671 results about "Crossover" patented technology

The invention discloses an intelligent path planning method and system for an unmanned aerial vehicle. The method comprises the following steps: initializing related parameters and optimization algorithm parameters of the unmanned aerial vehicle; coding the path points of the unmanned aerial vehicle; calculating fitness values of all individuals in the population; generating a new population through selection, crossover and mutation operations; carrying out local search on individuals with better fitness in the population; dynamically adjusting algorithm parameters according to population fitness changes; and outputting the path with the highest fitness value as the optimal flight path of the unmanned aerial vehicle. According to the method, a local search mechanism and a dynamic adjustment strategy are introduced on the basis of a traditional genetic algorithm, the method is suitable for a combinatorial optimization problem, excellent performance is shown in unmanned aerial vehicle path planning, and a path with the shortest time can be efficiently found.

The invention discloses a grid-connected inverter parameter optimization method based on an improved genetic particle swarm algorithm. The grid-connected inverter parameter optimization method comprises the following steps: constructing a target function; performing population position initialization; calculating a fitness value of each particle at an initial moment, recording an optimal solution, and storing initial global optimal particle and individual optimal particle information; random disturbance is added to the speed and the position of the particles; updating the speed and the position of the particle; performing multi-point crossover and mutation operation; according to the current optimal solution, carrying out local domain search by adopting a fine search mode based on Laplacian distribution and carrying out global neighborhood search by adopting a search mode based on Cauchy distribution; and checking convergence conditions, if convergence occurs, outputting an optimal solution, and obtaining optimal parameters of the grid-connected inverter system. The method has the remarkable effects that by optimizing three key parameters in a control system and utilizing a nonlinear dynamic strategy to enhance the global search capability and the convergence speed of the algorithm, various defects in the prior art can be overcome.

The invention discloses a renewable energy distribution optimization method based on a genetic algorithm. The method comprises the following steps: S1, generating an initial population; s2, defining a fitness function, and evaluating the fitness of each individual in the population; s3, an improved roulette selection method is adopted, and individuals are selected according to the fitness; s4, performing crossover operation on the selected individuals, and generating new individuals by adopting a multi-point crossover and uniform crossover combined method; s5, performing mutation operation on the new individuals, and adopting a self-adaptive mutation mechanism; s6, making a dynamic adjustment strategy; s7, new individuals generated through selection, intersection, variation and dynamic adjustment are used for replacing old individuals, and a new generation of population is formed; s8, repeating the steps S2-S7 until a preset termination condition is met; and S9, after a termination condition is met, outputting an individual with the highest fitness in the population as a final renewable energy source distribution scheme. Renewable energy distribution is optimized through the improved genetic algorithm, and efficient and stable distribution is achieved.

The invention belongs to the technical field of mixing precision quantification, and discloses a mixing precision quantification method of a deep neural network and a related device. The mixing precision quantification method comprises the following steps: carrying out diversity processing on a population by using a variation strategy in a differential evolution algorithm to obtain a variation population; recombining the sub-components of the variation individuals in the variation population with the sub-components of the target individuals in the current population by using binomial crossover operation to obtain a plurality of test individuals; performing hybrid precision quantification on the deep neural network based on the test individual and the target individual, and selecting a better individual from the test individual and the target individual; and based on the selected better individual, judging whether the differential evolution algorithm converges or not, if so, obtaining the deep neural network after final mixing precision quantification, and otherwise, selecting the obtained better individual as a population of the next generation. The method can realize mixing precision quantification of the deep neural network, and has the advantages of relatively low calculation complexity and relatively high adaptability.

