the worse the position. The formula used is as follows:
[0043] ;
[0044] ;
[0045] In the formula, v i (t+1) and v i (t) are the particle velocities at the t+1th iteration and the tth iteration respectively; x i (t+1) and x i (t) are the particle positions at the t+1th iteration and the tth iteration respectively; p i is the best historical position of the ith particle; rand is a random number between 0 and 1; best is the best individual position of the population; is the maximum movement weight; x r is the random individual position;
[0046] Step S47: Design a mutation strategy to perform position mutation on the 10% individuals with the highest fitness values in the population, and disturb the positions of the mutated individuals based on neighbor nodes. The formula used is as follows:
[0047] ;
[0048] In the formula, is the position after mutation, α is the mutation weight; r1 and r2 are both random numbers in the range of 0 to 1 and are independent of each other; x i+1 (t+1) and x i-1 (t+1) are the i+1th individual and the i-1th individual at the t+1th iteration;
[0049] Step S48: Inversion judgment. A fitness threshold is set in advance. When there is an individual fitness value lower than the fitness threshold, the inversion of the vertical profile of atmospheric ozone is completed, and the parameters represented by the particle position are the inversion modeling setting parameters; if the maximum number of iterations is reached, the population position is reinitialized; otherwise, the position is updated and the iterative search continues.
[0050] The beneficial effects achieved by the present invention using the above scheme are as follows:
[0051] (1) In view of the problems in the general atmospheric ozone vertical profile inversion method, such as improper initialization of the inversion population, insufficient diversity of individual positions, excessive randomness of the initial solution quality, and insufficient stability of the moving weights, which in turn lead to poor inversion effect, this scheme introduces the Hammersley low-discrepancy sequence and centroid adversarial learning strategy to improve the diversity of the initial positions of the population; introduces the T-distribution to design the moving weights to improve the accuracy and robustness of the algorithm during the optimization process.
[0052] (2) In view of the problems that the general atmospheric ozone vertical profile inversion method has improper learning rate design, resulting in slow inversion convergence speed, poor population movement diversity, and easy to fall into the local optimal solution, this scheme introduces an adaptive T distribution to correct the learning rate, which helps to balance the convergence speed and search accuracy of the algorithm. It also introduces a mutation strategy and uses neighbor nodes for perturbation, which helps to introduce diversity, avoid falling into the local optimal solution, and improve the global search ability of the algorithm, thereby improving the accuracy of the inversion process. BRIEF DESCRIPTION OF THE DRAWINGS
[0053] Figure 1 A schematic diagram of the process of the atmospheric ozone vertical profile inversion method based on the improved particle swarm algorithm provided by the present invention;
[0054] Figure 2 It is a schematic diagram of the process of step S4.
[0055] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION
[0056] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0057] In the description of the present invention, it should be understood that terms such as “upper”, “lower”, “front”, “back”, “left”, “right”, “top”, “bottom”, “inside” and “outside” indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as limiting the present invention.
[0058] Example 1, see Figure 1 The present invention provides an atmospheric ozone vertical profile inversion method based on an improved particle swarm algorithm, the method comprising the following steps:
[0059] Step S1: data collection, collecting atmospheric ozone vertical profile inversion data;
[0060] Step S2: Data cleaning, removing invalid data and noise from the collected data; and formatting the data;
[0061] Step S3: Inversion modeling, digitizing atmospheric radiation transmission, ozone concentration distribution and optical thickness, which are used for subsequent inversion of the vertical distribution of ozone in the atmosphere;
[0062] Step S4: Based on the inversion of the improved particle swarm, the population initialization introduces the Hammersley low-discrepancy sequence and the centroid opposition learning strategy; introduces the T-distribution to design the moving weight; introduces the adaptive T-distribution to correct the learning rate; introduces the mutation strategy and uses the neighbor nodes for perturbation; and finally realizes the inversion of the vertical profile of atmospheric ozone.
[0063] Example 2, see Figure 1 This embodiment is based on the above embodiment. In step S1, the atmospheric ozone vertical profile inversion data includes incident radiation intensity data, absorption coefficient data, actual observed radiation intensity data and ozone vertical distribution data.
