Optimization method of packaging can material ratio based on particle swarm optimization algorithm

Abstract

The present invention discloses a method for optimizing the material ratio of packaging cans based on a particle swarm optimization algorithm, comprising the following steps: initializing a particle swarm and dynamically adjusting the particle position and speed through a fuzzy logic control system to generate an initial material ratio scheme; constructing a fitness function based on a non-Euclidean geometry method, comprehensively considering multiple indicators such as strength, toughness, corrosion resistance and cost, and calculating the fitness value of each scheme; using a biologically inspired particle swarm optimization algorithm to gradually optimize the material ratio scheme; introducing a dynamic clustering mechanism to analyze and optimize different material ratio schemes; dynamically switching between global and local searches through a hybrid search strategy to further optimize the selected scheme; and finally outputting the optimized optimal material ratio scheme, and evaluating its reliability through experimental verification and simulation. The present invention improves the optimization efficiency and performance of the material ratio of packaging cans, and is suitable for the requirements of production automation and high efficiency.

Description

Technical Field

[0001] The invention relates to the technical field of packaging engineering, and in particular to a packaging can material ratio optimization method based on a particle swarm optimization algorithm. Background Art

[0002] In the modern packaging industry, packaging cans, as an important packaging container, are widely used in many fields such as food, beverages, and chemical products. In order to meet the requirements of strength, toughness, corrosion resistance, and cost of packaging cans in different application scenarios, the optimization of material ratio becomes the key in the design and manufacturing process. However, traditional material ratio optimization methods usually rely on experience and experiments, which have the following significant defects.

[0003] Traditional methods often rely on engineers’ experience and intuition to select and adjust material ratios. This experience-based approach makes it difficult to systematically consider the multi-dimensional factors of material performance, such as strength, toughness, corrosion resistance, and cost. Since the performance indicators of different materials are mutually restrictive, in actual operations, it is difficult for engineers to optimize multiple performance indicators at the same time, resulting in the final material ratio scheme often failing to achieve a balance in various performance indicators, limiting the overall performance improvement of the packaging can.

[0004] Traditional material ratio optimization methods usually adopt a step-by-step test method. This method is not only time-consuming and laborious, but the test results are subject to the complexity of the materials and changes in process conditions, and it is difficult to ensure repeatability and stability. In addition, due to the randomness and uncontrollability of the test, traditional methods cannot comprehensively search for potential optimal solutions in a wide range of material ratio spaces, which further limits the effect of material ratio optimization. Especially in multi-objective optimization problems, traditional methods find it difficult to find the global optimal solution in a complex multi-dimensional performance indicator space, and are often prone to falling into the local optimal solution, resulting in the material performance not being fully utilized.

[0005] Existing optimization methods lack intelligent and automated means. With the rapid development of industrial automation, traditional manual optimization methods are difficult to meet the requirements of modern industry for efficiency, precision and automation. In the actual production process, material ratio optimization often requires a large amount of experimental data and historical experience, and the processing and analysis of these data rely on the manual operation of engineers, which is inefficient and prone to errors. In addition, traditional optimization methods are difficult to respond quickly to changes in the production process, lack dynamic adjustment capabilities, and are difficult to adapt to the changing working conditions and needs in industrial production.

[0006] Therefore, how to provide a packaging can material ratio optimization method based on a particle swarm optimization algorithm is a problem that those skilled in the art need to solve urgently. Summary of the invention

[0007] One purpose of the present invention is to propose a method for optimizing the material ratio of packaging cans based on a particle swarm optimization algorithm. The present invention makes full use of fuzzy logic control systems, non-Euclidean geometry methods and bio-inspired particle swarm optimization algorithms, and describes in detail the entire process from the initial generation of a material ratio scheme to the final optimization. It has the advantages of high optimization efficiency, comprehensive performance considerations, and strong algorithm adaptability.

[0008] The method for optimizing the ratio of packaging can materials based on a particle swarm optimization algorithm according to an embodiment of the present invention comprises the following steps:

[0009] S1. Initialize the particle swarm, set the number of particle swarms, the initial position and speed of each particle, each particle represents a material ratio scheme, the initial position and speed of the particle are set based on the fuzzy logic control system, and by inputting experimental data and expert knowledge, the fuzzy logic control system is used to dynamically adjust the initial position and speed of the particle to generate a set of initial material ratio schemes;

[0010] S2. Using the initial material ratio scheme, define the fitness function. The fitness function comprehensively considers multiple performance indicators of the packaging can, including strength, toughness, corrosion resistance and cost. The fitness function is constructed using a non-Euclidean geometry method. According to the multi-dimensional performance indicators of the material ratio scheme, the weights and parameters are dynamically adjusted to calculate the initial fitness value of each ratio scheme.

[0011] S3. Based on the initial fitness value of each ratio scheme, the particle swarm optimization algorithm is used to update the position and velocity of the particles, and a new material ratio scheme is iteratively generated. The bio-inspired particle swarm optimization algorithm is introduced to adjust the behavior strategy of the particle swarm by imitating the dynamic change mechanism of the biological population, and gradually optimize the ratio scheme;

[0012] S4. In each iteration, the fitness value of each new material ratio scheme is calculated and updated, the fitness value of the current ratio scheme is compared with its historical optimal fitness value, and the particle swarm is clustered according to the dynamic clustering mechanism to identify different particle swarms. By adjusting the search strategy of particles in each cluster, a more optimized ratio scheme is generated;

[0013] S5. Among all the updated material ratio schemes, the scheme with the highest fitness value is selected as the global optimal scheme. Through the hybrid search strategy, the global and local searches are dynamically switched to further optimize the selected scheme so that it approaches the global optimal solution.

[0014] S6. When the preset number of iterations or fitness value convergence condition is reached, the iteration is terminated, the optimal material ratio scheme after final optimization is output, and the reliability of the optimization result is evaluated through experimental verification and simulation;

[0015] S7. Apply the final optimized optimal material ratio scheme to the design and manufacture of packaging cans, use the optimal material ratio scheme to guide material selection and ratio, optimize production processes and quality control measures, and realize automated and efficient production.

