32results about How to "Avoid premature convergence" patented technology

The application relates to the technical field of human resource management, in particular to a medical staff scheduling system based on medical staff skills, which comprises a standard skill input and classification device, which is used for inputting the working skills of each department, dividing the same working skills of each department into general skills, and dividing the special working skills of each department into special skills; a department patient input device, which obtains the treatment information of patients in each department ward, extracts the predicted workload of each patient in the next scheduling cycle, and counts the predicted workload of each department; the predicted workload comprises the execution times of each general skill and the execution times of each special skill; the system realizes cross-department dynamic scheduling by constructing a hospital-level skill resource pool and a demand prediction mechanism, standardizes the general skills into shared resources, and enables medical staff to flexibly support departments with high demand peaks.

The application belongs to the field of workshop scheduling, and particularly discloses a semiconductor wafer probe test scheduling method and system based on an improved genetic algorithm, which adopts double-layer process-machine coding, and combines auxiliary resource taboo coding as an individual, effectively converts a complex scheduling problem into a coding form through an innovative coding and decoding strategy, and avoids traversal search on all selectable auxiliary resources; global-local hybrid search is carried out through the improved genetic algorithm, on the basis of machine level exchange with the traditional critical path-key block, search on the auxiliary resource immediately preceding work is increased, and neighborhood exploration is carried out based on a load balancing strategy, so that the search efficiency and solution quality are improved. The application can quickly and efficiently solve the semiconductor wafer probe test scheduling scheme, effectively improve the equipment utilization, speed up the scheduling efficiency, and break through the bottleneck of traditional scheduling.

This invention discloses an optimization method and system for smoke screen chaff deployment based on a view cone masking model, belonging to the field of signal jamming technology. The method includes the following steps: S1, constructing an adversarial kinematic model of the missile, target, and chaff; S2, establishing a view cone masking quantitative evaluation model with masking completeness as the core; S3, defining a fitness function with the maximum effective masking time as the objective; and S4, using an adaptive genetic algorithm to solve for the optimal deployment strategy that satisfies spatiotemporal constraints. This invention solves two major bottlenecks in existing technologies: the disconnect between effect evaluation and decision control, and the excessive reliance of advanced decision models on preset conditions. It achieves accurate, direct, and decision-driven quantitative evaluation of masking effects from the first-person perspective of the missile seeker.

The present application relates to a transformer fault diagnosis method based on improved sparrow search algorithm optimization SVM, through the pretreatment and feature engineering of transformer oil chromatographic data; according to the data of pretreatment and feature engineering, the model based on oil chromatographic data is constructed, and the model parameters are optimized and trained; the input transformer oil chromatographic data is carried out transformer fault diagnosis based on the optimized and trained model.The present application improves the sparrow search algorithm by introducing the best point set strategy, the golden regular update rule, the differential mutation disturbance and the reverse learning mechanism, so as to comprehensively improve the parameter optimization process of support vector machine, the global exploration ability of the algorithm is enhanced, the risk of falling into local optimum is avoided, and the problem of premature convergence is effectively avoided.In addition, the improved algorithm pays attention to multi-link optimization, and the parameter setting is simple, the algorithm is low in use difficulty, so that the optimized support vector machine (SVM) can effectively improve the transformer fault prediction precision.

This invention discloses a method for monitoring the state trends of hydropower units based on two-stage signal decomposition and the IBiLSTM model. The method involves collecting vibration signals from the hydropower units for preprocessing; constructing the ITGCOA optimization algorithm and designing a fitness function; using the ITGCOA optimization algorithm to adaptively optimize the SVMD and BAACMD models, achieving initial decomposition of the preprocessed signal and secondary decomposition of the sub-mode component with the highest center frequency obtained from the initial decomposition; calculating the fuzzy entropy values ​​of the remaining sub-mode components for reconstruction; and fusing the high-frequency feature sub-sequences obtained from the secondary decomposition with the reconstructed feature sub-sequences to construct the input sequence of the prediction model. This input sequence is then used to construct the IBiLSTM prediction model for monitoring the state trends of hydropower units, achieving high-precision prediction. Compared with existing technologies, this invention improves the efficiency and accuracy of hydropower unit state trend prediction, accurately warns of abnormal unit operating conditions, ensures the safe and stable operation of the units, and improves the overall efficiency of the power plant.