[0064] Example 3, see Figure 1This embodiment is based on the above embodiment. In step S3, the inversion modeling specifically includes the following steps:
[0065] Step S31: The atmospheric radiation transfer equation describes the absorption and scattering effects of photons in the atmosphere during propagation. The specific form is:
[0066] ;
[0067] Where I(τ,μ) is the radiation intensity at optical thickness τ with direction cosine μ; I0(τ0,μ) is the incident radiation intensity at initial optical thickness τ0; k(·) is the absorption coefficient of ozone; S(·) is the source function; and are the first and second derivatives of τ respectively;
[0068] Step S32: Assuming that the source function is zero, that is, there is no internal source, the simplified atmospheric radiation transfer equation is:
[0069] ;
[0070] Step S33: The vertical distribution of ozone is the target function obtained by inversion, which is expressed as the change of ozone concentration with height. The formula used is as follows:
[0071] ;
[0072] In the formula, is the ozone concentration at the ground level; H is the scale height of ozone; z t is the height of the tropopause; n t is the ozone concentration at the tropopause; n(·) is the ozone concentration at the target altitude; z is the altitude;
[0073] Step S34: Calculate the optical thickness, which is the integral result of the vertical distribution of ozone. The formula used is as follows:
[0074] ;
[0075] In the formula, is the vertical height; is the wavelength.
[0076] Example 4, see Figure 1 and Figure 2 This embodiment is based on the above embodiment. In step S4, the inversion based on the improved particle swarm specifically includes the following steps:
[0077] Step S41: Pre-setting, based on the ozone concentration at the ground , ozone scale height H, tropopause height z t and the ozone concentration at the top of the troposphere, n tEstablish the search space; calculate the inversion value, the formula used is as follows:
[0078] ;
[0079] Where Y modeli,i1 is the inversion value of the ith particle for the ith data point; τ i and μ i are the optical thickness and direction cosines based on the position of the ith particle, respectively; I(·) is the radiation intensity;
[0080] Step S42: define a fitness value function, which is used to measure the error between the inversion value and the actual observation value. The formula used is as follows:
[0081] ;
[0082] Where f(·) is the fitness value of the particle position; x i is the position of the ith particle; Y obs,i1 is the actual observed value, i.e., the actual radiation intensity; n is the number of data points;
[0083] Step S43: Population initialization, introducing the Hammersley low-discrepancy sequence and centroid adversarial learning strategy to initialize the population, the formula used is as follows:
[0084] ;
[0085] ;
[0086] ;
[0087] ;
[0088] In the formula, is the i-th point in the Van der Corput sequence generated using base b, and is a vector of length M; a l (i) is the lth digit after conversion to base b; N is the total length of the sequence; and are the i-th point in the Van der Corput sequence generated using bases b1 and bn-1 respectively; H i is the i-th point of the Hammersley low-discrepancy sequence; lb and ub are the lower and upper bounds of the search space, respectively; , and are the 1st, i-th and N-th quasi-individual positions respectively;
[0089] Step S44: Design the moving weight and introduce T-distribution to enhance the search capability. The formula used is as follows:
[0090] ;
[0091] In the formula, and are the moving weights at the tth iteration and the t+1th iteration respectively, and q is the degree of freedom parameter; is the gamma function;
[0092] Step S45: designing a learning rate;
[0093] Step S46: Design a movement strategy. The higher the individual fitness value, the worse the position. The formula used is as follows:
[0094] ;
[0095] ;
[0096] In the formula, v i (t+1) and v i (t) are the particle velocities at the t+1th iteration and the tth iteration respectively; x i (t+1) and x i (t) are the particle positions at the t+1th iteration and the tth iteration respectively; p i is the best historical position of the ith particle; rand is a random number between 0 and 1; best is the best individual position of the population; is the maximum movement weight; x r is the random individual position;
[0097] Step S47: designing a mutation strategy;
[0098] Step S48: Inversion judgment. A fitness threshold is set in advance. When there is an individual fitness value lower than the fitness threshold, the inversion of the vertical profile of atmospheric ozone is completed, and the parameters represented by the particle position are the inversion modeling setting parameters; if the maximum number of iterations is reached, the population position is reinitialized; otherwise, the position is updated and the iterative search continues.
