Optimization method of mixed coal washing and blending structure based on genetic algorithm

Abstract

The present invention relates to a method for optimizing the structure of mixed coal washing and coal blending based on a genetic algorithm, which solves the defect that it is difficult to determine the best coal blending scheme in the process of mixed coal blending compared with the prior art. The present invention includes the following steps: constructing a multi-objective random optimization model for the structure of mixed coal washing and coal blending; setting target parameters and their value ranges for optimization of the multi-objective random optimization model for the structure of mixed coal washing and coal blending; solving the multi-objective random optimization model for the structure of mixed coal washing and coal blending; and obtaining the optimization results of the structure of mixed coal washing and coal blending. The present invention adopts a non-dominated sorting genetic algorithm with an elite strategy for multi-objective optimization. By combining non-dominated sorting, elite strategy and congestion calculation, it can effectively solve the multi-objective optimization problems in the field of coal blending structure optimization, improve the global search capability and convergence accuracy of the algorithm, and achieve the optimization of the structure of mixed coal washing and coal blending.

Description

Technical Field

[0001] The invention relates to the technical field of mixed coal blending structure optimization, and in particular to a mixed coal washing blending structure optimization method based on a genetic algorithm. Background Art

[0002] Mixed coal washing and blending technology is an important part of the coal industry. It scientifically blends coal of different qualities to achieve the goals of improving combustion efficiency, reducing costs, and reducing environmental pollution. In mixed coal washing and blending technology, how to optimize the coal blending structure to maximize economic benefits is very critical. Traditional coal blending methods often rely on empirical judgment and simple mathematical models. These methods often fail to fully consider the complex characteristics of various coal resources and the dynamic changes in market demand. Therefore, through mixed coal washing and blending technology, coal from different sources and different characteristics are mixed in a certain proportion to achieve the expected sulfur content, ash content and other indicators, while considering cost-effectiveness and environmental impact, which can effectively improve coal blending efficiency.

[0003] Genetic algorithm is an optimization algorithm that simulates the biological evolution process. It solves optimization problems by simulating biological evolution mechanisms such as natural selection, inheritance, crossover and mutation. Genetic algorithm can use its powerful search ability and global optimization characteristics to find the best coal blending ratio for coal blending optimization problem to maximize economic benefits. Compared with the single-objective coal blending optimization model, the multi-objective coal blending structure analysis method can take into account multiple objectives such as cost, environmental impact, and economic benefits at the same time, rather than simply pursuing cost minimization or quality optimization, providing decision makers with more choices and more comprehensive decision support.

[0004] Therefore, how to use genetic algorithms to optimize the structure of mixed coal washing and blending has become a technical problem that needs to be solved urgently. Summary of the invention

[0005] The purpose of the present invention is to solve the defect in the prior art that it is difficult to determine the best coal blending scheme in the mixed coal blending process, and to provide a mixed coal washing and blending structure optimization method based on genetic algorithm to solve the above problem.

[0006] In order to achieve the above object, the technical solution of the present invention is as follows:

[0007] A method for optimizing the structure of mixed coal washing and blending based on genetic algorithm comprises the following steps:

[0008] 11) Constructing a multi-objective stochastic optimization model for the structure of mixed coal washing and blending: Taking the lowest cost of mixed coal blending, the least amount of high-quality coal and the largest amount of low-quality coal as the objective function, and taking the quality of the mixed coal washed clean coal product meeting the requirements as the constraint condition, a multi-objective stochastic optimization model for the structure of mixed coal washing and blending is constructed;

[0009] 12) Setting the target parameters and their value ranges for the multi-objective stochastic optimization model of mixed coal washing and blending structure;

[0010] 13) Solution of multi-objective stochastic optimization model for mixed coal washing and blending structure: Establish a mixed coal washing and blending structure optimization analysis method based on genetic algorithm, and use non-dominated sorting genetic algorithm with elite strategy to solve the mixed coal washing and blending structure optimization model, and obtain the best coal blending scheme and various coal quality indicators of the mixed coal under this scheme;

[0011] 14) Obtaining the optimization results of the mixed coal washing and blending structure: Based on the raw coal screening and floating test data of the coal mining industry, the blending ratio, coal quality index information and coal price data of each single type of coal are obtained. In the case of three types of single coal blending, the data is input into the multi-objective stochastic optimization model of the mixed coal washing and blending structure, and the encoded population individual vector is obtained. As the optimization variable input, Indicates Type of single coal number, Indicates The original variables are optimized through multiple iterations of the non-dominated sorting genetic algorithm with elite strategy to obtain the Pareto frontier corresponding to the optimized solution set. , according to the coal blending price information corresponding to the optimized solution set and the proportion of high-quality coal and low-quality coal, the target solution of mixed coal blending is determined, and the solution closest to the target solution of mixed coal in the Pareto front is selected as the final coal blending optimization result.

[0012] The multi-objective stochastic optimization model for mixed coal washing and blending structure is constructed by the following steps:

[0013] 21) Based on the optimized mixed coal washing and blending structure to achieve the goal of maximizing economic benefits, assuming that Kind of raw coal Mix and wash, the ratio of mixed washing is , No. The cost price of coal is , then the cost price of mixed coal is as follows:

[0014] ,

[0015] The optimization goal for the lowest coal blending cost is:

[0016] ,

[0017] in, To design the price of coal, is the coal blending cost function for mixed coal;

[0018] Assume The first type of coal is high-quality coal, and its blending ratio is , No. The first type of coal is low-quality coal, and its blending ratio is , then the optimization goal of minimizing the use of high-quality coal is , the optimization goal of using the most inferior coal is ,in is the function of the proportion of high-quality coal in the mixed coal, is the proportion function of inferior coal in mixed coal;

[0019] 22) Setting coal quality indicators for mixed coal into clean coal includes: caking index , moisture , ash content , sulfur content and volatile matter ,

[0020] For mixed coal , assuming that the predicted value of the bonding index is ,but ,in are the lower and upper limits of the value range that the mixed coal caking index should meet, respectively.