This invention discloses an AI-based low-code runtime anomaly identification system, comprising a data acquisition module, a node runtime anomaly identification module, a process runtime anomaly identification module, and a low-code runtime anomaly management module. This invention relates to the field of data processing technology, specifically to an AI-based low-code runtime anomaly identification system. This solution innovatively proposes performing full-dimensional anomaly prediction before low-code deployment and operation, and conducting three-level progressive runtime anomaly identification, thereby improving the stability of low-code process operation. It introduces TT tensor decomposition technology to construct a Laplace matrix, and uses a dynamic correlation matrix matching the current node component type as the correlation weight for graph convolution operations, improving the accuracy of the low-code node anomaly identification output. It employs a sine and cosine periodic perturbation chaotic mapping strategy and introduces a dual adaptive step-size mechanism to improve the optimization algorithm, thereby enhancing the stability and identification accuracy of the process runtime anomaly identification model.

This invention discloses an underwater electric field source localization method based on a determinant orthogonal propagation operator and adaptive hybrid-driven differential evolution, belonging to the field of electromagnetic positioning and detection. This invention acquires and receives array signals and calculates the covariance matrix. Based on the propagation operator, it constructs a projection operator matrix orthogonal to the array manifold to replace the noise subspace, thereby constructing a determinant spatial spectrum function and defining a cost function. Then, based on this cost function, it constructs an adaptive hybrid-driven differential evolution framework, fusing the alpha evolution algorithm and an adaptive differential evolution algorithm based on successful historical records to solve for the target position coordinates. This invention eliminates the need for eigenvalue decomposition, reducing computational complexity, and accelerates convergence through an adaptive hybrid-driven search strategy, enabling fast and high-precision electric field source localization in complex underwater environments.

Embodiments of the present application disclose a satellite telemetry data feature selection method and device based on a multi-objective optimization algorithm and a medium; the method can include: splicing the original telemetry data of the satellite, and screening out constant values in numerical data and time alignment, and based on the Rayleigh criterion, the 53H algorithm and the median filtering algorithm, removing outliers and noise in the original telemetry data to obtain preprocessed telemetry data; improving the traditional fast non-dominated sorting genetic algorithm NSGA-II according to the non-uniform arithmetic crossover operator and the mutation operator based on the fitness level; determining the objective function for feature selection according to the feature quantity, the root mean square error and the correlation function; based on the objective function, selecting feature data for data completion, satellite state prediction and fault diagnosis from the preprocessed telemetry data by using the improved NSGA-II algorithm.

This application relates to the field of wireless communication technology, providing an antenna robustness design method and apparatus based on hybrid deep learning. This invention simulates manufacturing process errors by combining sampling methods within the range of process errors, obtaining design variables and... S Sensitivity analysis was performed on the Gaussian distribution curves of the statistical mapping relationship between parameter responses to identify variables more sensitive to manufacturing errors, thereby reducing the dimensionality of variables and decreasing the complexity of subsequent antenna robustness optimization. A hybrid deep learning model was used to construct an antenna response substitution model, replacing traditional electromagnetic simulation, significantly shortening the optimization cycle while ensuring the accuracy of response prediction and avoiding distortion of optimization results due to model errors. By constructing a robustness objective function and employing a genetic algorithm to optimize antenna parameter robustness, a highly robust optimal solution was found, avoiding premature convergence of the gradient method in multi-peaked environments caused by random errors.

The application relates to the field of wind power prediction, and discloses a wind power prediction method and system based on DIRMO and a differentiated target function. In order to solve the defects that a single model is used for prediction in the existing multi-step power prediction, time series modeling and feature processing cannot be simultaneously considered, the prediction accuracy is reduced, the model robustness is insufficient, and the search efficiency is low, the power and wind speed data after preprocessing are denoised, normalized and feature-extracted to obtain multi-scale features; the normalized power and wind speed data are subjected to time series modeling to generate a preliminary prediction result in the future; the multi-step prediction task is divided into several groups, and a multi-output problem is converted into a single-output problem for training and prediction; the hyperparameters of a LightGBM model are automatically optimized, the optimized hyperparameters are used for training the LightGBM models of the groups, and the final power prediction is carried out based on the preliminary prediction result of the GRU and the multi-scale features. The application is mainly used for predicting wind power.

