41 results about "Degree function" patented technology

The invention discloses an automatic instrument fault prediction method based on fuzzy logic, and belongs to the technical field of industrial automation and equipment fault diagnosis. Comprising the following steps: generating a preprocessed sensor data set; constructing a feature space fusing multi-dimensional information; constructing a three-layer fuzzy reasoning architecture comprising an input layer, a reasoning layer and an output layer; mapping membership function parameters and rule weights in the three-layer fuzzy reasoning architecture into a parameter optimization space comprising a parameter adjustment action, a rule weight optimization action and a reasoning structure adjustment action; self-adaptive adjustment of the membership function parameters is achieved; and calculating the fault probability according to the current optimized membership function parameter, and evaluating the optimization effect of the current parameter configuration in combination with the multi-target reward function. According to the method, through combination of multi-time window dynamic monitoring, multi-type membership function design and a strict hard-soft constraint mechanism, accurate evaluation and stable prediction of the equipment fault probability are realized.

The invention relates to the technical field of avionics, and discloses a friend behavior prediction and multi-platform autonomous decision-making method and system during communication interruption, and the method comprises the steps: constructing a multi-level fuzzy inference tree, and carrying out the fuzzy processing of multi-source heterogeneous input information, and forming a plurality of fuzzy sets; decomposing the complex input relation into a plurality of subsystems based on a hierarchical fuzzy reasoning mechanism, and obtaining a subtask response benefit value through layer-by-layer reasoning; input variables are fuzzified based on a triangular membership function, and a potential task allocation scheme of the friend formation is predicted in combination with a hierarchical rule reasoning mode. According to the method, centerless, low-dependence and high-robustness multi-platform autonomous task allocation can be realized, and the collaborative decision-making capability and task execution efficiency of formation members in a complex environment under an incomplete information condition are improved.

The invention provides a pipeline transaction bilateral contract matching method and system, belongs to the field of crossing of metering economics and computer algorithms, and is used for solving the problems of multi-parameter coupling, high research and development uncertainty and low matching efficiency of pipeline transaction contracts in related technologies. A core term is converted into a tensor model, a Markov chain is combined to quantify expected utility, multi-round intelligent matching is realized based on a fuzzy satisfaction function and a sub-scene convex optimization algorithm, a key parameter dynamic calibration mechanism is matched, a complete technical closed loop is formed, the matching accuracy and negotiation efficiency are objectively improved, and the negotiation efficiency is improved. And reliable technical support is provided for complex pipeline transactions.

The present application relates to a kind of video enhancement inference method, device, equipment and medium based on fuzzy control, method includes: initialization multiple different scale pre-training video inference model;Define the fuzzy semantic set of input variable and output variable and membership function;Based on prior knowledge, construct fuzzy rule table;Real-time acquisition input data and complete fuzzy inference, and accurate score is obtained by solving fuzzy output by barycenter method;According to the score value range and anti-shake strategy, the optimal video inference model of next frame is dynamically selected.This method makes full use of the continuity characteristics of target between video frames, realizes the dynamic balance of hardware resources and inference accuracy, the utilization rate of GPU is obviously improved, the target detection rate is also obviously improved, and the temperature is stable within the safety threshold, suitable for real-time video scene such as traffic monitoring, unmanned aerial vehicle inspection.

The invention discloses a multi-AGV formation keeping control method based on IPSO-fuzzy PID. The multi-AGV formation keeping control method is characterized by comprising the following steps: (1) establishing a dynamic self-adaptive multi-AGV pilot following formation control model; (2) constructing a fuzzy PID controller, setting a fuzzy domain of input and output variables, a membership function and a fuzzy control rule, and obtaining the correction of PID parameters through fuzzy reasoning and defuzzification; (3) constructing an improved particle swarm optimization algorithm with dynamic balance of global search and local optimization, improving the inertia weight and learning factor of a particle swarm, and setting a fitness function; and (4) optimizing the input quantization factor and the output scaling factor of the fuzzy PID controller by using an IPSO algorithm, assigning optimized parameters to the controller, and adjusting the linear velocity and the angular velocity of the following AGV. The formation keeping precision, the response speed and the anti-interference capability can be improved.