[0021] Assume The proportion of coal is Its coal resources are The total resources of each type of coal are , then the proportion of each type of coal is as follows:

[0022] ;

[0023] 23) Construct a multi-objective stochastic optimization model for the structure of mixed coal washing and blending. The specific model is as follows:

[0024] Objective function:

[0025] ,

[0026] Constraints:

[0027] ,

[0028] in, is the decision variable, i.e., the blending ratio of each single coal under different coal blending schemes; is the objective function, is the blending cost function of mixed coal, is the function of the proportion of high-quality coal in the mixed coal, is the proportion function of inferior coal in mixed coal, It is The amount of coal resources of a single type of coal, is the total resources of each type of coal, They are the lower limits of the range of values ​​for the caking index, moisture, ash, sulfur and volatile matter in the quality requirements of mixed coal, They are the upper limits of the range of values ​​for the caking index, moisture, ash, sulfur and volatile matter in the quality requirements of mixed coal. , , , They are the predicted values ​​of the mixed coal quality indicators caking index, moisture, ash, sulfur and volatile matter respectively.

[0029] The method of setting the target parameters and their value ranges for optimizing the multi-objective stochastic optimization model for mixed coal washing and blending structure comprises the following steps:

[0030] 31) Setting target parameters includes coal quality indicators, coal price, and the proportion of high-quality coal and low-quality coal. Coal quality indicators include ash content, sulfur content, caking index, moisture content and volatile matter value;

[0031] 32) The linear weighted method is used to predict the coal quality indexes of mixed coal, that is,

[0032] ,

[0033] , ;

[0034] in, Indicates The proportion of each type of coal, Indicates Types of single coal quality indicators, That is the predicted value of the quality index of mixed coal. , , , They are the predicted values ​​of the mixed coal quality indicators caking index, moisture, ash, sulfur and volatile matter, Separate Coal quality indicators of each type of coal include caking index, moisture, ash, sulfur and volatile matter;

[0035] 33) Apply step 32) to predict various coal quality indices of the mixed coal in the multi-objective stochastic optimization model of the mixed coal washing and blending structure, and use the predicted mixed coal indices as coal quality index constraint conditions of the multi-objective stochastic optimization model of the mixed coal washing and blending structure.

[0036] The solution of the multi-objective stochastic optimization model of the mixed coal washing and blending structure includes the following steps:

[0037] 41) Set the code for the mixed coal washing and blending structure:

[0038] The real number coding method is used to convert the actual coal blending problem into mathematical language. Different chromosomes correspond to different coal blending schemes in the mixed coal blending process. The genes on the chromosomes are represented by the numbers of each single coal in the mixed coal blending scheme and its blending ratio.

[0039] for In the case of single coal blending, the value range of coal type code is defined as The proportion of each type of coal in the mixture is , then the population individual code is The number of each type of coal randomly selected from the single coal warehouse and the proportion of each type of coal;

[0040] 42) Design a fitness function to judge the quality of individuals. The fitness function expression is as follows:

[0041] ,

[0042] in, The value is the fitness value of the individual in the population, , is the objective function in the multi-objective stochastic optimization model of mixed coal washing and blending structure. is the decision variable, i.e., the blending ratio of each single coal under different coal blending schemes;

[0043] 43) The genetic selection operator uses the tournament selection operator to apply the fitness function designed in step 42) to calculate the individual fitness. According to the fitness value of each individual, the individual with the highest fitness is selected to enter the next generation population;

[0044] Set the crossover operator to use The crossover operator randomly pairs selected individuals with a crossover probability Perform crossover operations to generate new individuals;

[0045] The mutation operator uses a polynomial mutation operator with mutation probability Introducing mutations to maintain the genetic diversity of the population and prevent the population from falling into local optima;

[0046] 44) Initialize the first generation population , the initial population contains Individuals, Represents the population evolution generation, the initial population , and set the maximum number of evolution generations to ;

[0047] 45) According to the fast non-dominated sorting method, the initial population is sorted to obtain the sorted population ; For each individual in the initial population Setting parameters and , Indicates the dominant individual in the population The number of solution individuals, Indicates that the individual The set of solution individuals dominated by

[0048] The specific non-dominated sorting steps are as follows:

[0049] 451) All the first generation population The individuals are stored in the collection middle;

[0050] 452) Yes Each individual in the set , the set of individuals it controls is ,set up There are For each individual , which will make The individuals are stored in the collection middle;

[0051] 453) As the first-level non-dominated individual set, all individuals in the set are given the same non-dominated order. , continue to Execute steps 451) and 452) until all individual classifications are completed;

[0052] 46) Define a comparison operator based on Euclidean distance to describe the density around a population individual, that is, The individual comparison operator expressions are as follows:

[0053] ,

[0054] in, Indicates The objective function, Indicates The individual in The value of the objective function, Indicates The individual in The value of the objective function, Indicates The comparison operator value based on Euclidean distance of each individual, Represents the number of objective functions. Individuals with the same value, The larger the value, the better, in order to maintain the diversity of the population;