The application discloses a kind of dual-machine formation 2v2 cooperative air combat auxiliary decision generation method, which realizes the efficiency promotion of dual-machine cooperative combat by the dual optimization of tactics and maneuver, and the dynamic analysis of air battlefield situation is realized by cooperative tactical decision-making, and the tactical formulation is carried out according to the situation relationship of enemy and our side, and then the combat target is selected;According to the tactics and the distribution of the combat target tactical role;Based on the target maneuver action library, the target prediction state is calculated, and then the final target prediction state is selected;Then the final target prediction state and the current state of our aircraft are input, the control amount of our aircraft in continuous domain is optimized by improved EDA algorithm, and the optimal flight trajectory of our aircraft is obtained.

The application discloses a kind of low-carbon scheduling method, system, equipment and medium of power system, belong to power grid low-carbon scheduling technical field, method includes initialization parameter and constructs collaborative space;By particle swarm module optimization reinforcement learning hyperparameter;Reinforcement learning module interactive experience data generation;Extract priority sample and elite particle realize two-way experience sharing;Converge after output scheduling strategy.System includes initialization module, collaborative space construction module, hyperparameter optimization module, interactive storage module, two-way experience sharing module and strategy output module.The application is coupled by constructing dynamic collaborative optimization space and two-way experience sharing mechanism, particle swarm and deep reinforcement learning.Feedback guides particle swarm to realize adaptive disturbance, avoid local optimum trap.At the same time, spontaneous optimization network parameter, significantly enhance the adaptability and generalization ability of algorithm.The scheme breaks through single algorithm decision bottleneck, improves low-carbon scheduling efficiency and power system reliability.

The application discloses a multi-virtual local area network spanning tree optimization method based on a large language model and an adaptive temperature genetic algorithm, and belongs to the technical field of network topology optimization, and comprises the following steps: acquiring physical topology information of a target network to construct a network graph model; initializing a population of the genetic algorithm and allocating a use quota of high-value edges to each virtual local area network based on the network graph model; iteratively optimizing the population by using an enhanced genetic algorithm framework integrated with a large language model; dynamically recommending a genetic operation strategy of the current generation by the large language model according to multi-dimensional state indexes of the population, and performing multi-objective fitness evaluation on individuals in the population; and dynamically adjusting a temperature parameter by using a bidirectional state-aware adaptive temperature scheduling mechanism; and collecting optimal edge selection schemes of each virtual local area network spanning tree obtained through final optimization as a multi-virtual local area network network configuration scheme to perform network configuration. The method can realize intelligent global optimal configuration of multi-VLAN spanning trees.

This invention discloses a design method, device, and system for an ultrasonic amplitude transformer based on a genetic algorithm, belonging to the field of manufacturing technology. Considering the strong global search capability and excellent convergence of genetic algorithms, this invention combines the principles of genetic algorithms with a fitness function to construct a fitness optimization algorithm based on genetic algorithms to optimize the contour parameters of the ultrasonic amplitude transformer. This ensures that the expected optimal amplification factor is obtained under different optimization objectives, effectively avoiding the local optimum problem common in traditional design methods and improving design efficiency and performance. By optimizing the contour of the amplitude transformer's geometry, the amplitude transformer can achieve the preset target amplification factor under operating conditions, thereby improving the efficiency of ultrasonic energy transfer and obtaining the expected output displacement amplitude. The final optimization result outputs the optimal geometric contour and parameters of the amplitude transformer, which can then be used for subsequent engineering manufacturing and application according to actual needs.

This invention discloses a method for locating stormwater and sewage mixing points and identifying pollution sources in drainage pipe networks based on particle swarm optimization (PSO) algorithm, belonging to the field of drainage pipe network monitoring and diagnosis technology. The method includes: acquiring basic data of the drainage pipe network and constructing a SWMM hydraulic and water quality model; deploying monitoring points to acquire real-time monitoring data; determining the range of pollution source parameters; initializing the particle swarm using the PSO algorithm and updating the model parameters; calling SWMM for simulation calculations; calculating the fitness value based on the objective function and updating the extreme values; checking the termination condition and outputting the optimal solution. This invention achieves multi-condition adaptive identification through the organic combination of the PSO algorithm and the SWMM model, effectively solving the problems of low positioning accuracy and poor computational efficiency in complex pipe networks using traditional methods. It has outstanding advantages such as high positioning accuracy, strong adaptability, and excellent computational efficiency, providing an effective technical means for the treatment of stormwater and sewage mixing in urban drainage pipe networks.

