41 results about "Premature convergence" patented technology

The invention discloses an unmanned aerial vehicle three-dimensional path planning method and system based on a multi-strategy improved black-wing optimization algorithm, and the method comprises the steps: generating an initial population through a Latin hypercube sampling method, thereby improving the distribution uniformity and diversity of the population; an adaptive weight factor is introduced to realize dynamic balance of exploration and development capabilities; meanwhile, a dynamic reverse learning strategy is combined, so that the global search capability is effectively enhanced, and premature convergence is avoided; the algorithm performance is further improved by fusing a warning person position updating formula in a sparrow search algorithm. The method is used for solving the problems that a traditional path planning problem is prone to falling into local optimum, the convergence speed is low, and the path is unstable. And a shorter and safer optimal flight path can be efficiently planned.

According to the large power grid reactive power optimization method and device, the storage medium and the computer equipment provided by the invention, the advantages of the two algorithms are fully exerted through the hybrid chaos quantum particle swarm optimization algorithm and the dimension-by-dimension convex space search algorithm. According to the chaotic quantum particle swarm algorithm, the global search capability and the capability of jumping out of local optimum of a particle swarm are enhanced by utilizing the characteristics of quantum behaviors and chaotic mapping, and the problem of premature convergence of a traditional heuristic intelligent algorithm is avoided. And according to the dimension-by-dimension convex space search algorithm, fine search is carried out on each excellent particle in different dimensions, a local optimal solution is determined, and the search precision and efficiency are further improved. According to the design of the hybrid algorithm, special optimization is carried out aiming at the characteristics of a reactive power optimization problem model, such as variable property difference, constraint complexity and the like, and the technical defects of poor optimization effect and optimization efficiency of an optimization solution algorithm in the prior art are effectively overcome.

The invention discloses a tunnel smooth blasting parameter selection method, which aims at the tunnel back-break and back-break problem and comprises the following steps of: constructing a blasting parameter optimization model by determining the back-break and back-break amount minimization as a target function and taking the tunnel section size, the lithology grade and the geological condition as constraint conditions; an improved genetic algorithm which introduces dynamic constraint, hierarchical coding and multi-stage fitness evaluation is utilized, and optimized blasting parameters under the minimum back break amount are screened out through continuous selection, intersection and mutation operations. Compared with a traditional genetic algorithm, global convergence and parameter practicability are improved, and meanwhile the problems of premature convergence and invalid solutions of the traditional genetic algorithm are solved.

The invention discloses a heterogeneous multi-core processor task scheduling method, system and device and a medium, and mainly relates to the technical field of task scheduling. The method is used for solving the problems that a traditional scheduling algorithm cannot process task dependency and communication overhead of a heterogeneous multi-core processor, a heuristic algorithm is prone to falling into local optimum, and a basic sparrow search algorithm has premature convergence in task scheduling. Comprising the following steps: taking a task scheduling sequence related to a DAG task scheduling graph, a communication frequency sum corresponding to the task scheduling sequence and a cross-core communication overhead sum as input data of a sparrow search algorithm; a discoverer in the sparrow population is updated through a random fractal search mechanism, and an addresser in the sparrow population is updated through a topology adaptive mechanism; meanwhile, a minimum scheduling length calculation function is used as a target function, and constraint conditions are configured; and when a preset iteration stopping condition is reached, outputting an optimal task scheduling sequence meeting the target function and the constraint condition.

The invention provides a task processing method and device based on an adaptive stagnation detection whale optimization algorithm, and aims to overcome the defect of insufficient adaptability of a static threshold value in a dynamic optimization scene in the prior art, introduce a stagnation counter to dynamically monitor an algorithm stagnation state in the whole aspect, and when the number of continuous unimproved iterations exceeds the dynamic threshold value, the algorithm stagnation state is dynamically monitored. The spiral probability is adaptively adjusted to enhance the global exploration capability, the exploration and development process of a nonlinear convergence factor coordination algorithm is continuously adopted in the aspect of details, the search efficiency is further improved by combining a random vector and a dynamic threshold mechanism, the dynamic balance of the algorithm between local optimization and global search is continuously ensured, and under the processing mechanism, the search efficiency is further improved. According to the method, premature convergence can be effectively avoided, the optimization precision and the convergence speed are remarkably improved, the method is suitable for solving complex optimization problems, and the high-quality target search requirement in practical application is met.