[0055] 47) Apply the tournament selection operator determined in step 43) Crossover operator and polynomial mutation operator on population Perform selection, crossover, and mutation operations to obtain the offspring population , ;

[0056] Merge population and Get population , whose size is , for the population All individuals in the non-dominated hierarchical sorting generate a series of non-dominated sorting sets ;

[0057] The non-dominated set after non-dominated sorting is the best, choose first Entering the new parent population ,if The number of individuals , then continue to Add non-dominated sets to , , until the non-dominated set is added back, , then The individuals in the The individual value is added to in, until The number of individuals reached , for the obtained population Perform selection, crossover, and mutation operations to generate new offspring populations ,like ,but , that is, continue to iterate and evolve;

[0058] 48) Repeat step 47) until the The generation of offspring population meets the algorithm termination requirement, that is, the population evolution generation reaches , and obtain the optimal coal blending scheme and various coal quality indicators of the mixed coal under this scheme.

[0059] The acquisition of the optimization result of the mixed coal washing and blending structure comprises the following steps:

[0060] 51) According to the raw coal screening floating and sinking experimental data of the coal mining industry, the proportion of each single type of coal, coal quality index information and coal price data are obtained. Based on the coal quality index constraints of the mixed coal washing and blending structure multi-objective stochastic optimization model obtained in step 33), in the case of three types of single coal blending, the proportion of each single type of coal, coal quality index information and coal price data are input into the mixed coal washing and blending structure multi-objective stochastic optimization model, and the encoded population individual vector is obtained. As the optimization variable input, Indicates Type of single coal number, Indicates Proportion of each type of coal;

[0061] 52) Apply the non-dominated sorting genetic algorithm with elite strategy to perform multiple iterations to optimize the original input variables and output the Pareto frontier corresponding to the optimized solution set ;

[0062] 53) According to the Pareto frontier The target solution of mixed coal blending is determined by the corresponding coal blending price information and the blending ratio of high-quality coal and low-quality coal. The solution closest to the target solution of mixed coal in the Pareto front is selected as the optimal solution output. The optimization result of the mixed coal washing blending structure is the number of the single coal in the mixed coal blending and the blending ratio of the corresponding coal type in the solution, so as to realize the optimization of the mixed coal washing blending structure.

[0063] Beneficial Effects

[0064] The mixed coal washing and blending structure optimization method based on genetic algorithm of the present invention adopts non-dominated sorting genetic algorithm with elite strategy for multi-objective optimization compared with the prior art. By combining non-dominated sorting, elite strategy and congestion calculation, it can effectively solve the multi-objective optimization problems in the field of coal blending structure optimization, improve the global search capability and convergence accuracy of the algorithm, and realize the optimization of mixed coal washing and blending structure.

[0065] The present invention first constructs a multi-objective random optimization model for the structure of mixed coal washing and blending, and seeks to achieve a balance between the lowest coal blending cost, the least amount of high-quality coal, and the maximum amount of low-quality coal under constraints in the process of mixed coal blending with frequent coal type fluctuations, so that the coal blending effect is optimal and the rationality of the coal blending scheme is improved. The multi-objective random optimization model for the structure of mixed coal washing and blending is solved by applying a non-dominated sorting genetic algorithm with an elite strategy to obtain the optimal coal blending scheme and various coal quality indicators of the mixed coal under the scheme, and a comparison operator based on Euclidean distance is introduced to measure the degree of crowding between individuals in the population, and a tournament selection operator is selected. The crossover operator and polynomial mutation operator are used to perform genetic evolution operations on the population, which largely maintains the diversity of population distribution. The Pareto solution set obtained is evenly distributed and has better diversity.

[0066] Solving the multi-objective optimization model of the mixed coal washing and blending structure belongs to a multi-objective optimization problem. Common multi-objective optimization algorithms include multi-objective particle swarm optimization, multi-objective algorithm based on decomposition, non-dominated sorting genetic algorithm, etc. Compared with other algorithms, the non-dominated sorting genetic algorithm with elite strategy involved in the present invention has better convergence, diversity and robustness in obtaining the Pareto solution set when solving the multi-objective optimization model of the mixed coal washing and blending structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 is a method sequence diagram of the present invention;

[0068] Figure 2 The flowchart of the non-dominated sorting genetic algorithm with elite strategy involved in the present invention;

[0069] Figure 3 This is a comparison chart of the Pareto optimal solution sets of the NSGA-II algorithm and the NSGA algorithm involved in the present invention. DETAILED DESCRIPTION

[0070] In order to have a further understanding and recognition of the structural features and the effects achieved by the present invention, a preferred embodiment and accompanying drawings are used for detailed description as follows:

[0071] like Figure 1 As shown, the method for optimizing the structure of mixed coal washing and blending based on genetic algorithm of the present invention comprises the following steps:

[0072] The first step is to construct a multi-objective stochastic optimization model for the structure of mixed coal washing and blending: taking the lowest cost of mixed coal blending, the least amount of high-quality coal and the largest amount of inferior coal as the objective function, and taking the quality of the mixed coal washed clean coal product meeting the requirements as the constraint condition to construct a multi-objective stochastic optimization model for the structure of mixed coal washing and blending. The present invention proposes a multi-objective stochastic optimization model for mixed coal washing and blending. Based on the fact that in the actual coal blending production process, the inventory of high-quality coal is far less than that of inferior coal, the model takes the lowest cost of mixed coal blending, the least amount of high-quality coal and the largest amount of inferior coal as the objective function, and maximizes the economic benefits of mixed coal blending with the purpose of saving high-quality coal and reducing costs. The best and most optimal blending scheme is more reasonable and has practical guiding significance for the actual production of mixed coal blending.