The invention provides a method for solving the computing power resource utilization rate based on a genetic algorithm, and relates to the technical field of computing power resource utilization. Comprising the following steps of S1, performing population initialization, randomly generating a group of initial solutions, and representing a computing power resource allocation scheme by each solution; s2, adaptability evaluation: defining a fitness function to calculate the fitness value of each individual; s3, genetic operation: performing selection operation, crossover operation and mutation operation to generate a new offspring; s4, generating a new-generation population, and selecting the most excellent individuals to form the new-generation population; s5, performing iterative computation, iteratively executing the genetic operation and the new generation population generation step, and continuously updating the population until a termination condition is met; and S6, outputting a result, and identifying an optimal solution or a group of excellent solutions from the current population. By simulating natural selection and genetic mechanisms in the biological evolution process, allocation of computing power resources is optimized, so that the resource utilization rate and the system performance are improved.

The invention discloses an equipment scheduling method and system based on dynamic priority and multi-objective optimization, and the method takes the minimization of a waiting time mean square error, the minimization of total waiting time and the minimization of maximum completion time as optimization objectives. A task dependency relationship, an equipment maintenance time window, a time window constraint, a student learning progress priority and a delay compensation mechanism constraint condition are comprehensively considered, a dynamic priority calculation mechanism is introduced, the priority is dynamically adjusted in real time according to task waiting time, deadline and urgency, and a high-quality initial population is generated through CMA-ES; a non-dominated sorting framework optimized by NSGA-II is combined, crossover variation of a traditional genetic algorithm is replaced by PSO updating, and an optimal solution is searched by using an elitism mechanism. The method can effectively improve the utilization rate of equipment resources, reduce the task completion time and optimize the fairness of task allocation, is suitable for equipment sharing and dynamic scheduling scenes, and has high practicability and popularization value.

The invention discloses an intelligent production scheduling method for a long-period operation chemical device, which takes a real-time state of a production device and a predicted future health condition as one of important basis of production scheduling, combines a traditional equipment failure mechanism model with a GBDT algorithm, can predict an equipment health level at a future moment, and improves the production scheduling efficiency. Guiding the dynamic adjustment of the equipment load and the production scheduling; and then iterative optimization of the scheduling scheme is carried out by using an adaptive genetic algorithm, dynamic adaptive crossover probability and mutation probability adjustment is carried out on the scheduling scheme in the intermediate process, and the rate of convergence of the algorithm to a globally optimal solution is improved. From the aspect of application effects, the long-period operation characteristic of the large chemical device and the continuity and stability of the production process can be considered while the production efficiency and the cost are ensured, and a scientific and feasible technical implementation scheme is provided for the production scheduling of the large device of a chemical enterprise.

The invention discloses a submarine searching sonar buoy array position optimization method based on a non-dominated sorting genetic algorithm NSGA-III, and the method comprises the steps: introducing a time scale factor in combination with the initial and maneuvering positions of a submarine, and obtaining a position probability distribution function of the submarine after maneuvering; a buoy position is generated in the whole search area according to probability distribution, more initial array position individuals are generated near a high-probability area, and the exploration tendency of a population to a key area is improved; when reference lines are selected during non-dominated sorting, fusion density estimation is added, so that the reference lines at individual sparse positions are denser; when the Pareto front is selected to be reserved, a reduced-proportion front-edge reservation method is selected, and except for the first front edge, subsequent individual reservation is gradually reduced in a reduced-proportion mode until the population scale is reached; introducing a gene adaptation entropy crossover threshold in the crossover part, adopting a priority retention strategy for the parent with high fitness, and then performing crossover operation on the remaining part; and the potential sonar buoy array position optimization is realized.

The invention discloses an unmanned aerial vehicle cluster formation configuration and spectrum resource joint optimization method, system, device, medium and product, and relates to the field of unmanned aerial vehicle planning control, and the method comprises the steps: dispatching a plurality of unmanned aerial vehicle formations to execute a plurality of tasks under the conditions of given spectrum resources and limited number of unmanned aerial vehicles, constructing an optimization problem model by using a coupling relationship among the number of the unmanned aerial vehicle formations, a target sensing sequence and spectrum resource allocation; solving the optimization problem model by using an improved genetic algorithm, and outputting an optimal unmanned aerial vehicle cluster formation configuration and spectrum resource joint optimization scheme; according to the improved genetic algorithm, a simulated annealing algorithm is embedded into the genetic algorithm, a crossover operator of the genetic algorithm is adjusted, and an elite strategy and a dynamically adjustable mutation rate are introduced into mutation operation. According to the invention, joint planning of unmanned aerial vehicle task planning, unmanned aerial vehicle formation number distribution and bandwidth selection can be realized, and the method has higher global optimization capability.