The invention relates to a factory job shop scheduling control optimization method and device, equipment and a medium, and the method comprises the steps: classifying workpieces irrelevant to all MPBPO sets into a first workpiece set, classifying the workpieces corresponding to each MPBPO set into one set to determine a second workpiece set, independently generating job task sorting sub-chromosome segments, and generating a job task sorting sub-chromosome segment; corresponding operation machine numbers and operation time are respectively filled into corresponding gene positions of the operation machine sub-chromosomes and the operation time sub-chromosomes, and the operation task sorting sub-chromosomes, the operation machine sub-chromosomes and the operation time sub-chromosomes are combined to obtain initial chromosomes so as to generate initial population individuals; genetic manipulation is carried out on the initial population individuals, and population individuals with the high fitness and the number of the population individuals being the population scale are selected as a next-generation initial population; and repeatedly executing, and outputting the production scheduling scheme with the optimal earliest completion time. According to the method, algorithm premature convergence can be avoided in a generalized job-shop scheduling problem with a forced parallel batch processing procedure.

The invention belongs to the field of medical data mining, and particularly provides a feature selection method of a particle swarm algorithm based on an improved dynamic multi-swarm strategy, on the basis of a binary particle swarm, in order to avoid premature convergence of the particle swarm, the dynamic multi-swarm strategy (DMS) is put forward, a non-linearly updated inertia weight coefficient is introduced, and the feature selection method of the particle swarm algorithm based on the improved dynamic multi-swarm strategy is provided. And the speed and the position of the particles are constrained by adopting a population grouping strategy, so that the particles are more suitable for exploration in the earlier stage of iteration and are more suitable for mining in the later stage of iteration to balance global search and local optimization capabilities, prevent the particles from falling into local optimum too early and ensure the convergence capability in the later stage of iteration. In order to verify the effectiveness of the method, a BPSO algorithm and a DMS-BPSO algorithm are adopted to process a data set Breaast Cancer and an Ovarian Cancer respectively, then processing results are input into a support vector machine, a K-nearest neighbor and a naive Bayes classifier for testing, and the results show that five evaluation indexes of the method are improved to a certain extent.

The invention discloses a ship plane segmentation workshop fuzzy scheduling method, which comprises the following steps of: firstly, establishing a ship plane segmentation fuzzy scheduling model, adding uncertainty of processing time to the traditional scheduling model, and representing the processing time by using a triangular fuzzy number; secondly, improving the Harlisia eagle algorithm, introducing nonlinear decreasing escape energy and simulated annealing temperature parameters, solving the problems of premature convergence and falling into a local optimal solution of the algorithm, and improving the quality of an initial population by adopting an improved NEH algorithm and chaotic mapping; and finally verifying the validity of the algorithm through a reference function, and applying the optimization method to an engineering example. The method is helpful for improving the efficiency of a ship plane section processing workshop, and has important practical significance.

The invention relates to the technical field of alfalfa screening, discloses a high-quality and high-yield alfalfa hybrid combination screening method fusing multiple technologies, provides a multi-dimensional coding and double-elite evolution mechanism, constructs an improved non-dominated sorting genetic algorithm-II, and provides a high-quality and high-yield alfalfa hybrid combination screening method by coding a hybrid combination into a continuous real number vector. A chromosome expression mode decoupled from a problem structure is constructed, a double-elite mechanism of elite retention and offset mating is introduced, the retention and heredity ability of a high-quality solution is improved, offset crossover and jitter operation is used, premature convergence is avoided, efficient global search of a strain combination space is achieved, and a high-quality solution is obtained. Meanwhile, a multi-omics association model of genes, metabolites and phenotypes is provided, transcriptome, metabolome and molecular marker data are comprehensively utilized, a multi-dimensional character prediction model is constructed, early-stage accurate screening of filial generations is achieved through correlation analysis and marker verification, the error selection rate is reduced, and the breeding period is shortened.