[0016] Optionally, the S1 specifically includes:

[0017] S11, initialize the particle swarm, set the number of particle swarms to N, where each particle P i , where i = 1, 2, ..., N represents a material ratio scheme;

[0018] S12, for each particle P i Randomly generate initial position X i (0) and initial velocity V i (0), where the initial position X i (0) represents the initial parameters of the material ratio scheme, the initial speed V i (0) represents the initial change rate of material ratio parameters;

[0019] S13, the experimental data set D = {d 1 ,d 2 ,…,d m} and expert knowledge base K = {k 1 ,k 2 ,…,k n} is input into the fuzzy logic control system to generate the fuzzy rule set R = {r 1 ,r 2 ,…,r p}, the fuzzy rule set R is used to guide the initial position X of the particle i (0) and initial velocity V i (0) Adjustment;

[0020] S14, based on the fuzzy reasoning process of the fuzzy logic control system, calculate the adjusted initial position X' of the particle i (0) and initial velocity V′ i (0):

[0021] X′ i (0) = μ X (Xi(0),R,D,K);

[0022] V′ i (0) = μ V (Vi(0),R,D,K);

[0023] Among them, the function μ X and μ V They represent the dynamic adjustment process of the fuzzy logic control system on the initial position and velocity of the particle, X′i (0) and V′ i (0) is the adjusted initial position and velocity of the particle;

[0024] S15, adjust the initial position of the particle X' i (0) and speed V′ i (0) As the initial input of the optimization algorithm, a set of initial material ratio schemes {P′ 1 ,P′ 2 ,…,P′ N}.

[0025] Optionally, the S2 specifically includes:

[0026] S21, the generated initial material ratio scheme {P′ 1 ,P′ 2 ,…,P′ N} Input the fitness function into the building module, and use the fitness function to evaluate the quality of each ratio scheme;

[0027] S22, define the fitness function F(P′ i ), the fitness function comprehensively considers various performance indicators of the packaging can, including strength S(P′ i ), toughness T(P′ i ), corrosion resistance C(P′ i ) and cost Cost(P′ i ):

[0028]

[0029] Among them, w 1 ,w 2 ,w 3 ,w 4 is the weight coefficient of the corresponding performance indicator, and is dynamically adjusted according to the priority of the multi-dimensional performance indicators;

[0030] S23, using non-Euclidean geometry method to construct fitness function, through the multi-dimensional performance index {S(P′ i ),T(P′ i ),C(P′ i ),Cost(P′ i )} to perform nonlinear transformation, model the distance and angle of each performance indicator in non-Euclidean space, and the non-Euclidean distance d(P′ i ) is defined as one of the metrics in the fitness function:

[0031]

[0032] Among them, g k(P′ i ) are the performance indicators obtained through nonlinear transformation, is the ideal target value;

[0033] S24, based on non-Euclidean geometry method to adjust the weights and parameters in the fitness function, to meet the optimization requirements of different material ratio schemes, by dynamically adjusting the weight coefficient {w 1 ,w 2 ,w 3 ,w 4}, so that the fitness function reflects the quality of the material ratio scheme;

[0034] S25, calculate each initial material ratio scheme P' i The fitness value F(P′ i ), and rank the material ratio schemes based on the calculation results to determine the advantages and disadvantages of each scheme.

[0035] Optionally, the S3 specifically includes:

[0036] S31, based on the sorted material ratio scheme {Q 1 ,Q 2 ,…,Q N} and its corresponding fitness value G(Q j ), where j = 1, 2, ..., N, update each particle Q in the particle swarm j Current location Y j (t) and speed W j (t);

[0037] S32, calculate each particle Q j The current fitness value G(Y j (t)) and the particle's own best historical position The corresponding fitness value For comparison, if G(Y j (t)) is better than Then the historical best position of the updated particle is

[0038] S33, introduce the biologically inspired particle swarm optimization algorithm to adjust the behavior strategy of the particle swarm by imitating the dynamic change mechanism of the biological population:

[0039] Y j (t+1)=Y j (t)+W j (t+1);

[0040] The speed update formula is:

[0041]

[0042] Among them, ω is the inertia weight, c 1 and c 2 is the acceleration constant, r 1 and r 2 Random numbers, is the global best position;

[0043] S34. When updating the particle position and velocity, the inertia weight ω and the acceleration constant c are adjusted according to the dynamic change characteristics in the particle swarm by combining the biological inspiration mechanism. 1 ,c 2 Perform adaptive adjustments to optimize the algorithm's convergence speed and global search capabilities;

[0044] S35. After each iteration, calculate the position Y of all particles j (t+1) corresponds to the fitness value G(Y j (t+1)), and compare these fitness values ​​with the global best fitness value Compare and if the new fitness value is better, update the global best position Generate a set of local optimal solutions {L 1 ,L 2 ,…,L M}, where each local optimal solution L m , where m = 1, 2, …, M represents a potential optimal solution found in the global optimization process.