A k-parallel row sorting problem solving method considering multiple channels relates to the technical field of disassembly line layout, and mainly comprises the following steps: determining an objective function, calculating channel coordinate information in a region, generating a population P1 and randomly generating a population P2 by using a greedy strategy, combining the population P1 and the population P2 into an initial population P, calculating the fitness of all solutions in the initial population, and obtaining a k-parallel row sorting problem; screening an elite solution, selecting solutions except the elite solution in the initial population based on the fitness by using a roulette mechanism to execute genetic circulation, sequentially carrying out crossover and mutation operation based on Q-learning, combining the population with the elite solution after mutation with the population P1 and the population P2 to update the population, according to the maximum number of iterations, it is judged that iterative calculation continues or a result is output; according to the variable domain genetic algorithm based on Q-learning, an optimal solution for solving kPROPP can be provided in a short time, the solving efficiency of the variable domain genetic algorithm is greatly superior to that of a conventional accurate solver, the solving effect of the variable domain genetic algorithm is superior to that of other methods, and reliable support is provided for solving the problem of parallel layout planning.

The invention provides an evaluation method and system based on soil buffer capacity, and relates to the technical field of soil evaluation.The method comprises the steps that a microbial community diversity index is coded to serve as a genotype of a genetic algorithm; randomly generating an initial population, wherein each individual represents one environment factor combination; calculating the fitness value of each individual, carrying out selection, crossover and mutation operations, and repeating genetic operations until a termination condition is met so as to obtain a final key environment factor combination; obtaining microorganism activity data according to the soil respiration rate in the microorganism sample; and according to the final key environment factor combination, the microbial activity data and the trained neural network model, predicting a soil buffer capability value. According to the method, the buffer capacity of the soil can be evaluated more accurately.

The invention relates to a visual-language model cue word evolution generation method based on a genetic algorithm, which comprises the following steps of: randomly generating different cue words under a target task to construct an initial cue word set, preprocessing the randomly generated cue words, and reserving N groups of cue words as an initial population; and designing a multi-dimensional fitness evaluation function to evaluate the performance quality of each group of cue words in a vision-language task, and selecting high-quality cue word individuals based on an elitism strategy. Performing crossover and mutation operation on the selected high-quality cue word individuals by referring to a genetic algorithm to generate a new-generation cue word population; and finally, carrying out iterative optimization for multiple times until a preset termination condition is met, and outputting an optimal cue word set. By the adoption of the visual-language model cue word evolution generation method based on the genetic algorithm, high-quality and diversified cue words can be automatically generated, the performance of a visual-language model is remarkably improved, and meanwhile the visual-language model cue word evolution generation method has good interpretability and adaptability.

The embodiment of the invention discloses a self-adaptive rapid non-dominated sorting genetic method and system, a storage medium and equipment. The method comprises the following steps: setting a population scale, a maximum iteration generation number and a generation channel parameter; selecting a preset number of individuals from the initial population as a parent population according to the individual fitness on the basis of the population scale and the generational channel parameters; determining a self-adaptive crossover probability and a self-adaptive mutation probability according to the non-dominated level and the congestion degree of the parent population; performing crossover variation according to the adaptive crossover probability and the adaptive variation probability in the progressive stage and the mutation stage; performing non-dominated solution sorting on individuals in the parent population after crossover variation to form a new offspring population; and determining a final filial generation population according to the non-dominated level and the congestion degree of the combined parent population and the new filial generation population until the maximum iteration generation number is reached, and outputting a non-dominated optimal solution set. And multiple crossover and mutation are carried out through the adaptive crossover probability and the adaptive mutation probability, so that the convergence speed is accelerated.