The present invention discloses a UAV path planning method based on a multi-strategy improved ant colony algorithm, and relates to the technical field of UAV path planning. In view of the problems existing in the traditional ant colony algorithm when performing UAV path planning, such as slow convergence speed, easy to fall into local optimal solution, complex parameter adjustment, etc., the present invention proposes a UAV path planning method based on a multi-strategy improved ant colony algorithm, including: introducing an angle function in the state transition probability to provide directional information, thereby achieving the purpose of accelerating convergence speed and reducing corners; adopting the strategy of inward expansion of pheromone and limiting pheromone concentration in the pheromone update rule to avoid the problem of premature convergence and falling into local optimal solution of the algorithm; dynamically adjusting the pheromone volatilization factor in an adaptive manner, and through pre-experimental processing, designing the objective function to dynamically adjust the value of the corresponding parameter to obtain the optimal parameter combination, thereby improving the convergence effect of the algorithm. The present invention can improve various performances in the UAV path planning process, and is more suitable for large-scale complex environments.

The invention discloses a photovoltaic MPPT (Maximum Power Point Tracking) control algorithm, which relates to the technical field of photovoltaic control algorithms, utilizes the characteristics of simple structure, high convergence speed, high global optimization capability, wide applicability and the like of a gradient optimization algorithm, and improves the defects of premature convergence and sensitivity to control parameters of the gradient optimization algorithm. Meanwhile, the photovoltaic output energy efficiency and the stability of the photovoltaic output energy efficiency serve as the evaluation basis, an MPPT algorithm performance evaluation model is constructed through an analytic hierarchy process, photovoltaic MPPT performance is evaluated, and GTO performance is verified and improved.

The invention provides a multi-AGV scheduling and path planning joint optimization method applied to intelligent warehousing, and the method comprises the steps: building a multi-AGV scheduling and path planning joint optimization model, and enabling the optimization target of the multi-AGV scheduling and path planning joint optimization model to be the minimum maximum operation completion time; and solving the proposed joint optimization model based on a biased random key genetic algorithm (BRKGA) to obtain a population optimal solution and corresponding maximum operation completion time of all AGVs. According to the method, a conflict-free path planning algorithm is adopted as a decoding scheme of population individuals in the biased random key genetic algorithm, so that a path planning scheme corresponding to a scheduling scheme represented by the individuals is obtained, and meanwhile, a local search strategy and a restart mechanism are added in the biased random key genetic algorithm; while the solving accuracy and convergence speed of the algorithm are improved, premature convergence of the population is prevented, so that a better result is obtained.

The application discloses a hysteretic parameter identification method of a BWBN model based on an improved particle swarm optimization, and belongs to the field of hysteretic parameter identification. The method comprises the following steps: determining the equivalent yield point of a RC pier under a pseudo-static force reciprocating load according to a skeleton curve of a measured hysteretic curve; giving boundary constraints of hysteretic parameters to be identified; obtaining optimal values of the hysteretic parameters to be identified by using an improved particle swarm optimization; optimizing a global historical optimal position by using a Levy flight strategy and a pattern search algorithm in each step of iteration of the improved particle swarm optimization, taking the global historical optimal position output in the last step as the value of each hysteretic parameter to be identified, and completing the hysteretic parameter identification of the BWBN model. The application solves the problems that the existing method is prone to falling into a local optimal solution in the later period, premature convergence, the convergence speed and fitting precision depend on the selection of parameters, and evolution lacks sufficient exploration ability and is difficult to guarantee the diversity of a particle population when dealing with complex high-dimensional space optimization problems.

The invention relates to a rectification system multi-objective optimization method based on dynamic NSGA-II, and belongs to the technical field of chemical process system optimization. The method comprises: determining a decision variable and a target function according to a rectification system model; a diversity index is obtained by calculating an invalid solution proportion of each generation of population and diversity of a target and decision space; and the crossover probability, the mutation probability and the population size are dynamically adjusted, a dynamic NSGA-II algorithm is constructed, the rectification system is optimized, and an optimal solution is obtained. According to the method, dynamic parameter adjustment is achieved by introducing the invalid solution proportion and the population diversity index, the number of non-convergent solutions can be effectively reduced, the population diversity is enhanced, premature convergence is avoided, the calculation cost is reduced while the optimization efficiency is improved, and an efficient and practical solution is provided for multi-objective optimization of a complex chemical system.

The invention discloses a determination method and device for a high-entropy alloy yield strength prediction model, equipment, a medium and a product. The method comprises the steps that an original feature sample set and an initial prediction model of a high-entropy alloy are obtained; screening the original feature sample set to determine an optimal feature subset; performing global search on hyper-parameters of the initial prediction model through a whale optimization algorithm to determine optimal hyper-parameters; and inputting the optimal feature subset and the optimal hyper-parameter into the initial prediction model for training, and determining a final prediction model. Firstly, key feature selection is carried out on an original feature sample set, an optimal feature subset is determined, global and local collaborative search is carried out on model hyper-parameters through a whale optimization algorithm, premature convergence is effectively avoided by utilizing an adaptive step length and position updating strategy, then model training is carried out, and a final prediction model is obtained. The calculation resource consumption in the model training and verification process is reduced, and the convergence speed and accuracy of yield strength prediction are improved.