[0099] By performing the above operations, this scheme introduces the Hammersley low-discrepancy sequence and centroid opposition learning strategy to improve the diversity of the initial position of the population; introduces the T-distribution to design the moving weights to improve the accuracy and robustness of the algorithm during the optimization process.
[0100] Example 5, see Figure 2 This embodiment is based on the above embodiment. In step S4, the specific contents of designing the learning rate and designing the mutation strategy are as follows:
[0101] Step S45: Design the learning rate and introduce the adaptive T distribution to modify the learning rate to balance the convergence speed and accuracy. The formula used is as follows:
[0102] ;
[0103] In the formula, and are the learning rate and initial learning rate at the t+1th iteration, respectively; T(·) is the T distribution; d is the scaling factor; C p is the critical value;
[0104] Step S47: Design a mutation strategy to perform position mutation on the 10% individuals with the highest fitness values in the population, and disturb the positions of the mutated individuals based on neighbor nodes. The formula used is as follows:
[0105] ;
[0106] In the formula, is the position after mutation, α is the mutation weight; r1 and r2 are both random numbers in the range of 0 to 1 and are independent of each other; x i+1 (t+1) and x i-1 (t+1) are the i+1th individual and the i-1th individual at the t+1th iteration respectively.
[0107] By performing the above operations, the general atmospheric ozone vertical profile inversion method has the problem of slow inversion convergence speed, poor population movement diversity, and easy falling into local optimal solution due to improper learning rate design. This scheme introduces adaptive T distribution to correct the learning rate, which helps to balance the convergence speed and search accuracy of the algorithm. The introduction of mutation strategy and the use of neighbor nodes for perturbation help to introduce diversity, avoid falling into local optimal solution, improve the global search ability of the algorithm, and thus improve the accuracy of the inversion process.
[0108] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.
[0109] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.
[0110] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.
Claims
1. An inversion method for atmospheric ozone vertical profile based on improved particle swarm algorithm. Features: The method comprises the following steps: Step S1: data collection, collecting atmospheric ozone vertical profile inversion data; Step S2: data cleaning; Step S3: inversion modeling; Step S4: Based on the inversion of the improved particle swarm, the population initialization introduces the Hammersley low-discrepancy sequence and the centroid opposition learning strategy; introduces the T-distribution to design the moving weight; introduces the adaptive T-distribution to correct the learning rate; introduces the mutation strategy and uses the neighbor nodes for perturbation; and finally realizes the inversion of the vertical profile of atmospheric ozone; Step S4 includes step S45: designing a learning rate, introducing an adaptive T distribution to modify the learning rate, and balancing the convergence speed and accuracy; the formula used is as follows: ; In the formula, and are the learning rate and initial learning rate at the t+1th iteration, respectively; T(·) is the T distribution; d is the scaling factor; C p is the critical value; q is the degree of freedom parameter; In step S4, the inversion based on the improved particle swarm specifically includes the following steps: Step S41: Pre-setting, based on the ozone concentration at the ground , ozone scale height H, tropopause height z t and the ozone concentration at the top of the troposphere, n t Establish the search space; calculate the inversion value, the formula used is as follows: ; Where Y modeli,i1 is the inversion value of the ith particle for the ith data point; τ i and μ i are the optical thickness and direction cosines based on the position of the ith particle, respectively; I(·) is the radiation intensity; Step S42: define a fitness value function, which is used to measure the error between the inversion value and the actual observation value. The formula used is as follows: ; Where f(·) is the fitness value of the particle position; x i is the position of the ith particle; Y obs,i1 is the actual observed value, i.e., the actual radiation intensity; n is the number of data points; Step S43: Population initialization, introducing the Hammersley low-discrepancy sequence and centroid adversarial learning strategy to initialize the population, the formula used is as follows: ; ; ; ; In the formula, is the i-th point in the Van der Corput sequence generated using base b, and is a vector of length M; a l (i) is the lth digit after conversion to base b; N is the total length of the sequence; and are the i-th point in the Van der Corput sequence generated using bases b1 and bn-1 respectively; H i is the i-th point of the Hammersley low-discrepancy sequence; lb and ub are the lower and upper bounds of the search space, respectively; , and are the 1st, i-th and N-th quasi-individual positions respectively; Step S44: Design the moving weight and introduce T-distribution to enhance the search capability. The formula used is as follows: ; In the formula, and are the moving weights at the tth iteration and the t+1th iteration respectively, and q is the degree of freedom parameter; is the gamma function; Step S45: designing a learning rate; Step S46: Design a moving strategy, the higher the individual fitness value, the worse the position; Step S47: Design a mutation strategy to perform position mutation on the 10% individuals with the highest fitness values in the population, and disturb the positions of the mutated individuals based on neighboring nodes; Step S48: Inversion judgment. A fitness threshold is set in advance. When there is an individual fitness value lower than the fitness threshold, the inversion of the vertical profile of atmospheric ozone is completed, and the parameters represented by the particle position are the inversion modeling setting parameters; if the maximum number of iterations is reached, the population position is reinitialized; otherwise, the position is updated and the iterative search continues.