[0045] Optionally, the S4 specifically includes:

[0046] S41, based on the local optimal solution set {L 1 ,L 2 ,…,L M}, for each local optimal solution L m , where m = 1, 2, ..., M is re-evaluated to calculate the fitness value G(L m ) and compare it with the global best position The fitness value of Make comparisons;

[0047] S42. In each iteration, based on the calculated fitness value, a dynamic clustering mechanism is used to cluster the local optimal solution set {L 1 ,L 2 ,…,L M} and the global optimal position Perform cluster analysis and classify the fitness value G(L m )and The similarity of these solutions is used to classify them and generate multiple clusters C k, where k = 1, 2, ..., K, and each cluster represents a characteristic category of a material ratio scheme;

[0048] S43. In each cluster C k Select the solution with the highest fitness value As the representative solution of the cluster, the fitness improvement space of the representative solution in the current cluster is further calculated as:

[0049]

[0050] L m′ ∈C k ;

[0051] Among them, ΔG(C k ) represents cluster C k The internal representation of the optimization potential of the solution, avg(G(L m′ )) represents cluster C k The average fitness value of all solutions in ;

[0052] S44, according to the fitness of each cluster to improve the space ΔG (C k ) size, dynamically adjust the weight coefficient w of each cluster k , optimize the search strategy for cluster representative solutions:

[0053] w k =w k +α·ΔG(C k );

[0054] Among them, α is the learning rate, which controls the amplitude of weight adjustment;

[0055] S45, the adjusted weight coefficient w k Applied to subsequent optimization iterations, continue to perform local and global searches on the representative solutions of each cluster to generate a more optimized material ratio scheme, and finally use the optimized representative solutions of the cluster to update the global optimal position

[0056] Optionally, the S5 specifically includes:

[0057] S51, in each cluster C k According to the adjusted weight coefficient w k and cluster representative solution The generated optimized material ratio scheme has the fitness value of all cluster representative solutions Compare and select the solution with the highest fitness value As a candidate global optimal solution;

[0058] S52, the candidate global optimal solution With the global best position For comparison, if Then the updated global best position is

[0059] S53. In the hybrid search strategy, first calculate the fitness improvement of the current iteration To evaluate the progress of search results;

[0060] S54, based on fitness improvement Dynamically adjust the scope of the search space. In the global search stage, when the fitness improvement is large, moderately expand the search space to avoid premature convergence. In the local search stage, when the fitness improvement is small, gradually narrow the search space to explore the potential optimal solution in the current area:

[0061]

[0062] Among them, S new is the adjusted search space range, β is the scaling factor, which controls the adjustment amplitude of the search space;

[0063] S55, switching between global search and local search according to the dynamically adjusted search space result;

[0064] S56, in the process of dynamic switching of the hybrid search strategy, the parameter ω of the search strategy g ,ω l ,c 1g ,c 2g ,c 1l ,c 2l Adaptive adjustment is performed, combined with an adaptive search space scaling mechanism, to optimize these parameters in real time according to the progress of the current search phase, optimizing the convergence speed and search accuracy of the algorithm at different search phases;

[0065] S57, the optimized global optimal position It is used to dynamically evaluate the global fitness value of each material ratio scheme, further optimize the selected scheme, and generate the optimized global optimal scheme.

[0066] Optionally, the S6 specifically includes:

[0067] S61. Based on the optimized global optimal solution, all material ratio solutions X i , where i = 1, 2, ..., M is used for real-time monitoring to determine whether the preset number of iterations T is reached max Or the fitness value convergence condition ΔG<δ, where ΔG represents the change in fitness value between the current iteration and the previous iteration, and δ is the set convergence threshold;

[0068] S62, when the fitness value change ΔG is less than the convergence threshold δ or the number of iterations reaches T max When , the iteration is terminated and the current global optimal material ratio scheme X is output opt ;

[0069] S63, output the optimal material ratio solution X opt Applied in actual production experiments to collect experimental data D test And calculate the experimental fitness value G test , in order to evaluate the actual application effect of the optimal material ratio scheme;

[0070] S64. Based on the experimental verification results, use simulation evaluation tools to evaluate the optimal material ratio scheme X opt Conduct simulation analysis to simulate its performance under various usage conditions and calculate the simulation fitness value G simul , and compare it with the experimental fitness value G test Make a comparison;

[0071] S65, the simulation evaluation result G simul The experimental verification results G test Comprehensive analysis is performed. If there is a significant difference between the two, the material ratio scheme X opt Adjust and optimize, and if the results are consistent and meet expectations, finally confirm X opt It is the global optimal material ratio scheme and is used in actual production as the final design scheme for the material ratio of packaging cans.

[0072] The beneficial effects of the present invention are:

[0073] The present invention introduces a fuzzy logic control system and dynamically adjusts the initial position and speed of particles according to experimental data and expert knowledge, thereby effectively improving the initial search efficiency and accuracy of the algorithm, making the optimization process more in line with practical application needs. The fitness function is constructed using a non-Euclidean geometry method, which can comprehensively consider the multi-dimensional performance indicators of the packaging can, and achieves accurate processing of multi-objective optimization problems by dynamically adjusting weights and parameters.

[0074] The present invention introduces a biologically inspired particle swarm optimization algorithm, which enhances the algorithm's adaptability and global search capabilities by imitating the dynamic change mechanism of biological populations, effectively avoiding the problem of the algorithm falling into the local optimal solution in high-dimensional optimization problems, and further improving the effect of global optimization. The application of the dynamic clustering mechanism enables the optimization process to classify particles according to their characteristics and optimize the particle swarms in each cluster separately. This mechanism improves the convergence speed and optimization efficiency of the algorithm. By dynamically switching between global search and local search through a hybrid search strategy, the selected material ratio scheme is further optimized, ensuring the flexibility and robustness of the algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0076] Figure 1 This is a flow chart of the packaging can material ratio optimization method based on the particle swarm optimization algorithm proposed by the present invention;

[0077] Figure 2 A schematic diagram of constructing a fitness function combining a fuzzy logic control system and a non-Euclidean geometry method in the present invention;

[0078] Figure 3 A schematic diagram of the combination of the biologically inspired particle swarm optimization algorithm and the dynamic clustering mechanism in the present invention;

[0079] Figure 4 Schematic diagram of the application of the hybrid search strategy in the present invention. DETAILED DESCRIPTION

[0080] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0081] refer to Figure 1-4 , a packaging can material ratio optimization method based on particle swarm optimization algorithm includes the following steps:

[0082] S1. Initialize the particle swarm, set the number of particle swarms, the initial position and speed of each particle, each particle represents a material ratio scheme, the initial position and speed of the particle are set based on the fuzzy logic control system, and by inputting experimental data and expert knowledge, the fuzzy logic control system is used to dynamically adjust the initial position and speed of the particle to generate a set of initial material ratio schemes;

[0083] S2. Using the initial material ratio scheme, define the fitness function. The fitness function comprehensively considers multiple performance indicators of the packaging can, including strength, toughness, corrosion resistance and cost. The fitness function is constructed using a non-Euclidean geometry method. According to the multi-dimensional performance indicators of the material ratio scheme, the weights and parameters are dynamically adjusted to calculate the initial fitness value of each ratio scheme.