The application discloses a wave band selection method, belonging to the technical field of spectral analysis. It includes: based on the spectral physical continuity prior and the regression model verification set performance index, constructing the task specialization genetic algorithm, which represents the wave band subset with a binary mask vector, takes the performance index as the fitness function, and adopts the structured search operator based on the continuity prior and the traditional genetic operator to form a hybrid search mechanism; based on the population evolution state representation and the fitness improvement feedback, constructing the reinforcement learning meta-control framework for synergistically regulating the algorithm hyperparameters and the structured operator; based on the recurrent neural network, constructing the timing-enhanced reinforcement learning agent; based on the spectral physical interference mechanism and the feature variation law, generating the simulation data for the iterative training of the agent; and taking the trained agent as the meta-controller to drive the algorithm to run and complete the wave band selection. The application solves the problems of the traditional genetic algorithm, such as lack of adaptive regulation, easy premature convergence and low screening precision.

This invention discloses a meme island optimization method and related apparatus for complementary sequence sets, belonging to the field of complementary sequence design. The method includes: constructing multiple island subpopulations that evolve in parallel; initializing a global archive for recording the history of search space region visits; performing population evolution in parallel on multiple islands; evaluating each candidate complementary sequence set according to a fitness function that incorporates region guidance; exchanging elite individuals among multiple islands when a preset migration condition is met; and outputting the optimal complementary sequence set found among all islands when an iteration condition is satisfied. By introducing region guidance based on the global archive into the fitness function, the search is actively guided to unexplored regions, systematically avoiding the search process from stagnating in the same basin, avoiding getting trapped in local optima, significantly improving global exploration capability, and reducing the risk of premature convergence.

The invention relates to the technical field of power system scheduling, and relates to a power system scheduling method based on particle swarm optimization, comprising the following steps: S1, establishing a scheduling model, and obtaining a total objective function; s2, constructing a power system constraint condition; s3, constructing an original artificial intelligence algorithm, initializing a population scale, and determining an iteration tolerance N1 and a local optimal tolerance N2; s4, performing population iteration until the number of iterations reaches N1, and recording the fitness of population particles; s5, screening local optimal particles, performing population iteration by taking the local optimal particles as parent particles, and querying population particle fitness in N2 iteration processes; s6, effective local optimal particles are determined, invalid local optimal particles are removed, and a population updating result is obtained; and S7, performing iteration by taking a population updating result as a parent particle, and outputting an optimal scheduling scheme. According to the method, premature convergence is avoided, the global search capability of the algorithm is remarkably enhanced, and the problem of easy local optimum in complex scheduling is solved.

The present invention discloses a method for identifying parameters of a piezoelectric actuator hysteresis model based on an MDE algorithm. The parameter identification method introduces an adaptive mechanism into the scaling factor in the mutation operation and the crossover probability in the crossover operation, so that the numerical values ​​of the scaling factor and the crossover probability change with the number of iterations, which helps the algorithm to jump out of the local optimal solution and improves the convergence speed of parameter identification and the quality of the solution. At the same time, the selection operation of the present invention copies the best individuals that appear in the evolutionary process to the next generation of original population for iteration, and gives individuals in the original population and the test population the opportunity to enter the next generation of original population, which makes up for the defect of premature convergence of the traditional DE algorithm and maintains the diversity of the population. In addition, the present invention can accurately identify the parameters in a complex hysteresis model, effectively solves the problem of low parameter prediction accuracy of the piezoelectric actuator hysteresis model, and can improve the output stability of the model parameters.

This invention discloses a generative adversarial network architecture search method and system based on a GA-PSO hybrid algorithm, belonging to the field of deep neural network technology. This method constructs a generative adversarial network supernet architecture, uses a genetic algorithm to globally explore the architecture population, combines particle swarm optimization to adjust local parameters of the elite architecture, dynamically balances inertia weights and learning factors, and ensures architecture compliance through discretized coding constraints. Finally, based on Pareto optimality, it integrates multi-objective evaluation results and cyclically optimizes the solution set using a primary and secondary collaborative ratio. Compared with traditional architecture search, this method solves the problem of premature convergence through a hybrid optimization mechanism, combines weight-sharing evaluation with layered coding to achieve global and local collaboration, significantly shortening the search cycle while improving generation quality.