This invention relates to the field of neural network architecture search and industrial intelligent processing technology, and discloses a structure search method for multi-output dendritic neuron models for industrial classification tasks. The method includes: initializing the synaptic connection weights and dendritic threshold parameters of the multi-input multi-output dendritic neuron model to generate an initial population; uniformly dividing the initial population into several subpopulations of equal size; calculating the temporal and spatial criteria for each subpopulation after task allocation and optimization, and dynamically allocating evolutionary computational resources for each subpopulation based on the fitness change rate represented by the temporal criterion and the population distribution state represented by the spatial criterion; implementing an accelerated sharing penalty mechanism to adjust fitness values ​​according to the crowding degree among individuals to maintain solution set diversity; and updating each subpopulation. This invention solves the problem of low convergence efficiency caused by uniform resource allocation in traditional large-scale multi-objective evolutionary algorithms during neural network architecture search.

The invention discloses a park integrated energy system scheduling method considering energy consumption of a central air conditioner cluster, and belongs to the technical field of energy management and optimization control. Generating a plurality of time sequence prediction samples in a future scheduling period; performing clustering analysis on each time sequence prediction sample, and selecting a plurality of typical day scene data; constructing a corresponding day-ahead scheduling optimization target based on the typical day scene data and the comfort degree deviation index; based on a day-ahead scheduling optimization target, solving through a day-ahead double-layer nested optimization model to obtain a scheduling result corresponding to each typical day scene; determining a next-day reference scheduling plan; and in the next-day park integrated energy system operation process, taking the next-day reference scheduling plan as an initial plan, and performing dynamic correction by adopting an intra-day rolling optimization and game coordination model. The method solves the problems that an existing park integrated energy system is insufficient in time scale coordination, the flexible utilization of a central air conditioner is limited, and the operation robustness is low.

The invention relates to an unmanned high-speed service area robot global path planning method and device and a robot. The method comprises the steps that 1, a path grid map is constructed; step 2, improving the Runyanlong optimization algorithm, including: generating an initial population; random key coding; judging the priority; introducing an affinity function; embedding a concentration adjusting mechanism; carrying out a double-variation strategy; step 3, bidirectional beetle antennae parallel search: taking a starting point and a target point corresponding to the initial path sequence generated after optimization in the step 2 as double search initial points, and synchronously executing the improved Rhouyanlong optimization algorithm; if the two-way search meets at the middle path points, directly butting to form an optimal path sequence; if not, selecting an optimal path sequence through a distance strategy and a time strategy; the distance strategy is to select a sequence with shorter path length, the time strategy is to select a sequence with shorter walking time consumption, and random selection is carried out when the length or time consumption is equal. The problem that the convergence speed is not high enough is solved.

This invention discloses a multi-model fusion method and system for permanent magnet motor fault diagnosis, belonging to the field of permanent magnet motor fault diagnosis technology. It aims to improve the flexibility and adaptability of models in multi-task learning and hierarchical classification by introducing multiple Softmax-layer CNNs and fusing them with multiple models, while effectively solving the problems of weak noise resistance and insufficient generalization ability of single models. Specifically, the technical solution of this invention first collects the vibration and current signals of the faulty permanent magnet motor, then uses Markov transfer fields to adaptively enhance the time-domain signals into images. Using a multi-Softmax-layer CNN and XGboost as base learners and an IWOA-SVM model as the meta-learner, a high-precision fault diagnosis result is finally obtained. This invention combines the advantages of multiple Softmax layers and multi-model fusion, effectively solving the multi-task fault diagnosis problem under complex working conditions, and improving the accuracy, reliability, and adaptability of fault diagnosis.

The invention discloses a software defect feature selection method, device and system and a storage medium. The method comprises the following steps: S1, carrying out feature coding on software measurement features of a software defect historical data set; s2, performing sectional probability initialization on the initial population by adopting a minimum redundancy and maximum correlation algorithm according to the coded software measurement characteristics to obtain a plurality of sub-populations; s3, performing parallel chromosome evolution operation and crossover mutation operation on each sub-population; and S4, after parallel evolution and overall crossover variation of each sub-population, putting a parent chromosome and a generated sub-chromosome sum into a chromosome pool, selecting an optimal individual chromosome according to a fitness descending order, then carrying out remodeling of the sub-population, and continuing to repeat genetic manipulation until the maximum number of iterations is reached. By adopting the technical scheme of the invention, a high-quality feature combination which is most effective to predict defects and has the lowest redundancy is quickly and accurately screened out from high-dimensional software features.