[0073] (1) Based on the optimized mixed coal washing and blending structure to achieve the goal of maximizing economic benefits, assuming that Kind of raw coal Mix and wash, the ratio of mixed washing is , No. The cost price of coal is , then the cost price of mixed coal is as follows:

[0074] ,

[0075] The optimization goal for the lowest coal blending cost is:

[0076] ,

[0077] in, To design the price of coal, is the coal blending cost function for mixed coal;

[0078] Assume The first type of coal is high-quality coal, and its blending ratio is , No. The first type of coal is low-quality coal, and its blending ratio is , then the optimization goal of minimizing the use of high-quality coal is , the optimization goal of using the most inferior coal is ,in is the function of the proportion of high-quality coal in the mixed coal, It is the proportion function of inferior coal in the mixed coal.

[0079] (2) Setting coal quality indicators for mixed coal into clean coal includes: caking index , moisture , ash content , sulfur content and volatile matter ,

[0080] For mixed coal , assuming that the predicted value of the bonding index is ,but ,in are the lower and upper limits of the value range that the mixed coal caking index should meet, respectively.

[0081] Assume The proportion of coal is Its coal resources are The total resources of each type of coal are , then the proportion of each type of coal is as follows:

[0082] ;

[0083] (3) Construct a multi-objective stochastic optimization model for the structure of mixed coal washing and blending. The specific model is as follows:

[0084] Objective function:

[0085] ,

[0086] Constraints:

[0087] ,

[0088] in, is the decision variable, i.e., the blending ratio of each single coal under different coal blending schemes; is the objective function, is the blending cost function of mixed coal, is the function of the proportion of high-quality coal in the mixed coal, is the proportion function of inferior coal in mixed coal, It is The amount of coal resources of a single type of coal, is the total resources of each type of coal, They are the lower limits of the range of values ​​for the caking index, moisture, ash, sulfur and volatile matter in the quality requirements of mixed coal, They are the upper limits of the range of values ​​for the caking index, moisture, ash, sulfur and volatile matter in the quality requirements of mixed coal. , , , They are the predicted values ​​of the mixed coal quality indicators caking index, moisture, ash, sulfur and volatile matter respectively.

[0089] The second step is to set the target parameters and their value ranges for the multi-objective stochastic optimization model of mixed coal washing and blending structure.

[0090] (1) The target parameters include coal quality indicators, coal price, and the proportion of high-quality coal and low-quality coal. Coal quality indicators include ash content, sulfur content, caking index, moisture content, and volatile matter.

[0091] (2) The linear weighted method is used to predict the coal quality indicators of the mixed coal, that is,

[0092] ,

[0093] in, , ;

[0094] in, Indicates The proportion of each type of coal, Indicates Types of single coal quality indicators, That is the predicted value of the quality index of mixed coal. , , , They are the predicted values ​​of the mixed coal quality indicators caking index, moisture, ash, sulfur and volatile matter, Separate The coal quality indicators of each type of coal include caking index, moisture, ash, sulfur and volatile matter.

[0095] (3) Apply step (2) to predict various coal quality indicators of the mixed coal in the multi-objective stochastic optimization model of the mixed coal washing and blending structure, and use the predicted mixed coal indicators as coal quality indicator constraints of the multi-objective stochastic optimization model of the mixed coal washing and blending structure.

[0096] The third step is to solve the multi-objective stochastic optimization model of the mixed coal washing and blending structure: establish an optimization analysis method for the mixed coal washing and blending structure based on genetic algorithm, and apply the non-dominated sorting genetic algorithm with elite strategy to solve the optimization model of the mixed coal washing and blending structure, and obtain the optimal coal blending scheme and various coal quality indicators of the mixed coal under this scheme.

[0097] Here, the use of a non-dominated sorting genetic algorithm with an elite strategy to solve the multi-objective stochastic optimization model of mixed coal washing and blending structure has significant technical advantages. The non-dominated sorting genetic algorithm with an elite strategy is suitable for different multi-objective optimization scenarios of mixed coal blending. Due to its efficient non-dominated sorting, crowding comparison operation performance and strong ability to maintain population diversity, the algorithm shows efficient global search capabilities in dealing with multi-objective optimization problems of mixed coal blending, while also avoiding falling into local optimality. A comparison operator based on Euclidean distance is introduced to measure the degree of crowding of individuals in the population. By considering the actual position of the individual in the target space, the crowding comparison operator of the individuals in the population is calculated using the idea of ​​Euclidean distance. By comparing the crowding of individuals in the population, it is ensured that individuals with greater advantages are retained in the genetic evolution process of the population, thereby ensuring the diversity of the population and avoiding premature convergence, so that the algorithm can explore a wider solution space. Use the tournament selection operator, The selection, crossover and mutation operations of the non-dominated sorting genetic algorithm with an elite strategy using the crossover operator and polynomial mutation operator can greatly improve the convergence speed of the algorithm when solving the multi-objective stochastic optimization model of mixed coal washing and blending structure. It has a more flexible mutation mechanism and the quality of the solution obtained is higher, more reasonable and effective.