[0084] S3. Based on the initial fitness value of each ratio scheme, the particle swarm optimization algorithm is used to update the position and velocity of the particles, and a new material ratio scheme is iteratively generated. The bio-inspired particle swarm optimization algorithm is introduced to adjust the behavior strategy of the particle swarm by imitating the dynamic change mechanism of the biological population, and gradually optimize the ratio scheme;

[0085] S4. In each iteration, the fitness value of each new material ratio scheme is calculated and updated, the fitness value of the current ratio scheme is compared with its historical optimal fitness value, and the particle swarm is clustered according to the dynamic clustering mechanism to identify different particle swarms. By adjusting the search strategy of particles in each cluster, a more optimized ratio scheme is generated;

[0086] S5. Among all the updated material ratio schemes, the scheme with the highest fitness value is selected as the global optimal scheme. Through the hybrid search strategy, the global and local searches are dynamically switched to further optimize the selected scheme so that it approaches the global optimal solution.

[0087] S6. When the preset number of iterations or fitness value convergence condition is reached, the iteration is terminated, the optimal material ratio scheme after final optimization is output, and the reliability of the optimization result is evaluated through experimental verification and simulation;

[0088] S7. Apply the final optimized optimal material ratio scheme to the design and manufacture of packaging cans, use the optimal material ratio scheme to guide material selection and ratio, optimize production processes and quality control measures, and realize automated and efficient production.

[0089] In this implementation manner, the S1 specifically includes:

[0090] S11, initialize the particle swarm, set the number of particle swarms to N, where each particle P i , where i = 1, 2, ..., N represents a material ratio scheme;

[0091] S12, for each particle P i Randomly generate initial position X i (0) and initial velocity V i (0), where the initial position X i (0) represents the initial parameters of the material ratio scheme, the initial speed V i (0) represents the initial change rate of material ratio parameters;

[0092] S13, the experimental data set D = {d 1 ,d 2 ,…,d m} and expert knowledge base K = {k 1 ,k 2 ,…,k n} is input into the fuzzy logic control system to generate the fuzzy rule set R = {r 1 ,r 2 ,…,r p}, the fuzzy rule set R is used to guide the initial position X of the particle i (0) and initial velocity V i(0) Adjustment;

[0093] S14, based on the fuzzy reasoning process of the fuzzy logic control system, calculate the adjusted initial position X' of the particle i (0) and initial velocity V′ i (0):

[0094] X′ i (0) = μ X (Xi(0),R,D,K);

[0095] V′ i (0) = μ V (Vi(0),R,D,K);

[0096] Among them, the function μ X and μ V They represent the dynamic adjustment process of the fuzzy logic control system on the initial position and velocity of the particle, X′ i (0) and V′ i (0) is the adjusted initial position and velocity of the particle;

[0097] S15, adjust the initial position of the particle X' i (0) and speed V′ i (0) As the initial input of the optimization algorithm, a set of initial material ratio schemes {P′ 1 ,P′ 2 ,…,P′ N}.

[0098] In this implementation, S2 specifically includes:

[0099] S21, the generated initial material ratio scheme {P′ 1 ,P′ 2 ,…,P′ N} Input the fitness function into the building module, and use the fitness function to evaluate the quality of each ratio scheme;

[0100] S22, define the fitness function F(P′ i ), the fitness function comprehensively considers various performance indicators of the packaging can, including strength S(P′ i ), toughness T(P′ i ), corrosion resistance C(P′ i ) and cost Cost(P′ i ):

[0101]

[0102] Among them, w 1 ,w 2 ,w3 ,w 4 is the weight coefficient of the corresponding performance indicator, and is dynamically adjusted according to the priority of the multi-dimensional performance indicators;

[0103] S23, using non-Euclidean geometry method to construct fitness function, through the multi-dimensional performance index {S(P′ i ),T(P′ i ),C(P′ i ),Cost(P′ i )} to perform nonlinear transformation, and model the distance and angle of each performance indicator in non-Euclidean space. The non-Euclidean distance d(P′ i ) is defined as one of the metrics in the fitness function:

[0104]

[0105] Among them, g k (P′ i ) are the performance indicators obtained through nonlinear transformation, is the ideal target value;

[0106] S24, based on non-Euclidean geometry method to adjust the weights and parameters in the fitness function, to meet the optimization requirements of different material ratio schemes, by dynamically adjusting the weight coefficient {w 1 ,w 2 ,w 3 ,w 4}, so that the fitness function reflects the quality of the material ratio scheme;

[0107] S25, calculate each initial material ratio scheme P' i The fitness value F(P′ i ), and rank the material ratio schemes based on the calculation results to determine the advantages and disadvantages of each scheme.