The application discloses a mechanical arm fuzzy PID control method based on quantum pigeon group optimization, first motion trajectory planning is carried out, and a mechanical arm is controlled to move according to the planned motion trajectory through a fuzzy PID controller; a QAPIO algorithm is used to optimize fuzzy PID control parameters, and each pigeon can be used as a PID control parameter of each joint of the mechanical arm. The QAPIO algorithm of the application modifies a compass operator update equation, a probability density function in quantum representation of real number coding and a quantum rotation gate in the original algorithm, so that the algorithm has a continuous updating capability, can continuously update an optimal value, improves global optimization capability, avoids premature local optimization, and solves the problem that the original algorithm is prone to falling into a [0, 0,..., 0, 0]T point, causing the optimization algorithm to fail. The application effectively realizes optimization of fuzzy PID controller parameters of the mechanical arm, and has good practicability.

This invention relates to a vehicle differential lock control method, terminal device, and storage medium. The method includes: S1: constructing a first fuzzy membership function, a second fuzzy membership function, and a fuzzy rule inference table; S2: real-time acquisition of environmental values, road grade, and road curvature values, obtaining the maximum value of road curvature; S3: obtaining all corresponding first fuzzy intervals and fuzzy values ​​from the first fuzzy membership function based on the road curvature ratio; S4: obtaining all corresponding second fuzzy intervals and fuzzy values ​​from the second fuzzy membership function based on the environmental values; S5: matching the first fuzzy intervals with the second fuzzy intervals to obtain a combined fuzzy value of the fuzzy interval combination; S6: taking the fuzzy interval combination corresponding to the maximum value as the fuzzy decision combination, and obtaining the decision result from the fuzzy rule inference table; S7: controlling the differential lock based on the decision result. This invention achieves timely differential locking of the vehicle in weather conditions such as rain and snow when slight slippage occurs during normal driving.

This invention relates to the field of control, and in particular to a method and system for controlling the operation of a reactor used in resin production. The method involves acquiring real-time operating parameters of the reactor, iteratively training a fuzzy neural network model using a hierarchical collaborative hybrid optimization strategy. The antecedent membership function of the fuzzy neural network model is composed of a Gaussian mixture model, and the consequent is a nonlinear polynomial function of the input variables. The training aims to minimize the comprehensive objective function. The optimization strategy involves using an improved quantum particle swarm optimization algorithm for global search to determine the Gaussian mixture model parameters, and using a recursive least squares algorithm for local search after each iteration to determine the consequent coefficients. During training, importance indices are calculated based on the average activation degree of the fuzzy rules and their contribution to the prediction error. Rules with importance consistently below a preset threshold are removed. After training, control commands for controlling the heating system power and material feed rate of the reactor at the next moment are generated and executed based on the real-time parameters.

The embodiment of the application discloses a kind of based on FG-DBN's multi-electric aircraft airborne system reliability analysis method, it is related to reliability design field, can be in fuzzy uncertain condition, to the complex polymorphic system in multi-electric aircraft airborne equipment is carried out dynamic reliability analysis.This application includes: the DBN model of complex polymorphic system in airborne equipment is established, the node in DBN model corresponds the failure state of component in complex polymorphic system;Fuzzy gradient function model containing fuzzy support radius is established, fuzzy gradient function model includes: fuzzy membership model of node in corresponding time slice and state transition matrix corresponding to failure mode;The conditional probability table is established for the failure relationship between nodes in DBN model;The output result of fuzzy gradient function model containing fuzzy support radius and the conditional probability table established are used, the node of weak link in complex polymorphic system is identified and reliability is analyzed.

The application belongs to the technical field and discloses a control method and equipment of a high-speed motor and an electromagnetic bearing system, the method comprising the following steps: selecting input variables and output variables of a fuzzy controller; fuzzifying the input variables and the output variables, determining a quantization factor, a basic domain and a fuzzy domain; defining fuzzy subsets and membership functions in the fuzzy domain of the input variables and the output variables; calculating the membership degrees through the membership functions; substituting the membership degrees into a fuzzy value formula to obtain fuzzy values of the output variables, obtaining accurate control amounts of the output variables through defuzzification; and outputting the accurate control amounts of the output variables to an electromagnetic bearing model for control. The method can effectively reduce the oscillation and vibration of the system, improve the stability of the overall system, and adapt to large disturbances or complex environments.