System and method for hybrid analysis of quantum and classical genetic algorithms is disclosed. The method includes, receiving an input bitstring, the input bitstring being an output of a genetic optimization module, processing the input bitstring to generate quantum processed bitstrings, mutating the input bitstring, mutating the quantum processed bitstrings as a function of the input bitstring and the mutated input bitstring, performing crossover on a combination of the mutated input bitstring and the mutated quantum processed bitstrings, and selecting a set of individuals from the quantum processed bit strings and output of the crossover. The method further includes, determining, after selecting the set of individuals, if the hybrid analysis is complete or incomplete based on predetermined criteria, returning, in response to the hybrid analysis being incomplete, the set of individuals as the input bit string, else, outputting the set of individuals as a result of the hybrid analysis.

The invention discloses a micro spring washer production method based on genetic algorithm optimization, and the method comprises the following steps: S1, initializing a technological parameter set, and generating a preliminary optimization scheme through a morphological optimization control model; s2, mapping high-dimensional parameters to a low-dimensional morphological feature space through self-organizing mapping, and implementing adaptive variation adjustment; s3, integrating the processed morphological characteristics and technological parameters into chromosome codes, and performing multi-generation iterative optimization; s4, combining an evolutionary game model and fuzzy logic control, and dynamically adjusting a process parameter combination; s5, cooperatively regulating and controlling technological parameters through crossover and mutation operations; s6, performing fitness evaluation, and adjusting genetic algorithm parameters; and S7, after a convergence condition is reached, outputting an optimal process parameter, and applying the optimal process parameter to actual production. The method can significantly improve the production efficiency and the product quality, and is suitable for the precision manufacturing field of complex process optimization.

A technique for solving combinatorial problems, such as vehicle routing for multiple vehicles integrates evolutionary algorithms and reinforcement learning. A genetic algorithm maintains a set of solutions for the problem and improves the solutions using mutation (modify a solution) and crossover (combine two solutions). The best solution is selected from the improved set of solutions. A system that integrates evolutionary algorithms, such as a genetic algorithm, and reinforcement learning comprises two components. A first component is a beam search technique for generating solutions using a reinforcement learning model. A second component augments a genetic algorithm using learning-based solutions that are generated by the reinforcement learning model. The learning-based solutions improve the diversity of the set which, in turn, improves the quality of the solutions computed by the genetic algorithm.

The invention discloses a multi-logistics-vehicle dispatching method based on an accessibility matching cross genetic algorithm, and the method comprises the steps: building a multi-logistics-vehicle dispatching model with access constraints based on a scene that a customer has preference demands for logistics vehicle services due to the difference of logistics categories in practical application; a preference relationship of clients to logistics vehicles is described by using an accessibility matrix, crossover operation and mutation operation based on accessibility matching are designed, and a dispatching scheme of multiple logistics vehicles is optimized by embedding a genetic algorithm. Through the above operation, customer preference demands can be comprehensively considered in the iteration process, the total path of the multiple logistics vehicles and the path balance of the logistics workshops are optimized at the same time, and a globally optimal multiple logistics vehicle dispatching scheme is quickly found. According to the method, the solving efficiency of the multi-logistics vehicle dispatching problem with the access constraint is greatly improved, the dispatching efficiency of the logistics vehicles is effectively improved, and the transportation cost is reduced.

The invention belongs to the technical field of mobile crowd sensing, and particularly relates to a task allocation method based on a whale optimization algorithm framework, and the method comprises the steps: constructing an immune genetic whale algorithm framework based on the whale optimization algorithm framework; designing a whale individual structure representing a task allocation strategy; defining a whale individual fitness evaluation function; initializing a first generation whale individual population; designing a crossover operator, a mutation operator and an immune operator based on an immune genetic method; designing a repairing operator for repairing the invalid whale individual; specific operations such as population initialization, crossover, variation, immunization and repair are integrated into an immune genetic whale algorithm framework, so that the design of a task allocation algorithm is completed. According to the method, the local and global optimization capabilities of the algorithm are considered, so that the platform can generate a task allocation strategy capable of maximizing profit at a relatively high speed, and the task allocation rate and the activeness of participants are improved.