This invention discloses a prompt word optimization control method combining difficulty grading and simulated annealing. The method includes: collecting failed case data to construct a multi-dimensional feature vector, obtaining a difficulty score, and allocating resource parameters and setting an initial annealing temperature accordingly; generating prompt word variants using a mutation strategy and calculating the pass rate; constructing a dynamic threshold model containing an iteration time power-law term and a hyperbolic tangent term for remaining difficulty to screen the parent population; monitoring changes in the pass rate to adaptively adjust the annealing temperature, and calculating the acceptance probability based on a strategy difference correction term to determine whether to accept inferior solutions. This invention achieves dynamic resource allocation and adaptive adjustment of the search process, effectively improving the accuracy of prompt word optimization and global optimization capability.

The invention discloses a two-stage constraint multi-objective optimization evolution method based on dynamic weight vector selection, and relates to the technical field of multi-objective optimizing.The method comprises the steps that in the first stage, a full-constraint population, a constraint relaxation population and an unconstraint population are maintained in parallel, and the populations are guided to penetrate through an infeasible area through a weak cooperation mechanism; and in the second stage, the unconstrained population is stopped, the distribution information of the constrained relaxation population is utilized, a potential search direction is extracted through a dynamic weight vector selection mechanism, and a search mechanism based on a dynamic vector is adopted to carry out refined environment selection on the main population so as to collaboratively approach a real leading edge. In addition, a high-quality feasible solution is reserved through an external file. The method shows better convergence, diversity and constraint satisfaction capability on a plurality of complex constraint optimization problems, and is suitable for practical engineering optimization scenes such as vehicle path planning and energy scheduling.

The invention discloses an aviation composite material solidification scheduling method and system considering multilayer space and time window constraints, and the method aims at multilayer board equipment space constraints, workpiece time window restrictions and batch processing demands in aviation composite material solidification production, and minimizes the maximum completion time as a target function. A single-machine batch scheduling model fusing multilayer space placement, laminate capacity and time window constraint is constructed, and meanwhile, an evolutionary algorithm combined with adaptive neighborhood search is adopted for solving; the improved strategies such as population initialization with time compatibility and random combination, a multi-dimensional neighborhood search operator and double-layer self-adaptive neighborhood selection are designed according to the characteristics of the composite material solidification scheduling, the global search and local optimization capability of the evolutionary algorithm is improved, and the solving quality and efficiency of the composite material solidification scheduling under the complex constraint are remarkably improved. The method can be widely applied to complex manufacturing scenes with multi-layer processing space, time window constraint and batch processing characteristics, such as aviation composite material forming and heat treatment furnace scheduling.

This invention proposes a method and system for optimizing warehouse locations in logistics hubs for multi-production line collaborative production, belonging to the field of manufacturing logistics and warehousing optimization. The method includes: constructing a logistics hub warehouse location optimization model for multi-production line collaborative production; performing multi-objective collaborative optimization using the NSGA-II algorithm based on the logistics hub warehouse location optimization model; outputting a Pareto optimal solution set representing the trade-offs between different objectives; performing simulation verification and sensitivity analysis on the obtained Pareto optimal solution set, predicting the performance of each solution under different production scenarios through logistics simulation, evaluating the impact of changes in key parameters on the optimization objectives, and outputting a final warehouse location configuration scheme with scientific decision support. This invention solves the problems of traditional methods neglecting production line collaboration, insufficient multi-objective balancing ability, and poor dynamic adaptability, achieving global, efficient, and adaptive optimization of logistics hub warehouse locations.

The invention discloses a fuzzy PID (Proportion Integration Differentiation) control parameter online optimization method based on multi-population collaborative PSO (Particle Swarm Optimization), which comprises the following steps of: establishing a mathematical model of a control object, and solving a transfer function of the control object according to structural characteristics and a working principle of a system; establishing a fuzzy PID controller, setting a fuzzy rule and configuring related parameter values; operating an improved PSO particle swarm algorithm, wherein an algorithm optimization space is a value interval of a proportionality coefficient, an integral coefficient and a differential coefficient corresponding to the fuzzy rule; the speed and the position of each particle are updated, the fitness value of each particle is obtained in the learning process, and the extreme values of individual particles and population particles are updated in each iteration; and optimizing the output parameters of the fuzzy controller according to the result of the PSO particle swarm algorithm. According to the method, a gradient-oriented local optimization mechanism is introduced, the global search capability and convergence stability of the PSO algorithm are improved, parameters of the fuzzy controller are dynamically adjusted by improving the PSO algorithm, and the control performance of the system is improved.