The invention relates to the technical field of intelligent control of rail transit vehicle-mounted equipment, in particular to a quantum fuzzy control method and system for a rail transit vehicle-mounted air conditioner, and the method comprises the steps: collecting the operation data of the vehicle-mounted air conditioner in real time, and carrying out the Kalman filtering noise reduction and EMC anti-interference compensation preprocessing; performing four-dimensional variable quantum bit mapping coding on the preprocessed operation data, and constructing a quantum superposition state including disturbance compensation; a fuzzy membership function is optimized based on quantum probability distribution, a quantum fuzzy rule base is called, and an air conditioner control instruction is generated through quantum annealing algorithm iterative reasoning; an execution mechanism is driven to execute the air conditioner control instruction, the adjusted operation data, the comfort index and the state of the solid state contactor are collected, and deviation is calculated; and if the deviation meets a preset threshold value, maintaining the current parameter, otherwise, returning to re-execute quantum mapping coding and quantum fuzzy reasoning. According to the invention, accurate control of the vehicle-mounted air conditioner of the rail transit vehicle can be realized.

The invention relates to the technical field of rectifier power dynamic allocation methods, in particular to a CC-Dickson fusion rectifier power dynamic allocation method based on self-evolution fuzzy control, which comprises the following steps of: acquiring actual operation parameters of a rectification system; determining a load power dynamic distribution coefficient; establishing a load characteristic model; constructing a fuzzy control rule; generating a membership function relation matrix; establishing a proportional-integral control algorithm model; a multi-level self-adaptive control algorithm based on a Riemann-Zeta function is executed; adjusting the working mode and parameters of the CC-Dickson fusion rectifier based on the output of the multi-level adaptive control algorithm; the system performance is continuously monitored, the multi-level self-adaptive control algorithm is executed circularly, and the projection of the Riemann-Zeta function is matched with the mapping of transformation, so that the calculation complexity is reduced, and the key nonlinear characteristics of the system are kept.

The invention provides a multi-factor fuzzy comprehensive evaluation method for the emergency degree of a collapse scene area, and belongs to the field of rescue area evaluation. Establishing a factor set influencing the importance degree of the collapse scene area; key parameter assignment of a continuous membership function and enumeration mapping of a discrete membership function are realized; constructing an initial fuzzy rule base; the initial fuzzy rule weight is optimized, and an updated fuzzy rule base is obtained based on the optimized weight; and carrying out comprehensive evaluation on the input collapse scene area information by utilizing the updated fuzzy rule base to obtain a result. According to the method, a fuzzy system is constructed to comprehensively analyze multiple factors such as casualty conditions, building danger degrees, collapse structures and rescue operation conditions in different regions in a collapse scene, accurate identification of regional emergency rescue emergency degrees is realized, and a scientific basis is provided for rescue personnel to formulate a rescue strategy.

The application discloses a damaged ancient Zhuang character detection and recognition method based on fuzzy logic and belongs to the technical field of ancient character intelligent recognition. First, the damaged ancient Zhuang character image is collected, and preprocessing is completed through adaptive denoising, stroke loss detection and completion; then, multi-dimensional fuzzy features such as stroke integrity, contour similarity, structural connectivity and texture consistency are extracted, corresponding membership functions are constructed, and membership values are calculated; fuzzy processing, rule matching and defuzzification processing are completed based on a fuzzy inference rule base, and the damage degree grade and candidate recognition confidence are output; then, high-dimensional semantic features are extracted through a deep learning model, and the fuzzy inference result is weighted and fused to obtain the final recognition result; finally, the modern character mapping database of ancient Zhuang characters is queried to output the interpretation. The application combines fuzzy logic and deep learning, effectively processes uncertain problems such as stroke loss and contour fuzziness, significantly improves the recognition accuracy and robustness of damaged ancient Zhuang characters, is suitable for damaged scenes such as epigraphy, ancient books and cliff inscriptions, and provides a reliable technical scheme for the digital protection of ancient Zhuang character cultural heritage.

An arithmetic apparatus includes an arithmetic unit configured to adjust an intensity schedule function in quantum annealing of a constraint term expressed by many-body interactions in an LHZ model by a hybrid algorithm that uses a value of the intensity schedule function at one time point or each of a plurality of time points as a variational parameter.

This invention discloses a fuzzy multi-attribute decision-making evaluation method and system. The method includes defining a hesitant fuzzy language terminology set; determining the evaluation object and its influencing factors based on the hesitant fuzzy language terminology set to form a factor set, and assigning weights to each influencing factor to determine the membership function of each factor; performing fuzzy synthesis operations on the membership function of each factor and its corresponding weights to obtain a comprehensive membership value; and performing fuzzy evaluation on the evaluation object based on the comprehensive membership value to obtain a fuzzy comprehensive evaluation result of the evaluation object. This invention's multi-attribute decision-making method, capable of handling incomplete and imprecise fuzzy information, plays a crucial role in ensuring the rationality and accuracy of investment risk assessment.