The invention discloses a method for solving a task-platform optimization configuration problem based on QGA quantum enhancement, and the method comprises the steps: firstly, employing the true randomness and entanglement of quantum, achieving the coding of a corresponding relation between a task and a platform through a quantum state, and determining an optimal individual in a population through a fitness function; secondly, a quantum logic gate is used for achieving selection operation in the algorithm, and the rotation angle and direction of the quantum logic gate are determined through the fitness difference value between each chromosome and the optimal chromosome; then, a quantum logic gate X and a quantum logic gate SWAP are used for achieving mutation operation and crossover operation in the algorithm respectively; the method provided by the invention has high accuracy and timeliness, the number of algorithm iterations and the time complexity are greatly reduced, and the method can be effectively used for solving the task-platform optimization configuration problem in the military C2 organization.

The invention discloses a multi-objective optimization method for performance of plate-shaped fuel. The method comprises the following steps: 1, carrying out model related setting in a fuel performance analysis program; 2, determining decision variables, optimization objectives and constraint conditions required by the multi-objective optimization method; 3, determining the number of individuals of each generation of population and the maximum number of iterations; 4, selecting a required multi-objective optimization algorithm, and performing population initialization operation; 5, introducing an elitist strategy, performing non-dominated sorting and distributing fitness; 6, calculating a non-dominated set optimization individual congestion degree and a constraint violation degree; 7, inputting the population individuals into a fuel performance analysis program, and respectively carrying out irradiation thermal hydraulic coupling calculation to obtain an optimization target value; 8, performing operations such as selection, crossover and variation on the current population individuals, and evolving to generate a filial generation population; 9, repeating the steps 5-8 until the algorithm judges that the optimization process is converged or the maximum number of iterations is reached; and 10, outputting an optimal solution set, namely design parameter information of p groups of dispersion particle diameters and pellet matrix thickness containing the optimized target value.

The invention discloses an 8-bit S box construction method based on an improved genetic algorithm, and the method comprises the following steps: (1) forming a composite chaotic system through Logistic chaotic mapping and Tent chaotic mapping, and generating an initialized population; (2) designing a fitness function by combining the nonlinearity, the difference uniformity and the strict avalanche criterion of the S box; (3) designing crossover operation for ensuring bijection of the S box; (4) designing a reverse search method to search variation points, combining a hill climbing algorithm and a random search method, designing variation operation based on 2N rounds of search, carrying out two-point variation point search in the first N rounds, adopting three-point variation point search in the last N rounds, and designing a table look-up statistical method to further evaluate and optimize the property of the S box; and (5) constructing an 8-bit password S box based on an improved genetic algorithm. The method is based on the improved genetic algorithm, and the 8-bit S box with excellent cryptographic properties can be generated.

The invention discloses a Mal-POBM malicious code confrontation sample generation method based on a genetic algorithm. The Mal-POBM malicious code confrontation sample generation method comprises the following steps: modifying static characteristics of an original sample; in adversarial sample generation, grouping optimization is carried out on a population, and crossover and variation are carried out at a local adaptive probability to increase the diversity of the population and increase the probability of finding an optimal solution; and when the population is sunk into the local optimal solution, catastrophe is carried out on the population by using the Cauchy probability. The malicious code adversarial sample can be accurately and quickly generated, and the method has important significance for evaluating and improving a malicious code detection system.