[0098] Compared with traditional genetic algorithms, non-dominated sorting genetic algorithms with elite strategies show superior performance in solving the multi-objective stochastic optimization model of mixed coal washing and blending structure. Traditional genetic algorithms usually only focus on a single optimization objective, which makes it difficult to effectively handle multi-objective optimization problems, and are prone to fall into local optimal solutions. They cannot fully explore the solution space to solve the true optimal coal blending scheme and the various coal quality indicators of mixed coal under the optimal coal blending scheme. This makes the actual production guidance significance of the optimal coal blending scheme obtained by using traditional genetic algorithms to solve the multi-objective stochastic optimization model of mixed coal washing and blending structure. The non-dominated sorting genetic algorithm with elite strategy introduces the concept of non-dominated sorting, which can handle multiple objective functions at the same time. The crowding comparison operator based on Euclidean distance can maintain the diversity of the population and avoid premature convergence to the local optimal solution. At the same time, by combining non-dominated sorting and crowding comparison operators, it can largely ensure that the algorithm converges to the Pareto optimal solution set more quickly. The non-dominated sorting genetic algorithm with elite strategy has the adaptability and flexibility to solve complex problems in different scenarios under the scenarios of multi-objective functions and multi-constraint conditions in mixed coal blending. The algorithm converges faster and the optimal coal blending results are of higher quality, which effectively improves the product quality and production efficiency of mixed coal blending, and provides a more reliable technical guarantee for the efficient utilization of coal resources.

[0099] (1) Setting the code for the mixed coal washing and blending structure:

[0100] The real number coding method is used to convert the actual coal blending problem into mathematical language. Different chromosomes correspond to different coal blending schemes in the mixed coal blending process. The genes on the chromosomes are represented by the numbers of each single coal in the mixed coal blending scheme and its blending ratio.

[0101] The use of real number coding can reduce the complexity of the algorithm coding process. At the same time, real number coding makes the algorithm move more smoothly in the search space and converge to the optimal solution faster. At the same time, real number coding can more accurately represent the subtle differences in the solution, especially in optimization problems that require high precision, and can better capture the changes in the objective function.

[0102] for In the case of single coal blending, the value range of coal type code is defined as The proportion of each type of coal in the mixture is , then the population individual code is It is composed of the number of single coal types randomly selected from the single coal warehouse and the proportion of each single coal type.

[0103] (2) Design a fitness function to judge the quality of individuals. The fitness function expression is as follows:

[0104] ,

[0105] in, The value is the fitness value of the individual in the population, , is the objective function in the multi-objective stochastic optimization model of mixed coal washing and blending structure. is the decision variable, that is, the mixing ratio of each single coal under different coal blending schemes.

[0106] (3) The genetic selection operator uses the tournament selection operator and applies the fitness function designed in step (2) to calculate the individual fitness. Based on the fitness value of each individual, the individual with the highest fitness is selected to enter the next generation population.

[0107] Set the crossover operator to use The crossover operator randomly pairs selected individuals with a crossover probability Perform crossover operations to generate new individuals;

[0108] The mutation operator uses a polynomial mutation operator with mutation probability Introduce mutations to maintain the genetic diversity of the population and prevent the population from falling into local optima.

[0109] The tournament selection operator has the advantages of simple operation, strong adaptability, and maintaining population diversity. It can also largely avoid the negative impact of the algorithm falling into elitism. That is, in some selection strategies, such as the roulette selection strategy, individuals with higher fitness will be over-selected, resulting in a lack of diversity in the population. The tournament selection can better balance the selection pressure through local competition and avoid over-emphasizing the influence of elitism. The crossover operator can control the intensity of crossover, that is, it can maintain the diversity of the population by controlling the degree of similarity between the generated new offspring individuals and the parent individuals. At the same time, it can adaptively adjust the search strategy in different solution space areas to adapt to different optimization problems; the polynomial mutation operator can flexibly adjust the intensity of mutation, control the amplitude and probability of mutation, and thus has the characteristics of adapting to different optimization problems and solution spaces.

[0110] (4) Initialize the first generation population , the initial population contains Individuals, Represents the population evolution generation, the initial population , and set the maximum number of evolution generations to .

[0111] (5) According to the fast non-dominated sorting method, the initial population is sorted to obtain the sorted population ; For each individual in the initial population Setting parameters and , Indicates the dominant individual in the population The number of solution individuals, Indicates that the individual The set of solution individuals dominated by

[0112] The specific non-dominated sorting steps are as follows:

[0113] A1) All the first generation population The individuals are stored in the collection middle;

[0114] A2) Yes Each individual in the set , the set of individuals it controls is ,set up There are For each individual , which will make The individuals are stored in the collection middle;

[0115] A3) As the first-level non-dominated individual set, all individuals in the set are given the same non-dominated order. , continue to Repeat steps A1 and A2 until all individuals have been graded.

[0116] Non-dominated sorting allows the algorithm to make trade-offs between multiple objectives, thereby effectively approaching the Pareto frontier. Through non-dominated sorting, individuals in the population are divided into different dominant layers. This hierarchical structure enables the algorithm to give priority to non-dominated individuals for selection, thereby ensuring the retention and dissemination of excellent individuals.

[0117] (6) Define a comparison operator based on Euclidean distance to describe the density around individuals in a population, that is, The individual comparison operator expressions are as follows:

[0118] ,

[0119] in, Indicates The objective function, Indicates The individual in The value of the objective function, Indicates The individual in The value of the objective function, Indicates The comparison operator value based on Euclidean distance of each individual, Represents the number of objective functions. Individuals with the same value, The larger the value, the better, in order to maintain the diversity of the population.