The application discloses a servo press motion control method and system, and relates to the technical field of servo press control. The method comprises the following steps: acquiring displacement, velocity, acceleration and jerk boundary parameters of stamping section start and end points, and constructing a seven-degree polynomial function; determining zero to third order low order term coefficients based on start point parameters; adopting an intelligent optimization algorithm to iteratively solve fourth to seventh order high order term coefficients to minimize a multi-objective fitness function; pre-storing the coefficient set in a controller register; in each interpolation period, the controller calls the coefficients to calculate displacement instructions and outputs, simultaneously collects actual slider displacement, calculates deviation and adjusts the low order term coefficients in the register in real time, which are used for instruction calculation at the next moment. By adjusting only the low order term coefficients and keeping the high order term unchanged in real-time control, the application effectively avoids curve oscillation caused by online adjustment of high order coefficients, and improves reliability.

The embodiment of the invention provides a fuzzy query method based on an exponential distribution probability density function and a processor. The method comprises the steps of obtaining sample information in a database; the sample information is an attribute value of an object associated with the user, and the attribute value obeys the index distribution; dividing a fuzzy interval according to the sample information and the discourse domain; calculating a maximum value of a membership function corresponding to each fuzzy interval according to the threshold value and the weight of each fuzzy interval and an inverse function of the exponential distribution probability density function; according to the maximum value of the membership function corresponding to each fuzzy interval and the discourse domain boundary, matching the membership function corresponding to each fuzzy interval; and performing fuzzy query in the database according to a preset fuzzy query condition, and performing defuzzification processing based on the matched membership function to obtain a target query result. According to the scheme, an accurate query result can be obtained during fuzzy query.

This application provides a data processing method, electronic device, storage medium, and product. The method includes acquiring multiple attribute data corresponding to data to be processed; calculating the comprehensive fuzzy proximity degree corresponding to each attribute data based on the multiple attribute data of the data to be processed; and calculating the attribute fusion result corresponding to the data to be processed based on the comprehensive fuzzy proximity degree. The method obtains the comprehensive fuzzy proximity degree of multiple attribute data corresponding to any position of the data to be processed in the data matrix through a membership function, then calculates the weight value of each attribute data based on the comprehensive fuzzy proximity degree, and finally performs data fusion to obtain a comprehensive attribute fusion result. Although the attribute fusion result loses the physical meaning inherent in the original attribute data, it retains the commonalities and advantages of each attribute data, enabling a more accurate, comprehensive, and reliable description of the same target. Therefore, it improves the accuracy of describing and characterizing underground geological structures.

This invention discloses a software quality evaluation method based on fuzzy theory. The method includes: obtaining a set of quality factors based on the McCall software quality model; determining an evaluation level set and dividing the evaluation level set into five levels; constructing a membership function using kernel density estimation; constructing a judgment matrix using triangular fuzzy numbers and calculating defuzzification weights; converting the score into a fuzzy composite value through fuzzy synthesis operations, and generating a fuzzy comprehensive evaluation score through weighted averaging and normalization. This method innovatively combines fuzzy theory with the McCall model, using fuzzy mathematics to handle the uncertainty in subjective and objective evaluations. By constructing a membership function and weight set based on historical data of similar software, the objectivity, scientific rigor, and repeatability of the evaluation results are effectively improved. This method is applicable to quality assessment and optimization decisions for various complex software systems.

The embodiment of the invention provides an interestingness prediction method and device, equipment, a storage medium and a program product, and relates to the technical field of financial science and technology or the technical field of artificial intelligence. The method comprises the following steps: acquiring a plurality of fuzzy sets, wherein the plurality of fuzzy sets are related to influence factors of interestingness of a user on a product; according to the user data and the multiple membership functions, first membership degrees are determined, and the first membership degrees comprise multiple membership degrees, corresponding to the multiple fuzzy sets one by one, of the user data; according to the first membership degree, determining a feature vector of the user data; according to the feature vectors, fuzzy C-means clustering is executed, second membership degrees are obtained, and the second membership degrees comprise C membership degrees corresponding to the C clustering clusters one by one by the user; and carrying out weighted sum processing on the second membership degree and preset weights corresponding to the C clustering clusters to obtain the interestingness of the user on the product. According to the method, the accuracy of interestingness prediction is improved.