This invention belongs to the field of agricultural machinery operation scheduling technology, specifically involving an agricultural machinery operation scheduling method, system, equipment, and medium based on EDIM and NMPD, which can balance uncertainty modeling, high-quality scheduling solution capabilities, and engineering feasibility. This invention achieves a dynamic adaptive balance between global exploration and local development through the directional guidance of the EDIM mechanism and the multi-level neighborhood mining of the NMPD mechanism, effectively avoiding premature convergence. This allows for the generation of scheduling schemes with significantly shorter total travel distances in various real-world farmland scenarios, resulting in a significant improvement in operational efficiency. The system of this invention, by integrating the core optimization engine of the aforementioned scheduling method, can automatically process large-scale farmland and agricultural machinery data and output high-quality scheduling schemes with low empty-running rates.

The application discloses a kind of considering multi-layer space and time window constraint's aviation composite curing scheduling method and system, the method is for the space constraint of multi-layer board equipment in aviation composite curing production, workpiece time window limit and batch processing demand, with minimizing maximum completion time as objective function, construct the single machine batch scheduling model of fusion multi-layer space placement, layer board capacity, time window constraint, while solving by using the evolutionary algorithm of combining adaptive neighborhood search, population initialization of time compatibility and random combination is designed for composite curing scheduling characteristics, multi-dimensional neighborhood search operator, double-layer adaptive neighborhood selection and other improvement strategies, improve the global search and local optimization ability of evolutionary algorithm, significantly improve the solution quality and efficiency of composite curing scheduling under complex constraint.The application can be widely applied to aviation composite forming, heat treatment furnace scheduling and other complex manufacturing scenarios with multi-layer processing space, time window constraint and batch processing characteristics.

This invention provides a distributed resource dynamic allocation method oriented towards task evolution, comprising the following steps: S1, constructing a task demand evolution model and recursively calculating the probability distribution of task demand events at each stage based on conditional probability; S2, receiving the probability distribution of task demand events at each stage, constructing a distributed resource allocation model oriented towards task evolution, the distributed resource allocation model taking maximizing the resource system's processing capacity in response to task demands as its primary objective and balancing the task demands at each stage as its secondary objective, establishing a multi-objective nonlinear programming model; S3, inputting the multi-objective nonlinear programming model into a heuristic solution algorithm, using a segmented symbolic encoding multi-strategy genetic algorithm for solution, and outputting the optimal resource allocation scheme. This invention, by combining with multi-stage conditional probability recursive calculation, achieves a quantitative description of the stage-wise evolution law of task demands, providing accurate probabilistic input for dynamic resource allocation.

The application provides a prefabricated component distribution scheduling optimization method based on a time window and multiple vehicle types, and relates to the technical field of distribution scheduling. Parameter information of multiple transport vehicles and distribution time windows of construction sites are obtained and utilized. An optimization function with the dual objectives of minimizing distribution cost and maximizing vehicle average loading rate is constructed. By simultaneously optimizing the two objectives, a balanced deployment scheme between economy and resource utilization efficiency can be obtained, avoiding the one-sidedness that may be caused by single-objective optimization. A multi-objective ant colony-genetic hybrid optimization algorithm is used for solving, generating multiple deployment schemes with different trade-offs for users to select. Through comprehensive consideration of multiple vehicle types, time windows, dual-objective optimization and hybrid intelligent algorithms, an efficient and flexible prefabricated component distribution scheduling optimization solution is provided, effectively improving resource utilization efficiency in the multiple vehicle type mixed distribution scene and improving the practicality of the distribution scheme.

The invention discloses a mechanical arm configuration optimization method based on an improved squirrel search algorithm. The method comprises the steps that S1, a joint module library of a mechanical arm is established; s2, initializing a population, decoding each individual vector and establishing a kinematics and dynamics model; s3, loads are gradually increased under zero configuration, configurations meeting the load capacity are screened out according to torque constraint, and the configurations which do not reach the standard are directly discarded; s4, calculating a comprehensive fitness value based on a plurality of preset performance indexes; s5, sorting according to the fitness values, reserving an optimal solution and a suboptimal solution, and updating the population through a seasonal factor and a decline disturbance strategy; and S6, outputting a global optimal configuration after the maximum number of iterations is reached. The joint parameters, the connecting rod size, the motor parameters and the like of the mechanical arm are optimized by improving the squirrel search algorithm, so that a plurality of performance indexes are well balanced, and the comprehensive performance of the mechanical arm is improved.