[0108] In this implementation, S3 specifically includes:

[0109] S31, based on the sorted material ratio scheme {Q 1 ,Q 2 ,…,Q N} and its corresponding fitness value G(Q j ), where j = 1, 2, ..., N, update each particle Q in the particle swarm j Current location Y j (t) and speed W j (t);

[0110] S32, calculate each particle Q j The current fitness value G(Y j(t)) and the particle's own best historical position The corresponding fitness value For comparison, if G(Y j (t)) is better than Then the historical best position of the updated particle is

[0111] S33, introduce the biologically inspired particle swarm optimization algorithm to adjust the behavior strategy of the particle swarm by imitating the dynamic change mechanism of the biological population:

[0112] Y j (t+1)=Y j (t)+W j (t+1);

[0113] The speed update formula is:

[0114]

[0115] Among them, ω is the inertia weight, c 1 and c 2 is the acceleration constant, r 1 and r 2 Random numbers, is the global best position;

[0116] S34. When updating the particle position and velocity, the inertia weight ω and the acceleration constant c are adjusted according to the dynamic change characteristics in the particle swarm by combining the biological inspiration mechanism. 1 ,c 2 Perform adaptive adjustments to optimize the algorithm's convergence speed and global search capabilities;

[0117] S35. After each iteration, calculate the position Y of all particles j (t+1) corresponds to the fitness value G(Y j (t+1)), and compare these fitness values ​​with the global best fitness value Compare and if the new fitness value is better, update the global best position Generate a set of local optimal solutions {L 1 ,L 2 ,…,L M}, where each local optimal solution L m , where m = 1, 2, …, M represents a potential optimal solution found in the global optimization process.

[0118] In this implementation, S4 specifically includes:

[0119] S41, based on the local optimal solution set {L 1 ,L 2 ,…,LM}, for each local optimal solution L m , where m = 1, 2, ..., M is re-evaluated to calculate the fitness value G(L m ) and compare it with the global best position The fitness value of Make comparisons;

[0120] S42. In each iteration, based on the calculated fitness value, a dynamic clustering mechanism is used to cluster the local optimal solution set {L 1 ,L 2 ,…,L M} and the global optimal position Perform cluster analysis according to the fitness value G(L m )and The similarity of these solutions is used to classify them and generate multiple clusters C k , where k = 1, 2, ..., K, and each cluster represents a characteristic category of a material ratio scheme;

[0121] S43. In each cluster C k Select the solution with the highest fitness value As the representative solution of the cluster, the fitness improvement space of the representative solution in the current cluster is further calculated as:

[0122]

[0123] L m′ ∈C k ;

[0124] Among them, ΔG(C k ) represents cluster C k The internal representation of the optimization potential of the solution, avg(G(L m′ )) represents cluster C k The average fitness value of all solutions in ;

[0125] S44, according to the fitness of each cluster to improve the space ΔG (C k ) size, dynamically adjust the weight coefficient w of each cluster k , optimize the search strategy for cluster representative solutions:

[0126] w k =w k +α·ΔG(C k );

[0127] Among them, α is the learning rate, which controls the amplitude of weight adjustment;

[0128] S45, the adjusted weight coefficient w kApplied to subsequent optimization iterations, continue to perform local and global searches on the representative solutions of each cluster to generate a more optimized material ratio scheme, and finally use the optimized representative solutions of the cluster to update the global optimal position

[0129] In this implementation manner, S5 specifically includes:

[0130] S51, in each cluster C k According to the adjusted weight coefficient w k and cluster representative solution The generated optimized material ratio scheme has the fitness value of all cluster representative solutions Compare and select the solution with the highest fitness value As a candidate global optimal solution;

[0131] S52, the candidate global optimal solution With the global best position For comparison, if Then the global best position is updated to

[0132] S53. In the hybrid search strategy, first calculate the fitness improvement of the current iteration To evaluate the progress of search results;

[0133] S54, based on fitness improvement Dynamically adjust the scope of the search space. In the global search stage, when the fitness improvement is large, moderately expand the search space to avoid premature convergence. In the local search stage, when the fitness improvement is small, gradually narrow the search space to explore the potential optimal solution in the current area:

[0134]

[0135] Among them, S new is the adjusted search space range, β is the scaling factor, which controls the adjustment amplitude of the search space;

[0136] S55, switching between global search and local search according to the dynamically adjusted search space result;

[0137] S56, in the process of dynamic switching of the hybrid search strategy, the parameter ω of the search strategy g ,ω l ,c 1g ,c 2g ,c 1l ,c 2lAdaptive adjustment is performed, combined with an adaptive search space scaling mechanism, to optimize these parameters in real time according to the progress of the current search phase, optimizing the convergence speed and search accuracy of the algorithm at different search phases;

[0138] S57, the optimized global optimal position It is used to dynamically evaluate the global fitness value of each material ratio scheme, further optimize the selected scheme, and generate the optimized global optimal scheme.

[0139] In this implementation manner, S6 specifically includes:

[0140] S61. Based on the optimized global optimal solution, all material ratio solutions X i , where i = 1, 2, ..., M is used for real-time monitoring to determine whether the preset number of iterations T is reached max Or the fitness value convergence condition ΔG<δ, where ΔG represents the change in fitness value between the current iteration and the previous iteration, and δ is the set convergence threshold;

[0141] S62, when the fitness value change ΔG is less than the convergence threshold δ or the number of iterations reaches T max When , the iteration is terminated and the current global optimal material ratio scheme X is output opt ;

[0142] S63, output the optimal material ratio solution X opt Applied in actual production experiments to collect experimental data D test And calculate the experimental fitness value G test , in order to evaluate the actual application effect of the optimal material ratio scheme;

[0143] S64. Based on the experimental verification results, use simulation evaluation tools to evaluate the optimal material ratio scheme X opt Conduct simulation analysis to simulate its performance under various usage conditions and calculate the simulation fitness value G simul , and compare it with the experimental fitness value G test Make a comparison;

[0144] S65, the simulation evaluation result G simul The experimental verification results G test Comprehensive analysis is performed. If there is a significant difference between the two, the material ratio scheme X opt Adjust and optimize, and if the results are consistent and meet expectations, finally confirm X opt It is the global optimal material ratio scheme and is used in actual production as the final design scheme for the material ratio of packaging cans.

[0145] Embodiment 1:

[0146] In order to verify the feasibility of the present invention in practice, the present invention is applied in the food and beverage industry. Packaging cans are widely used in the packaging of liquid foods, such as juice, carbonated beverages and canned foods. These products have very high requirements for packaging cans, which not only need to ensure the safety and shelf life of the contents, but also need to control costs. Therefore, how to optimize the material ratio of packaging cans to achieve the best balance in multiple performance aspects such as strength, toughness, and corrosion resistance, while reducing production costs, has become an urgent problem to be solved in the industry.