This invention discloses a container scheduling method and system based on entropy weighting and multi-strategy particle swarm optimization. The method includes: constructing a multi-objective container scheduling model with the objective function of minimizing the data exchange rate between different servers and the average load imbalance of servers; acquiring server set, container set, and historical scheduling instance data, and constructing an objective priority matrix based on the multi-objective container scheduling model to determine the weights of each objective function; transforming the multi-objective container scheduling model into a single objective using a linear weighting method, and iteratively solving the single-objective container scheduling model using the multi-strategy particle swarm optimization algorithm to obtain the final optimal container scheduling solution. This solves the objective balance problem and the problems of premature convergence and insufficient population diversity in the basic particle swarm optimization algorithm, resulting in a more balanced server load and lower network communication load, achieving the optimal balance effect.

The invention discloses an NSGA-II (Non-dominated Sorting Genetic Algorithm-II) algorithm and a forest coverage type classification method. The algorithm comprises the steps of population initialization, non-dominated sorting, fitness evaluation, tournament selection, crossover and mutation operation, elitist retention and termination condition judgment. According to the invention, based on the improved NSGA-II algorithm, simplified congestion distance calculation is adopted, so that the required convergence speed is improved by 30-40%, the calculation time is reduced, and meanwhile, through cooperation of the adaptive crossover probability and the mutation probability, dynamic adjustment can be carried out according to the population evolution state, global exploration and local development are effectively balanced, premature convergence is avoided, and the convergence efficiency is improved. And the precision of outputting the optimal solution is improved.

The invention belongs to the technical field of artificial intelligence, and provides an arc model parameter optimization method based on an improved Asian wolf optimization algorithm, and the method comprises the steps: collecting an arc current signal based on a current sensor, and extracting the spectrum capability waveform of the arc current signal; establishing a direct-current arc noise model according to the arc current signal spectrum capability waveform; constructing an objective function based on the root-mean-square error, and determining a to-be-optimized parameter; and executing the improved Asian wolf optimization algorithm, and optimizing the to-be-optimized parameters. According to the method, the Asian wolf optimization algorithm is introduced through the chaos self-scaling learning mechanism, adaptive interaction of population individual position information is achieved through the chaos self-scaling learning mechanism, the optimization algorithm can be effectively prevented from being caught in premature convergence, the convergence speed is increased, and the method has high recognition precision for parameters of the direct-current arc noise model.

A computational architecture employing self-pruning fractal branch management for achieving supercomputer-class performance on standard hardware and noisy intermediate-scale quantum (NISQ) devices. Unlike conventional parallel computing systems requiring massive hardware resources or genetic algorithms requiring extensive population evolution, this invention utilizes hierarchical fractal doubles—modular computational units organized in self-similar tree structures—with real-time adaptive pruning eliminating non-promising solution branches based on geometric performance metrics computed via √2-scaled fractal analysis. Controlled perturbations (branch shaking) inject stochastic exploration preventing premature convergence while pruning maintains computational efficiency. The system achieves quantum-competitive performance on classical hardware through fractal interference patterns mimicking quantum superposition, and enables NISQ quantum computers to operate effectively despite hardware noise by pruning decoherence-corrupted branches before they contaminate computation. Core innovation: geometric pruning criterion comparing branch trajectory fractal dimension against optimal threshold, triggering instant elimination of branches exhibiting non-productive exploration patterns. Applications include neural architecture search, protein folding simulation, quantum system modeling, combinatorial optimization, and multi-agent coordination—all achieving 10-100× speedup versus conventional approaches while consuming 60-80% less energy through aggressive branch elimination. Technical advantages: (1) no training dataset required (deterministic pruning), (2) hardware-agnostic (runs on CPU / GPU / QPU), (3) noise-tolerant (quantum error mitigation via pruning), (4) energy-efficient (eliminates wasted computation), (5) scalable (fractal recursion to arbitrary depth).