2. According to claim 1, the atmospheric ozone vertical profile inversion method based on improved particle swarm algorithm, Features: In step S4, the formula used to design the mobile strategy is as follows: ; ; In the formula, v i (t+1) and v i (t) are the particle velocities at the t+1th iteration and the tth iteration respectively; x i (t+1) and x i (t) are the particle positions at the t+1th iteration and the tth iteration respectively; p i is the best historical position of the ith particle; rand is a random number between 0 and 1; best is the best individual position of the population; is the maximum movement weight; x r is a random individual position.
3. The atmospheric ozone vertical profile inversion method based on improved particle swarm algorithm according to claim 1, Features: In step S4, the formula used for designing the mutation strategy is as follows: ; In the formula, is the position after mutation, α is the mutation weight; r 1 and r 2 are all random numbers in the range of 0 to 1 and are independent of each other; x i+1 (t+1) and x i-1 (t+1) are the i+1th individual and the i-1th individual at the t+1th iteration respectively.
4. The atmospheric ozone vertical profile inversion method based on improved particle swarm algorithm according to claim 1, Features: In step S3, the inversion modeling specifically includes the following steps: Step S31: The atmospheric radiation transfer equation describes the absorption and scattering effects of photons in the atmosphere during propagation. The specific form is: ; Where I(τ,μ) is the radiation intensity at optical thickness τ with direction cosine μ; I 0 (τ 0 ,μ) is the initial optical thickness τ 0 is the incident radiation intensity at ; k(·) is the absorption coefficient of ozone; S(·) is the source function; and are the first and second derivatives of τ respectively; Step S32: Assuming that the source function is zero, that is, there is no internal source, the simplified atmospheric radiation transfer equation is: ; Step S33: The vertical distribution of ozone is the target function obtained by inversion, which is expressed as the change of ozone concentration with height. The formula used is as follows: ; In the formula, is the ozone concentration at the ground level; H is the scale height of ozone; z t is the height of the tropopause; n t is the ozone concentration at the tropopause; n(·) is the ozone concentration at the target altitude; z is the altitude; Step S34: Calculate the optical thickness, which is the integral result of the vertical distribution of ozone. The formula used is as follows: ; In the formula, is the vertical height; is the wavelength.
5. The atmospheric ozone vertical profile inversion method based on improved particle swarm algorithm according to claim 1, Features: In step S1, the atmospheric ozone vertical profile inversion data includes incident radiation intensity data, absorption coefficient data, actual observed radiation intensity data and ozone vertical distribution data.
6. The atmospheric ozone vertical profile inversion method based on improved particle swarm algorithm according to claim 1, Features: In step S2, the data cleaning is to process the invalid data and noise collected; and format the data to ensure that the data conforms to the inversion modeling.
Citation Information
Patent Citations
Generalized extremum distribution-based prestack non-Gaussian AVO inversion method
CN110568494A
Troposphere ozone profile calculation method
CN111339665A
Electromagnetic monitoring crack parameter identification method based on improved marine predator algorithm
CN115293039A
Automatic XAI (autoxai) with evolutionary NAS techniques and model discovery and refinement
US20220398460A1