[0120] Compared with the common crowding operator, the crowding comparison operator using Euclidean distance measures the degree of crowding around individuals in the real sample space. In the selection process, individuals located in less crowded areas are given priority, which can effectively avoid the population from being too concentrated in a specific area, promote the uniform distribution of solutions, maintain the diversity of the population, and enable the algorithm to quickly converge to the Pareto optimal solution.

[0121] (7) Apply the tournament selection operator determined in step (3), Crossover operator and polynomial mutation operator on population Perform selection, crossover, and mutation operations to obtain the offspring population , ;

[0122] Merge population and Get population , whose size is , for the population All individuals in the non-dominated hierarchical sorting generate a series of non-dominated sorting sets ;

[0123] The non-dominated set after non-dominated sorting is the best, choose first Entering the new parent population ,if The number of individuals , then continue to Add non-dominated sets to , , until the non-dominated set is added back, , then The individuals in the The individual value is added to in, until The number of individuals reached , for the obtained population Perform selection, crossover, and mutation operations to generate new offspring populations ,like ,but , that is, continue to iterate and evolve.

[0124] (8) Repeat step (7) until the The offspring population of the first generation meets the algorithm termination requirements, the population evolves, and the optimal coal blending scheme and various coal quality indicators of the mixed coal under the scheme are obtained.

[0125] The fourth step is to obtain the optimization results of the mixed coal washing and blending structure: According to the raw coal screening and floating test data of the coal mining industry, the proportion of each single coal, coal quality index information and coal price data are obtained. In the case of three types of single coal blending, the data are input into the multi-objective stochastic optimization model of the mixed coal washing and blending structure, and the encoded population individual vector is obtained. As the optimization variable input, Indicates Type of single coal number, Indicates The original variables are optimized through multiple iterations of the non-dominated sorting genetic algorithm with elite strategy to obtain the Pareto frontier corresponding to the optimized solution set. , according to the coal blending price information corresponding to the optimized solution set and the proportion of high-quality coal and low-quality coal, the target solution of mixed coal blending is determined, and the solution closest to the target solution of mixed coal in the Pareto front is selected as the final coal blending optimization result.

[0126] (1) Based on the raw coal screening and floating test data of the coal mining industry, the proportion of each single type of coal, coal quality index information and coal price data are obtained. Based on the coal quality index constraints of the mixed coal washing and blending structure multi-objective stochastic optimization model obtained in the third step, in the case of three types of single coal blending, the proportion of each single type of coal, coal quality index information and coal price data are input into the mixed coal washing and blending structure multi-objective stochastic optimization model, and the encoded population individual vector is converted into As the optimization variable input, Indicates Type of single coal number, Indicates The proportion of each type of coal.

[0127] (2) Apply the non-dominated sorting genetic algorithm with elite strategy to perform multiple iterations to optimize the original input variables and output the Pareto frontier corresponding to the optimized solution set. .

[0128] (3) According to the Pareto frontier The target solution of mixed coal blending is determined by the corresponding coal blending price information and the blending ratio of high-quality coal and low-quality coal. The solution closest to the target solution of mixed coal in the Pareto front is selected as the optimal solution output. The optimization result of the mixed coal washing blending structure is the number of the single coal in the mixed coal blending and the blending ratio of the corresponding coal type in the solution, so as to realize the optimization of the mixed coal washing blending structure.

[0129] like Figure 2As shown, it is a flow chart of the mixed coal washing and blending structure optimization method based on genetic algorithm involved in the present invention. In the mixed coal washing and blending optimization process, based on the goal of maximizing the economic benefits of the optimized mixed coal blending structure, a multi-objective stochastic optimization model of the mixed coal washing and blending structure is constructed with the minimum mixed coal blending cost, the minimum amount of high-quality coal and the maximum amount of low-quality coal as the objective function, and the mixed coal washing clean coal product quality meeting the requirements as the constraint condition, to provide a theoretical basis for the optimization of the mixed coal washing and blending structure. On this basis, the non-dominated sorting genetic algorithm with elite strategy is used to solve the multi-objective stochastic optimization model of the mixed coal washing and blending structure, and the optimal coal blending scheme and various coal quality indicators of the mixed coal under the scheme are obtained. The non-dominated sorting genetic algorithm with elite strategy first obtains the first generation population through initialization; the fitness function is applied to evaluate the individuals in the first generation population, and the fitness values ​​on each target are calculated for subsequent non-dominated sorting calculations; the individuals in the first generation population are non-dominated hierarchical sorting combined with the crowding comparison operator based on Euclidean distance to perform multiple evolutionary iterations to optimize the population. As the iteration proceeds, it is determined whether the number of generations of genetic evolution of the population has reached the maximum number of generations of genetic evolution, that is, whether the termination condition of the algorithm is met. Finally, the Pareto optimal solution set is output, and the final optimal coal blending scheme and various coal quality indicators of the mixed coal under the scheme are determined in combination with the needs of decision makers. Through this process, we can quickly and efficiently obtain the optimization results of mixed coal washing and coal blending, and the obtained optimization results of mixed coal washing and coal blending are more valuable for guiding the actual coal blending production, more conducive to the accurate allocation and effective utilization of coal resources, and can adapt to different complex mixed coal blending production scenarios, and meet the needs of various industries and fields to produce high-quality coal products with strict standards.