A general assembly project scheduling optimization method based on importance and a resource unavailable time window comprises the following steps: analyzing topological characteristics and information transfer capability of a process in a complex equipment general assembly network through a process importance evaluation method, and after the process importance is obtained, generating a global approximate optimal scheduling scheme according to a hybrid genetic-variable neighborhood search algorithm. According to the method, through an optimization module for process importance quantification, global search of a hybrid genetic algorithm and local search of an improved variable neighborhood search algorithm, heuristic information mining and scheduling optimization of a complex equipment general assembly system are integrated, so that the algorithm search efficiency is remarkably improved, and meanwhile, premature convergence to a local optimal solution is avoided; and a new thought is provided for solving a production scheduling optimization problem in a complex production environment.

The invention discloses a code development system and method with self-debugging capability, and relates to the technical field of artificial intelligence and software engineering. The system comprises a code defect acquisition module, a self-debugging model module, a verifiable evaluation module and a strategy updating module. The self-debugging model module generates a candidate code repair scheme based on a large language model; and the strategy updating module adopts an adaptive entropy guide reinforcement learning method, and dynamically adjusts a dominant function for training by calculating the variable quantity of the model strategy entropy so as to excite the model to realize intelligent balance between exploration of a new repair path and utilization of a known effective method. The corresponding method performs model training and code repair based on the system. According to the method, the problems that an existing code debugging method based on reinforcement learning is insufficient in exploration capability and a repair strategy is converged too early are solved, and the diversity and robustness of a model generation repair scheme can be improved.

The application relates to a compiler automatic tuning method based on a discrete particle swarm algorithm, which improves the particle swarm algorithm to adapt to a discrete solution space of a compiler optimization problem, and proposes a speed updating strategy combining a crowded distance and fitness, balances the moving speed of particles by evaluating the crowded distance and fitness of solution samples in a region where the current particles are located, selects the best optimization sequence, and thus improves the performance during compiler automatic tuning; solves the premature convergence problem existing in many current discrete particle swarm algorithms and the problem of insufficient global search capability when the problem dimension is high, and significantly improves the tuning effect of compiler automatic tuning.

The application discloses a heterogeneous multi-core processor task scheduling method, system, device and medium, mainly relates to the technical field of task scheduling, and is used to solve the problems that the traditional scheduling algorithm cannot process the task dependency and communication overhead of the heterogeneous multi-core processor, the heuristic algorithm is easy to fall into local optimization, and the basic sparrow search algorithm has the problem of premature convergence in task scheduling. The application comprises the following steps: taking the task scheduling sequence related to the DAG task scheduling graph, the communication frequency sum corresponding to the task scheduling sequence, and the cross-core communication overhead sum as the input data of the sparrow search algorithm; updating the discoverer in the sparrow population through a random fractal search mechanism, and updating the joiner in the sparrow population through a topology adaptive mechanism; meanwhile, taking the minimum scheduling length calculation function as the objective function, and configuring the constraint condition; when the preset stop iteration condition is reached, the optimal task scheduling sequence meeting the objective function and the constraint condition is output.

The application discloses a code development system and method with self-debugging capability, and relates to the technical field of artificial intelligence and software engineering. The system comprises a code defect acquisition module, a self-debugging model module, a verifiable evaluation module and a strategy updating module. The self-debuging model module generates a candidate code repair scheme based on a large language model. The strategy updating module adopts an adaptive entropy-guided reinforcement learning method, calculates the change amount of the model strategy entropy, dynamically adjusts the advantage function used for training, and stimulates the model to intelligently balance between exploring new repair paths and utilizing known effective methods. The corresponding method performs model training and code repair based on the system. The application solves the problems of insufficient exploration capability and premature convergence of repair strategies in the existing code debugging method based on reinforcement learning, and can improve the diversity and robustness of the repair scheme generated by the model.

The application discloses a tunnel smooth blasting parameter selection method, aiming at the problem of tunnel overbreak and underbreak, constructs a blasting parameter optimization model by taking the minimization of overbreak and underbreak as an objective function, and taking the tunnel section size, lithology grade and geological condition as constraint conditions, and uses an improved genetic algorithm with dynamic constraints, hierarchical coding and multi-stage fitness evaluation to screen out the optimized blasting parameters under the minimum overbreak and underbreak through continuous selection, crossover and mutation operations. Compared with the traditional genetic algorithm, the application improves the global convergence and parameter practicability, and avoids the problems of premature convergence and invalid solution of the traditional genetic algorithm.