The invention relates to a hybrid evolutionary neural network search method, which comprises the following steps of: endowing initial weights to candidate operations of a plurality of neural network architectures, and optimizing the initial weights through a gradient descent algorithm to form a preliminary search direction; based on the preliminary search direction, generating a plurality of neural network architecture individuals by using Gaussian distribution sampling, and constructing an initial population; and performing fitness evaluation on the initial population, selecting individuals with better performance, generating new individuals through crossover and mutation operations, and constructing a new population. And the mean vector and the covariance matrix of the population are updated by using the distribution information of the excellent individuals so as to optimize the population distribution direction, and the neural network model with the optimal performance on the verification set is finally selected through iterative optimization search and updating of the neural network architecture. According to the method, the problem that a single gradient descent method is easy to fall into local optimum is effectively avoided, and the efficiency of neural network architecture search and the model performance are improved.

The invention provides a multi-scene vehicle scheduling optimization algorithm based on a parallel evolution framework, and provides the multi-scene vehicle scheduling optimization algorithm based on the parallel evolution framework through interactive population dynamic evolution and a Knee point evaluation strategy on the basis of a non-dominated sorting genetic algorithm (NSGAII) in combination with a KnEA algorithm. The established multi-scene multi-target vehicle path model with capacity constraint is optimized, and a compromise delivery route set (PCOS) which achieves a good effect in multiple scenes is found out. In order to solve the limitation of single crossover operator optimization, a self-adaptive crossover mechanism for a migration solution and a Knee point is designed on a multi-scene vehicle scheduling optimization algorithm framework of a parallel evolution framework. A three-population crossover operator is designed, and the updating of an optimized population quantity proportion matrix is realized by calculating the distance between a Knee point generated after the population is optimized by each operator and a migration solution to an extreme value line, so that the quantity of the population optimized by each operator is dynamically adjusted, and the quality of the PCOS can be remarkably improved.

The invention relates to the technical field of panel processing, in particular to a panel furniture workpiece typesetting method based on an improved genetic algorithm, which comprises the following steps: acquiring parameter information of a to-be-typeset workpiece; and performing classification based on the parameter information of all the workpieces to be typeset to obtain a plurality of workpiece categories, endowing each workpiece category with a unique type code, constructing an initial population according to the category and the number of the workpieces to be typeset, and inputting the initial population into the improved genetic algorithm to obtain an optimal chromosome. According to the method, the workpieces with the same size and shape are classified into workpieces with the same specification, and a repeated gene coding mode is adopted, so that the defect that actual repeated chromosomes are generated when a plurality of workpieces with the same specification are processed in a traditional unique coding mode is effectively avoided, and the algorithm efficiency is improved. And secondly, in the gene crossover operation, an improved ordered crossover and order crossover strategy is introduced, and a random deletion mechanism is combined, so that the algorithm population diversity and the solution space coverage are effectively improved, and the search divergence and convergence efficiency are improved.

The invention belongs to the technical field of petroleum and natural gas drilling engineering, and particularly relates to a borehole friction coefficient segmented inversion method based on a genetic algorithm and string mechanics, which comprises the following steps of: acquiring specified data of a construction well, dividing a borehole of the construction well to be predicted into well sections with fixed lengths, defining depth ranges of the well sections, and setting a global search range of a friction coefficient; initializing the population size, the maximum number of iterations, the selection probability, the crossover probability and the mutation probability of the genetic algorithm, and performing downward inversion section by section from a wellhead section: initializing a candidate friction coefficient population of a current well section, setting a sectional friction coefficient, calculating a hook load at the tail of the current section, and calculating the maximum number of iterations; iterative optimization is carried out until a termination condition is met, and then the optimal friction coefficient of the section is output and fixed; and iteratively executing until inversion of all well sections is completed, outputting a full-borehole segmented friction coefficient profile, and realizing intelligent segmented inversion of the borehole friction coefficient by combining a tubular column mechanical model and a genetic algorithm, so that the prediction precision of the friction coefficient is improved, and the well drilling design and construction are optimized.