[0130] like Figure 3As shown, it is a comparison diagram of the Pareto optimal solution set of the non-dominated sorting genetic algorithm with elite strategy and the traditional non-dominated sorting genetic algorithm in the mixed coal washing and blending structure optimization method based on genetic algorithm involved in the present invention. It can be seen that the individual distribution in the Pareto optimal solution set output by the non-dominated sorting genetic algorithm with elite strategy is located at the lower left of the Pareto optimal solution set of the non-dominated sorting genetic algorithm, that is, the quality of the solution in the Pareto optimal solution set obtained by the non-dominated sorting genetic algorithm with elite strategy is higher. On this basis, some individuals in the Pareto optimal solution set output by the two algorithms are similar, but the distribution of the Pareto optimal solution set output by the non-dominated sorting genetic algorithm with elite strategy is smoother and more uniform, indicating that the algorithm performs better in convergence and computational efficiency, and has significantly improved the quality and diversity of the solution set. The individuals in the Pareto optimal solution set output by the non-dominated sorting genetic algorithm show a non-smooth trend, and the individual distribution shows discrete non-uniformity, which may be due to the poor convergence of the algorithm, requiring more iterations to find a better solution, especially when solving complex multi-objective optimization problems, the algorithm performance is poor. In general, the non-dominated sorting genetic algorithm with elite strategy can obtain a higher quality Pareto optimal solution set when dealing with complex multi-objective optimization problems, has better spatial search ability and convergence, and has superior algorithm performance.

[0131] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for optimizing the structure of mixed coal washing and blending based on genetic algorithm, characterized in that: The following steps are involved: 11) Constructing a multi-objective stochastic optimization model for the structure of mixed coal washing and blending: Taking the lowest cost of mixed coal blending, the least amount of high-quality coal and the largest amount of low-quality coal as the objective function, and taking the quality of the mixed coal washed clean coal product meeting the requirements as the constraint condition, a multi-objective stochastic optimization model for the structure of mixed coal washing and blending is constructed; The multi-objective stochastic optimization model for mixed coal washing and blending structure is constructed by the following steps: 111) Based on the optimized mixed coal washing and blending structure to achieve the goal of maximizing economic benefits, assuming that Kind of raw coal Mix and wash, the ratio of mixed washing is , No. The cost price of coal is , then the cost price of mixed coal is as follows: ， The optimization goal of minimizing coal blending cost is ， in, To design the price of coal, is the coal blending cost function for mixed coal; Assume The first type of coal is high-quality coal, and its blending ratio is , No. The first type of coal is low-quality coal, and its blending ratio is , then the optimization goal of minimizing the use of high-quality coal is , the optimization goal of using the most inferior coal is ,in is the function of the proportion of high-quality coal in the mixed coal, is the proportion function of inferior coal in mixed coal; 112) Setting coal quality indicators for mixed coal into clean coal includes: caking index , moisture , ash content , sulfur content and volatile matter , For mixed coal , assuming that the predicted value of the bonding index is ,but ,in are the lower and upper limits of the value range that the mixed coal caking index index must meet, Assume The proportion of coal is Its coal resources are The total resources of each type of coal are , then the proportion of each single type of coal is as follows: ； 113) Construct a multi-objective stochastic optimization model for the structure of mixed coal washing and blending. The specific model is as follows: Objective function: ， Constraints: ， in, is the decision variable, i.e., the blending ratio of each single coal under different coal blending schemes; is the objective function, is the blending cost function of mixed coal, is the function of the proportion of high-quality coal in the mixed coal, is the proportion function of inferior coal in mixed coal, It is The amount of coal resources of a single type of coal, is the total resources of each type of coal, They are the lower limits of the range of values ​​for the caking index, moisture, ash, sulfur and volatile matter in the quality requirements of mixed coal, They are the upper limits of the range of values ​​for the caking index, moisture, ash, sulfur and volatile matter in the quality requirements of mixed coal. , , , They are the predicted values ​​of the coal quality indicators of mixed coal, namely, caking index, moisture, ash, sulfur and volatile matter; 12) Setting the target parameters and their value ranges for the multi-objective stochastic optimization model of mixed coal washing and blending structure; 13) Solution of multi-objective stochastic optimization model for mixed coal washing and blending structure: Establish a mixed coal washing and blending structure optimization analysis method based on genetic algorithm, and use non-dominated sorting genetic algorithm with elite strategy to solve the mixed coal washing and blending structure optimization model, and obtain the best coal blending scheme and various coal quality indicators of the mixed coal under this scheme; 14) Obtaining the optimization results of the mixed coal washing and blending structure: Based on the raw coal screening and floating test data of the coal mining industry, the blending ratio, coal quality index information and coal price data of each single type of coal are obtained. In the case of three types of single coal blending, the data is input into the multi-objective stochastic optimization model of the mixed coal washing and blending structure, and the encoded population individual vector is obtained. As the optimization variable input, Indicates Type of single coal number, Indicates The original variables are optimized through multiple iterations of the non-dominated sorting genetic algorithm with elite strategy to obtain the Pareto frontier corresponding to the optimized solution set. , according to the coal blending price information corresponding to the optimized solution set and the proportion of high-quality coal and low-quality coal, the target solution of mixed coal blending is determined, and the solution closest to the target solution of mixed coal in the Pareto front is selected as the final coal blending optimization result.

2. The method for optimizing the structure of mixed coal washing and blending based on genetic algorithm according to claim 1, characterized in that: The method of setting the target parameters and their value ranges for optimizing the multi-objective stochastic optimization model for mixed coal washing and blending structure comprises the following steps: 21) Setting target parameters includes coal quality indicators, coal price, and the proportion of high-quality coal and low-quality coal. Coal quality indicators include ash content, sulfur content, caking index, moisture content and volatile matter value; 22) The linear weighted method is used to predict the coal quality indexes of mixed coal, that is, ， ， ； in, Indicates The proportion of each type of coal, Indicates Types of single coal quality indicators, That is the predicted value of the quality index of mixed coal. , , , They are the predicted values ​​of the mixed coal quality indicators caking index, moisture, ash, sulfur and volatile matter, Separate Coal quality indicators of each type of coal include caking index, moisture, ash, sulfur and volatile matter; 23) Apply step 22) to predict various coal quality indicators of the mixed coal in the multi-objective stochastic optimization model of the mixed coal washing and blending structure, and use the predicted mixed coal indicators as coal quality indicator constraints of the multi-objective stochastic optimization model of the mixed coal washing and blending structure.