[0147] To meet this challenge, we conducted a field application in a large beverage manufacturing plant. The plant produces more than 1 billion cans of beverages annually, and the material cost and quality of packaging cans are directly related to the market competitiveness of products and the profits of enterprises. The traditional material ratio method mainly relies on the experience and experiments of engineers, and often fails to take into account various performance indicators, resulting in insufficient strength, poor corrosion resistance or excessive cost of packaging cans in actual use.

[0148] In this scenario, we applied a method for optimizing the material ratio of packaging cans based on the particle swarm optimization algorithm. First, we conducted a detailed analysis of the performance requirements of the packaging cans. According to production requirements, the packaging cans need to have sufficient strength to resist internal and external pressures, good toughness to prevent damage during transportation, and excellent corrosion resistance to ensure food safety. At the same time, the material cost must be controlled to maximize economic benefits. In the optimization process, we first initialize the particle swarm using a fuzzy logic control system. By inputting experimental data and expert experience, the initial position and speed of each particle are dynamically adjusted to generate a set of initial material ratio schemes. At this stage, the system considered a variety of material combinations, including aluminum alloys, tin-plated steel sheets, and plastic composites in different proportions. These materials performed differently in the initial experiments. The system dynamically adjusted the initial parameters of each material through fuzzy reasoning to generate an initial ratio scheme that better meets actual needs.

[0149] We constructed a fitness function that uses non-Euclidean geometry to handle multi-dimensional performance indicators. This fitness function takes into account the strength, toughness, corrosion resistance, and cost of the packaging can. By comparing the experimental data of different material combinations, we found that the fitness function constructed using non-Euclidean geometry can more accurately reflect the comprehensive performance of material properties. During the optimization process, the fitness function dynamically adjusts the weights according to the importance of each performance indicator, thereby avoiding excessive bias towards a single performance indicator in multi-objective optimization.

[0150] Under the guidance of the bio-inspired particle swarm optimization algorithm, we continuously update the position and velocity of the particles and iteratively generate new material ratio schemes. In the experiment, we observed that the bio-inspired mechanism effectively enhanced the adaptability of the particle swarm and avoided the risk of the algorithm falling into the local optimal solution. Through this process, we successfully optimized a variety of material ratio schemes, which showed better overall performance in preliminary experiments. We introduced a dynamic clustering mechanism to further optimize these material ratio schemes. The dynamic clustering mechanism divides the material ratio schemes into different categories according to their characteristics, and optimizes the schemes in each category separately. In this process, the system can automatically identify material combinations with similar performance, and adjust the optimization strategy according to the clustering results, so that the final generated material ratio scheme achieves a better balance in different performance indicators.

[0151] In order to verify the actual effect of the final optimization scheme, we selected three different material ratio schemes and tested them on the actual production line of the factory. The test results show that compared with the traditional material ratio scheme, the optimized packaging cans have an increase of about 15% in strength, about 12% in toughness, and 20% in corrosion resistance, while the material cost is reduced by about 8%. These data show that the present invention not only improves the performance of the packaging cans, but also effectively reduces the production cost, bringing significant economic benefits to the enterprise. During the optimization process, we further verified the reliability of the optimization scheme through experimental data and simulation evaluation. In multiple actual production batches, the optimized packaging cans showed good consistency and stability, and almost no quality problems occurred. At the same time, through simulation evaluation, we simulated the long-term storage of packaging cans under different storage environments. The results showed that the optimized packaging cans can maintain good physical properties under various extreme conditions, further proving the practicality of the optimization method.

[0152] Table 1 Performance comparison before and after material ratio optimization

[0153]

[0154] Table 2 Experimental and simulation evaluation results of optimized material ratio scheme

[0155]

[0156]

[0157] Through the data display of the above two tables, we can clearly see the significant improvements brought about by the material ratio optimization method of the packaging can based on the particle swarm optimization algorithm of the present invention. In terms of material performance, the optimized material ratio scheme shows significant improvements in multiple key performance indicators. For example, the strength is increased by up to 16%, the toughness is increased by up to 12.5%, and the corrosion resistance is increased by up to 21%. These data show that through the method of the present invention, the physical properties of the packaging can in actual use have been significantly improved, and can better meet the production and use requirements. At the same time, the optimized material ratio scheme also shows excellent effects in reducing production costs, with the cost reduced by up to 8.2%. This result proves that the present invention can not only improve the overall performance of the packaging can, but also bring significant economic benefits in reducing production costs. In addition, the data in Table 2 further verify the practical application effect and reliability of the optimization scheme. The qualified rate of the optimized material ratio scheme in the actual production experiment is significantly improved, and at the same time, in different simulation environments, the optimized material ratio scheme shows excellent adaptability and stability. These results comprehensively demonstrate the significant advantages of the method of the present invention in improving the performance of the packaging can, reducing production costs, and ensuring product quality consistency and reliability, which fully proves the practical application value and innovation of the present invention.

[0158] The present invention proposes a method for optimizing the material ratio of packaging cans based on a particle swarm optimization algorithm. By introducing a fuzzy logic control system, a non-Euclidean geometry method and a bio-inspired particle swarm optimization algorithm, the efficiency and effect of material ratio optimization are effectively improved. In the embodiment, the present invention is actually applied in a large beverage manufacturing plant. By intelligently optimizing the ratio of packaging can materials, the strength, toughness and corrosion resistance of the packaging cans are significantly improved, while the production cost is reduced. In a specific application, the initial position and speed of the particle swarm are dynamically adjusted according to experimental data and expert experience by a fuzzy logic control system, and a set of reasonable initial material ratio schemes are generated. The fitness function constructed by the non-Euclidean geometry method can comprehensively consider the multi-dimensional performance indicators of the packaging can and realize the multi-objective optimization of the material ratio scheme. The bio-inspired particle swarm optimization algorithm further enhances the global search capability of the algorithm and avoids the problem of falling into a local optimal solution. In the entire optimization process, a dynamic clustering mechanism and a hybrid search strategy are also combined to ensure that the material ratio scheme finally generated achieves the best balance in various performance indicators. In general, the present invention provides an efficient and intelligent material ratio optimization solution for the packaging industry, which has broad application prospects and important commercial value.