The invention discloses a wind power generation tower drum steel-concrete composite structure design method based on genetic algorithm optimization. The method specifically comprises the steps that operation data and design parameters of different types of wind power generation tower drums under different working conditions are obtained to construct an original data set; secondly, determining a design target function and quantifying, and setting each target weight coefficient; a genetic algorithm population is initialized, and individuals are represented by design parameter codes; decoding individuals of the population to obtain an actual design scheme, performing structural mechanical analysis, and calculating mechanical property indexes; calculating an individual fitness value according to the target function and the weight coefficient; new individuals are generated through selection, crossover and mutation operation, whether termination conditions are met or not is judged, if yes, the design scheme corresponding to the individual with the highest fitness is output, and if not, iterative optimization continues. According to the method, the genetic algorithm is used for iterative optimization, the comprehensive performance of the design scheme of the steel-concrete composite structure of the wind power generation tower under multiple targets is effectively improved, and a better design result is obtained.

The invention belongs to the field of computational lithography, and particularly discloses a lithography model calibration method based on a proxy model assisted hybrid genetic algorithm, which comprises the following steps: determining to-be-optimized parameters of a lithography model, and generating an initial population according to the to-be-optimized parameters; performing parallel simulation on the current population to obtain a photoetching model simulation result of the current population; performing parallel evaluation on the current population to obtain photoetching model evaluation indexes and fitness distribution of the current population; screening excellent individuals in the current population to construct an agent model, and performing optimal prediction to obtain a first type of individuals; screening excellent individuals in the current population and carrying out crossover and variation to obtain a second type of individuals; and combining the first type of individuals and the second type of individuals, carrying out parallel simulation and evaluation on the next-generation population, and carrying out loop iteration updating until a preset termination condition is met, so as to obtain a photoetching model parameter set of the optimal individual. According to the invention, the calibration process of the photoetching model can be simplified, and the automation, precision efficiency and global search capability are improved.

The air freight loading efficiency and the flight safety are crucial to the efficient operation of a transportation system. Therefore, aiming at the problem of multi-cargo loading optimization under the complicated constraint condition of the cargo hold, a multi-objective optimization model containing business load limitation, gravity center shift, volume utilization rate and rotational inertia is constructed, and a solution method based on a grey wolf genetic hybrid algorithm is provided. A high-quality initial population is generated by adopting a two-stage initialization strategy, and evaluation is performed through a fitness function. And dynamically adjusting a step length coefficient and a mutation probability to realize self-adaptive adjustment of the search capability. In the selection stage, excellent individuals are reserved by using a tournament strategy, and meanwhile, the population quality is improved through an alpha wolf direct inheritance mechanism. In the breeding stage, the feasibility and diversity of solutions are enhanced through single-point crossing and mutation operation based on the gravity center distance. And after iterative updating, outputting the scheme with the highest evaluation value as the optimal loading scheme. According to the algorithm, the loading balance and the calculation efficiency are improved, and the high loading rate of the space in the cabin is ensured.

The invention discloses a heterogeneous cluster task scheduling fusion method based on Q learning and a genetic algorithm, and belongs to the technical field of intelligent scheduling. Collaborative optimization is carried out through a dynamic feedback mechanism of Q-Learning and the global search capability of the genetic algorithm: in an initialization stage, a cluster state space and an action space are defined, and a multi-Q-value table and an initial population are generated; during task scheduling, nodes are selected based on a # imgabs0 #-greedy strategy, reward values are calculated, and a Q value table is updated; meanwhile, the scheduling scheme is coded into chromosomes, a new population is generated through roulette selection, crossover and variation, and a Q value table is optimized; in the iteration process, the convergence is judged according to the Q value change or the population fitness change, and the strategy is dynamically adjusted. According to the method, the advantages of double algorithms are fused, local optimum is avoided, the task processing efficiency and the resource utilization rate are remarkably improved, adaptive strategy updating during cluster state change is supported, and the method is suitable for an efficient task scheduling scene of a large-scale heterogeneous cluster.