The invention discloses a multi-modal remote sensing image classification method and device and a storage medium, and the method comprises the steps: constructing a fuzzy enhanced gating fusion Mama network which comprises a feature extraction and fuzzy processing main network, a dual-branch gating fusion unit and a classifier; the feature extraction and fuzzy processing main network comprises a spatial Mama block, a spectrum Mama block and an HSI exclusive fuzzy block; the double-branch gating fusion unit is a gating cross-modal fusion block; the classifier is composed of an average pooling layer, a linear layer and a Softmax layer; wherein the HSI exclusive fuzzy block dynamically models fuzzy distribution of spectral characteristics through a learnable membership function based on a Gaussian fuzzy theory and a fuzzy mathematics principle; calculating the difference between the prediction category and the real category by adopting a cross entropy loss function, and updating network parameters through back propagation iteration by utilizing an Adam optimizer until the loss function converges or reaches a preset number of iterations, so as to complete network training; and classifying the multi-modal remote sensing images based on the trained network.

The invention discloses a target identification basic probability assignment generation method based on an evidence theory, and the method comprises the steps: calculating a membership degree corresponding to an interval endpoint value through employing a triangular fuzzy number according to a target attribute interval value observed by a sensor, representing a membership degree function of a target attribute as a triple, according to the interval endpoint values, the numerical relationship of the triple and the membership degrees corresponding to the interval endpoint values, obtaining the initial interval membership degrees of the target attribute interval values belonging to the targets; performing normalization processing on the initial interval membership degrees of the various types of targets to obtain basic probability assignment; and according to the detection reliability of each sensor under the interference condition, correcting the basic probability assignment to obtain the basic probability assignment of each sensor after correction of various types of targets. According to the embodiment of the invention, the problem of basic probability assignment of how to convert the detection result into the evidence focus element when the target identification detection value is the interval value in the information confrontation environment is solved.

The invention belongs to the technical field of hard disk tuning, and particularly relates to an HDD hard disk automatic tuning method based on a fuzzy algorithm, which comprises the following steps of: mapping core parameters into low, medium and high fuzzy subsets by monitoring and preprocessing multi-dimensional operation data such as I / O load, disk rotating speed, temperature and the like of an HDD hard disk in real time; a load type-tuning parameter fuzzy rule base is called, a tuning parameter fuzzy output set is generated through Mamdani fuzzy reasoning, a gravity center method is adopted for defuzzification, and specific tuning decisions such as the pre-reading size and the I / O scheduler type are obtained and automatically executed. And meanwhile, wear state parameters such as redistributed sector counting are collected in real time, and after fuzzy risk assessment, parameters of the medium / high wear risk hard disk are dynamically corrected and optimized. The limitation of traditional fixed threshold adjustment and optimization is broken through, self-adaptive matching of a load scene is achieved, the performance and the service life of the hard disk are balanced, whole-process automatic execution is achieved, and adjustment and optimization accuracy and stability are improved.

This invention provides a method and system for controlling construction waste crushing based on load feedback. The method includes: acquiring the stator current signal of the crusher's main shaft motor; extracting the active current component representing torque; constructing a time sliding window sequence; calculating the algebraic difference between the current active current component and a preset target value to obtain the original load deviation; constructing a weighted function to correct the load deviation amplitude; obtaining the load change rate by taking the first derivative with respect to time; building a fuzzy logic controller, using a Gaussian membership function in the fuzzification stage to establish a mapping between sample entropy and the width of the Gaussian distribution, and setting the membership function width in real time; adjusting the fuzzy subset of PID parameters according to fuzzy rules; using the peak-to-mean ratio of the sliding window sequence as an impact factor to weight and correct the centroid method defuzzification results; superimposing the adjustment increment onto the PID reference parameters; and combining the corrected load deviation to calculate the control command and adjust the output frequency of the feed motor inverter. This invention enables multi-dimensional adjustment of the crusher.

The application discloses a kind of based on fuzzy comprehensive evaluation's advantage seepage channel identification method, it is related to oil reservoir development technical field.The method includes: obtaining oil reservoir related data, screening evaluation index and constructing evaluation index system;Using analytic hierarchy process to determine the weight of each index, and obtaining weight vector;According to the distribution characteristics of each index, membership function is constructed;Membership degree is calculated to each index observation value, and fuzzy relation matrix is formed;The weight vector and fuzzy relation matrix are carried out fuzzy matrix multiplication, and the fuzzy comprehensive evaluation result vector is obtained;Whether there is advantage seepage channel in oil reservoir area is determined according to the comparison result of evaluation value and preset threshold value, and dynamic data and / or numerical simulation result is verified.The application can effectively handle the fuzziness and uncertainty of oil reservoir data, improve the accuracy and reliability of advantage seepage channel identification, and optimize oil reservoir development decision.