[0159] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A packaging can material ratio optimization method based on particle swarm optimization algorithm, characterized in that: The steps include: S1. Initialize the particle swarm, set the number of particle swarms, the initial position and speed of each particle, each particle represents a material ratio scheme, the initial position and speed of the particle are set based on the fuzzy logic control system, and by inputting experimental data and expert knowledge, the fuzzy logic control system is used to dynamically adjust the initial position and speed of the particle to generate a set of initial material ratio schemes; S2. Using the initial material ratio scheme, define the fitness function. The fitness function comprehensively considers multiple performance indicators of the packaging can, including strength, toughness, corrosion resistance and cost. The fitness function is constructed using a non-Euclidean geometry method. According to the multi-dimensional performance indicators of the material ratio scheme, the weights and parameters are dynamically adjusted to calculate the initial fitness value of each ratio scheme. S3. Based on the initial fitness value of each ratio scheme, the particle swarm optimization algorithm is used to update the position and velocity of the particles, and a new material ratio scheme is iteratively generated. The bio-inspired particle swarm optimization algorithm is introduced to adjust the behavior strategy of the particle swarm by imitating the dynamic change mechanism of the biological population, and gradually optimize the ratio scheme; S4. In each iteration, the fitness value of each new material ratio scheme is calculated and updated, the fitness value of the current ratio scheme is compared with its historical optimal fitness value, and the particle swarm is clustered according to the dynamic clustering mechanism to identify different particle swarms. By adjusting the search strategy of particles in each cluster, a more optimized ratio scheme is generated; S5. Among all the updated material ratio schemes, the scheme with the highest fitness value is selected as the global optimal scheme. Through the hybrid search strategy, the global and local searches are dynamically switched to further optimize the selected scheme so that it approaches the global optimal solution. S6. When the preset number of iterations or fitness value convergence condition is reached, the iteration is terminated, the optimal material ratio scheme after final optimization is output, and the reliability of the optimization result is evaluated through experimental verification and simulation; S7. Apply the final optimized optimal material ratio scheme to the design and manufacture of packaging cans, use the optimal material ratio scheme to guide material selection and ratio, optimize production processes and quality control measures, and realize automated and efficient production.

2. The method for optimizing the ratio of packaging can materials based on particle swarm optimization algorithm according to claim 1, characterized in that: The S1 specifically includes: S11, initialize the particle swarm, set the number of particle swarms to N, where each particle P i , where i = 1, 2, ..., N represents a material ratio scheme; S12, for each particle P i Randomly generate initial position X i (0) and initial velocity V i (0), where the initial position X i (0) represents the initial parameters of the material ratio scheme, the initial speed V i (0) represents the initial change rate of material ratio parameters; S13, the experimental data set D = {d1, d2, ..., d m } and expert knowledge base K = {k1, k2, …, k n } is input into the fuzzy logic control system to generate the fuzzy rule set R = {r1, r2, …, r p }, the fuzzy rule set R is used to guide the initial position X of the particle i (0) and initial velocity V i (0) Adjustment; S14, based on the fuzzy reasoning process of the fuzzy logic control system, calculate the adjusted initial position X' of the particle i (0) and initial velocity V′ i (0): X′ i (0)=μ X (X i (0),R,D,K); V′ i (0)=μ V (V i (0),R,D,K); Among them, the function μ X and μ V They represent the dynamic adjustment process of the fuzzy logic control system on the initial position and velocity of the particle, X′ i (0) and V′ i (0) is the adjusted initial position and velocity of the particle; S15, adjust the initial position of the particle X' i (0) and speed V′ i (0) As the initial input of the optimization algorithm, a set of initial material ratio schemes {P′1,P′2,…,P′ N }.

3. The method for optimizing the ratio of packaging can materials based on particle swarm optimization algorithm according to claim 1, characterized in that: The S2 specifically includes: S21, the generated initial material ratio scheme {P′1, P′2,…, P′ N } Input the fitness function into the building module, and use the fitness function to evaluate the quality of each ratio scheme; S22, define the fitness function F(P′ i ), the fitness function comprehensively considers various performance indicators of the packaging can, including strength S(P′ i ), toughness T(P′ i ), corrosion resistance C(P′ i ) and cost Cost(P′ i ): Among them, w1, w2, w3, w4 are weight coefficients of corresponding performance indicators, and are dynamically adjusted according to the priority of multi-dimensional performance indicators; S23, using non-Euclidean geometry method to construct fitness function, through the multi-dimensional performance index {S(P′ i ),T(P′ i ),C(P′ i ),Cost(P′ i )} to perform nonlinear transformation, model the distance and angle of each performance indicator in non-Euclidean space, and the non-Euclidean distance d(P′ i ) is defined as one of the metrics in the fitness function: Among them, g k (P′ i ) are the performance indicators obtained through nonlinear transformation, is the ideal target value; S24, based on the non-Euclidean geometry method, the weights and parameters in the fitness function are adjusted to meet the optimization requirements of different material ratio schemes, and the fitness function is made to reflect the advantages and disadvantages of the material ratio scheme by dynamically adjusting the weight coefficients {w1, w2, w3, w4}; S25, calculate each initial material ratio scheme P' i The fitness value F(P′ i ), and rank the material ratio schemes based on the calculation results to determine the advantages and disadvantages of each scheme.