3. The method for optimizing the structure of mixed coal washing and blending based on genetic algorithm according to claim 1, characterized in that: The solution of the multi-objective stochastic optimization model of the mixed coal washing and blending structure includes the following steps: 31) Set the code for the mixed coal washing and blending structure: The real number coding method is used to convert the actual coal blending problem into mathematical language. Different chromosomes correspond to different coal blending schemes in the mixed coal blending process. The genes on the chromosomes are represented by the numbers of each single coal in the mixed coal blending scheme and its blending ratio. for In the case of single coal blending, the value range of coal type code is defined as The proportion of each type of coal in the mixture is , then the population individual code is The number of each type of coal randomly selected from the single coal warehouse and the proportion of each type of coal; 32) Design a fitness function to judge the quality of individuals. The fitness function expression is as follows: ， in, The value is the fitness value of the individual in the population, , is the objective function in the multi-objective stochastic optimization model of mixed coal washing and blending structure. is the decision variable, i.e., the blending ratio of each single coal under different coal blending schemes; 33) The genetic selection operator uses the tournament selection operator to apply the fitness function designed in step 32) to calculate the individual fitness. According to the fitness value of each individual, the individual with the highest fitness is selected to enter the next generation population; Set the crossover operator to use The crossover operator randomly pairs selected individuals with a crossover probability Perform crossover operations to generate new individuals; The mutation operator uses a polynomial mutation operator with mutation probability Introducing mutations to maintain the genetic diversity of the population and prevent the population from falling into local optima; 34) Initialize the first generation population , the initial population contains Individuals, Represents the population evolution generation, the initial population , and set the maximum number of evolution generations to ; 35) According to the fast non-dominated sorting method, the initial population is sorted to obtain the sorted population ; For each individual in the initial population Setting parameters and , Indicates the dominant individual in the population The number of solution individuals, Indicates that the individual The set of solution individuals dominated by The specific non-dominated sorting steps are as follows: 351) All the first generation population The individuals are stored in the collection middle; 352) Yes Each individual in the set , the set of individuals it controls is ,set up There are For each individual , which will make The individuals are stored in the collection middle; 353) As the first-level non-dominated individual set, all individuals in the set are given the same non-dominated order. , continue to Execute steps 351) and 352) until all individual classifications are completed; 36) Define a comparison operator based on Euclidean distance to describe the density around a population individual, that is, The individual comparison operator expressions are as follows: ， in, Indicates The objective function, Indicates The individual in The value of the objective function, Indicates The individual in The value of the objective function, Indicates The comparison operator value based on Euclidean distance of each individual, Represents the number of objective functions. Individuals with the same value, The larger the value, the better, in order to maintain the diversity of the population; 37) Apply the tournament selection operator determined in step 33) Crossover operator and polynomial mutation operator on population Perform selection, crossover, and mutation operations to obtain the offspring population , ; Merge population and Get population , whose size is , for the population All individuals in the non-dominated hierarchical sorting generate a series of non-dominated sorting sets ; The non-dominated set after non-dominated sorting is the best, choose first Entering the new parent population ,if The number of individuals , then continue to Add non-dominated sets to , , until the non-dominated set is added back, , then The individuals in the The individual value is added to In, until The number of individuals reached , for the obtained population Perform selection, crossover, and mutation operations to generate new offspring populations ,like ,but , that is, continue to iterate and evolve; 38) Repeat step 37) until the The generation of offspring population meets the algorithm termination requirement, that is, the population evolution generation reaches , and obtain the optimal coal blending scheme and various coal quality indicators of the mixed coal under this scheme.

4. The method for optimizing the structure of mixed coal washing and blending based on genetic algorithm according to claim 1, characterized in that: The obtaining of the optimization result of the mixed coal washing and blending structure comprises the following steps: 41) According to the raw coal screening floating and sinking experimental data of the coal mining industry, the proportion of each single type of coal, coal quality index information and coal price data are obtained. Based on the coal quality index constraints of the mixed coal washing and blending structure multi-objective stochastic optimization model obtained in step 33), in the case of three types of single coal blending, the proportion of each single type of coal, coal quality index information and coal price data are input into the mixed coal washing and blending structure multi-objective stochastic optimization model, and the encoded population individual vector is obtained. As the optimization variable input, Indicates Type of single coal number, Indicates Proportion of each type of coal; 42) Apply the non-dominated sorting genetic algorithm with elite strategy to perform multiple iterations to optimize the original input variables and output the Pareto frontier corresponding to the optimized solution set ; 43) According to the Pareto frontier The target solution of mixed coal blending is determined by the corresponding coal blending price information and the blending ratio of high-quality coal and low-quality coal. The solution closest to the target solution of mixed coal in the Pareto front is selected as the optimal solution output. The optimization result of the mixed coal washing blending structure is the number of the single coal in the mixed coal blending and the blending ratio of the corresponding coal type in the solution, so as to realize the optimization of the mixed coal washing blending structure.

Images