The invention relates to the technical field of rail transit vehicle control, in particular to a quantum fuzzy control method and system for a rail transit vehicle door controller, and the method comprises the steps: collecting the position, speed and other four-dimensional operation data of a door body, and carrying out the Kalman filtering and EMC compensation preprocessing; mapping and coding the four-dimensional variable quantum bits, and constructing a quantum superposition state containing disturbance compensation; a membership function is optimized based on quantum probability distribution, and a control instruction is generated through quantum annealing algorithm iterative reasoning; and the brushless motor is driven to execute and form closed-loop feedback. The system is developed based on an FPGA and comprises a data acquisition module, a preprocessing module, a quantum fuzzy calculation module, an instruction execution module and a closed-loop feedback module. According to the invention, the problems of poor dynamic load adaptation, weak anti-interference and the like of the existing gate controller can be solved, and accurate and stable control of the gate controller under dynamic load and strong electromagnetic interference scenes can be realized.

The invention relates to a substation equipment state evaluation system based on AHP-entropy weight fusion and improved fuzzy comprehensive evaluation. The substation equipment state evaluation system comprises a feature acquisition module, an adaptive scoring module, a membership calculation module, a weight calculation module, a fuzzy synthesis module and a classification optimization module. The invention further relates to a substation equipment state evaluation method based on AHP-entropy weight fusion and improved fuzzy comprehensive evaluation. The method comprises the steps that S1, a multi-strategy nonlinear scoring rule is established; s2, constructing a dynamic membership function; s3, carrying out an AHP-entropy weight fusion algorithm; s4, improving a fuzzy synthesis operator; and S5, integrating a classification optimization strategy. According to the method, the degree of fitting of the equipment degradation law is improved to 92% through nonlinear scoring, and the weight error is reduced by 35% through the AHP-entropy weight fusion algorithm; the evaluation system provided by the invention is tested on 5000 groups of simulation data, the recall rate of a'serious anomaly 'class reaches 93.5%, and meanwhile, the time consumed by single evaluation is as long as lt; 2ms.

The application relates to the technical field of machining, and discloses a tool intelligent selection method based on a fuzzy ontology, which comprises the following steps: receiving macro-geometric parameters of a target gear and extracting tool nominal parameters to construct an ontology instance; combining micro-feature classification to generate a boundary compensation coefficient and a compression rate factor, performing translation and contraction calculation on a basic trapezoidal membership function model to output a reconstructed trapezoidal membership function; adjusting a model for processing tool nominal parameters by using a degree fuzzy limiting operator to generate a multi-dimensional membership vector set; judging a matching state, generating rough machining and fine machining task individuals when there is a matching conflict, issuing a fuzzy modification instruction to reset the operator strength and performing secondary logical deduction; extracting a minimum comprehensive deviation distance individual after secondary logical deduction as a result output to generate process coordination data. The application decouples the physical performance contradiction under the multivariable constraint condition, avoids logical deadlock in single tool selection and matching, and improves the matching accuracy of a process planning sequence.

The application discloses a multi-objective vehicle path optimization method based on fuzzy time window satisfaction degree, which is suitable for a distribution scene containing a distribution center, vehicles and customer sites. The method first acquires customer coordinates, demand, time window and other data and calculates a distance matrix, generates a disturbed time window after disturbance on the basis of random disturbance of the deterministic time window, and represents the disturbed time window as a three-parameter fuzzy number [a, m, b]. The fuzzy arrival and service time are obtained through fuzzy operator recursion. Then, a quasi-trapezoidal membership function is constructed to calculate customer satisfaction and average satisfaction through barycenter and similarity. A heuristic and roulette are used to generate an initial population, and a distribution cost Cost and average satisfaction AvgMu are used as double targets. After non-dominated sorting, tournament selection, crossover and mutation evolution, the solution quality is optimized by combining local search, and the Pareto optimal solution set is output after the termination condition is met. The application integrates the uncertainty of the time window into optimization, balances the cost and satisfaction, and improves the path robustness.