4. The method for optimizing the ratio of packaging can materials based on particle swarm optimization algorithm according to claim 1, characterized in that: The S3 specifically includes: S31, based on the sorted material ratio scheme {Q1, Q2, ..., Q N } and its corresponding fitness value G(Q j ), where j = 1, 2, ..., N, update each particle Q in the particle swarm j Current location Y j (t) and speed W j (t); S32, calculate each particle Q j The current fitness value G(Y j (t)) and the particle's own best historical position The corresponding fitness value For comparison, if G(Y j (t)) is better than Then the historical best position of the updated particle is S33, introduce the biologically inspired particle swarm optimization algorithm to adjust the behavior strategy of the particle swarm by imitating the dynamic change mechanism of the biological population: Y j (t+1)=Y j (t)+W j (t+1); The speed update formula is: Among them, ω is the inertia weight, c1 and c2 are acceleration constants, r1 and r2 are random numbers, is the global best position; S34. When updating the particle position and velocity, the inertia weight ω and acceleration constants c1, c2 are adaptively adjusted according to the dynamic change characteristics in the particle swarm in combination with the biological inspiration mechanism to optimize the convergence speed and global search capability of the algorithm; S35. After each iteration, calculate the position Y of all particles j (t+1) corresponds to the fitness value G(Y j (t+1)), and compare these fitness values ​​with the global best fitness value Compare and if the new fitness value is better, update the global best position Generate a set of local optimal solutions {L1,L2,…,L M }, where each local optimal solution L m , where m = 1, 2, …, M represents a potential optimal solution found in the global optimization process.

5. The method for optimizing the ratio of packaging can materials based on particle swarm optimization algorithm according to claim 1, characterized in that: The S4 specifically includes: S41, based on the local optimal solution set {L1,L2,…,L M }, for each local optimal solution L m , where m = 1, 2, ..., M is re-evaluated to calculate the fitness value G(L m ) and compare it with the global best position The fitness value of Make comparisons; S42. In each iteration, based on the calculated fitness value, a dynamic clustering mechanism is used to cluster the local optimal solution set {L1, L2, ..., L M } and the global optimal position Perform cluster analysis and classify the fitness value G(L m )and The similarity of these solutions is used to classify them and generate multiple clusters C k , where k = 1, 2, ..., K, and each cluster represents a characteristic category of a material ratio scheme; S43. In each cluster C k Select the solution with the highest fitness value As the representative solution of the cluster, the fitness improvement space of the representative solution in the current cluster is further calculated as: L m′ ∈C k ; Among them, ΔG(C k ) represents cluster C k The internal representation of the optimization potential of the solution, avg(G(L m′ )) represents cluster C k The average fitness value of all solutions in ; S44, according to the fitness of each cluster to improve the space ΔG (C k ) size, dynamically adjust the weight coefficient w of each cluster k , optimize the search strategy for cluster representative solutions: In k =in k +α ΔG(C k ); Among them, α is the learning rate, which controls the amplitude of weight adjustment; S45, the adjusted weight coefficient w k Applied to subsequent optimization iterations, continue to perform local and global searches on the representative solutions of each cluster to generate a more optimized material ratio scheme, and finally use the optimized representative solutions of the cluster to update the global optimal position 6. The method for optimizing the ratio of packaging can materials based on particle swarm optimization algorithm according to claim 1, characterized in that: The S5 specifically includes: S51, in each cluster C k According to the adjusted weight coefficient w k and cluster representative solution The generated optimized material ratio scheme has the fitness value of all cluster representative solutions Compare and select the solution with the highest fitness value As a candidate global optimal solution; S52, the candidate global optimal solution With the global best position For comparison, if Then the updated global best position is S53. In the hybrid search strategy, first calculate the fitness improvement of the current iteration To evaluate the progress of search results; S54, based on fitness improvement Dynamically adjust the scope of the search space. In the global search stage, when the fitness improvement is large, moderately expand the search space to avoid premature convergence. In the local search stage, when the fitness improvement is small, gradually narrow the search space to explore the potential optimal solution in the current area: Among them, S new is the adjusted search space range, β is the scaling factor, which controls the adjustment amplitude of the search space; S55, switching between global search and local search according to the dynamically adjusted search space result; S56, in the process of dynamic switching of the hybrid search strategy, the parameter ω of the search strategy g ,ω l ,c 1g ,c 2g ,c 1l ,c 2l Adaptive adjustment is performed, combined with an adaptive search space scaling mechanism, to optimize these parameters in real time according to the progress of the current search phase, optimizing the convergence speed and search accuracy of the algorithm at different search phases; S57, the optimized global optimal position It is used to dynamically evaluate the global fitness value of each material ratio scheme, further optimize the selected scheme, and generate the optimized global optimal scheme.

7. The method for optimizing the ratio of packaging can materials based on particle swarm optimization algorithm according to claim 1, characterized in that: The S6 specifically includes: S61. Based on the optimized global optimal solution, all material ratio solutions X i , where i = 1, 2, ..., M is used for real-time monitoring to determine whether the preset number of iterations T is reached max Or the fitness value convergence condition ΔG<δ, where ΔG represents the change in fitness value between the current iteration and the previous iteration, and δ is the set convergence threshold; S62, when the fitness value change ΔG is less than the convergence threshold δ or the number of iterations reaches T max When , the iteration is terminated and the current global optimal material ratio scheme X is output opt ; S63, output the optimal material ratio solution X opt Applied in actual production experiments to collect experimental data D test And calculate the experimental fitness value G test , in order to evaluate the actual application effect of the optimal material ratio scheme; S64. Based on the experimental verification results, use simulation evaluation tools to evaluate the optimal material ratio scheme X opt Conduct simulation analysis to simulate its performance under various usage conditions and calculate the simulation fitness value G simul , and compare it with the experimental fitness value G test Make a comparison; S65, the simulation evaluation result G simul The experimental verification results G test Comprehensive analysis is performed. If there is a significant difference between the two, the material ratio scheme X opt Adjust and optimize, and if the results are consistent and meet expectations, finally confirm X opt It is the global optimal material ratio scheme and is used in actual production as the final design scheme for the material ratio of packaging cans.

Images