47results about How to "Solving Multi-Objective Optimization Problems" patented technology

The invention provides a route planning method based on a multi-target glowworm swarm algorithm and belongs to the technical field of route planning. The method includes: modeling a route planning problem, initializing the multi-target glowworm swarm algorithm, updating the glowworm position, determining a non-inferior solution set, updating an external archived file, judging whether a preset maximum iteration number is achieved or not, and determining a Pareto optimal route. The basic glowworm swarm algorithm is improved based on the Pareto dominant conception, and global searching and parallel computing capacities of the glowworm swarm algorithm are well used. Multiple route performance indexes are considered simultaneously during planning, a group of Pareto optimal solution sets can be obtained by planning once, and high flexibility is achieved. Further, the route panning method is different from traditional route planning methods for a single target and route planning methods of using a weighing method to convert multiple targets into the single target, and can better meet practical requirements of route planning.

The invention belongs to the field of path planning technology and specifically relates to a three-dimensional multi-UAV coordinated path planning method based on SAS. The method comprises the following steps: carrying out modeling on a path planning environment; initializing multi-target SAS calculation parameters consisting of the length of a minimum path section, a maximum turning angle, a maximum angle of climb / glide, a minimum safe distance of UAVs and a minimum flight altitude of UAVs; initializing the positions of the UAVs, wherein each UAV represents a path; updating the positions of the UAVs; expanding a current node; determining whether a path section collides with other path sections; updating a node table of the path section; executing a step (8) if minimum path cost set in the step (2) is met, and otherwise, executing the step (3); determining a cooperated planned optimal path so as to complete path planning. The method overcomes multi-objective optimization problems, has versatility, provides a reasonable optimal solution for a decision maker and better accords with practical needs.

The invention discloses an optimal deployment method of a large-scale industrial wireless sensor network based on differential evolution algorithm, which ensures the system reliability through carrying out optimization deployment on nodes, and relates to two fields of industrial wireless sensor network and intelligent computation. The method comprises the following steps of: carrying out automatic coordination on spaces according to actual spaces of industrial sites, obstacles, wireless sensor power and accurate requirement; and using the total number of nodes and minimum load standard difference of cluster heads as targets, wherein a node deployment model is established for restriction conditions based on the redundancy requirement, '1' represents arrangement of the cluster heads corresponding to mesh points, and '0' represents no arrangement. The invention provides a new binary differential evolution algorithm for optimizing and solving the model. By using a new probability prediction operator, the population is updated by a generated binary variation individual. The method can ensure the system reliability, and can reduce the construction cost of the system at the same time, balances the system energy consumption and prolongs the network life cycle through the optimization deployment of the nodes.

The invention discloses a spectrum resource management method based on deep reinforcement learning, and mainly aims to solve the problem that incomplete channel state information cannot be effectivelyutilized for spectrum and power allocation and spectrum resource management multitarget optimization in the prior art. According to the implementation scheme, the method comprises the following steps: constructing an adaptive deep neutral network which takes channel gain and noise power as weight parameters by taking spectrum efficiency maximization as an optimization target; and initializing theweight parameters, observing user access information and interference information, calculating a loss function according to the energy efficiency and fairness of a communication network, updating thechannel gain and the noise power layer after layer along the gradient descent direction of the loss function, repeatedly training the adaptive deep neural network, and outputting an optimal spectrumresource management strategy when a training end condition is satisfied. Through adoption of the spectrum resource management method, the optimal spectrum resource management strategy can be obtainedon the basis of the incomplete channel state information, and the spectrum efficiency, energy efficiency and fairness of the communication network are improved effectively. The spectrum resource management method can be applied to spectrum and power allocation in wireless communication.

The invention discloses a method for scheduling multi-target testing task based on decomposition and optimal solution following strategies, which belongs to the field of scheduling parallel testing task. The method comprises the following steps: initializing parameter setting; decomposing a multi-target problem into a series of subproblems in an objective function space through the utilization of the method based on the decomposition, and updating and evolving the subproblems through exchanging informations with neighborhood areas; updating reference points and the neighborhood areas of every subproblem; adopting the optimal solution following strategy to allow a solution set to be improved as a whole; and obtaining a more optimal solution through repeated iterations, thereby obtaining a more optimal testing task sequence and a relevant testing scheme collection. The method solves the multi-target optimizing problem based on decomposition strategy, avoids the using of weighted sum method, reduces the effect of human factors, allows the quality of every generation of solution set to be improved as a whole with the adding of the optimal solution following strategy, and finally improves the efficiency of the method for scheduling the multi-target testing task.

The invention relates to a multi-target probabilistically optimal power flow fuzzy modelling and solving method for a power distribution network. The multi-target probabilistically optimal power flow fuzzy modelling and solving method comprises the following steps: obtaining n (n is less than t) random variables in t nodes in a network, and establishing a random variable model; obtaining 2<n> sample points corresponding to the n random variables and weighting coefficients omega<1> and k corresponding to the sample points by adopting a two-point estimation method; establishing a multi-target probabilistically optimal power flow nonlinear programming model; converting the multi-target probabilistically optimal power flow model into a single-target nonlinear programming model on the basis of a fuzzy set theory, and obtaining optimal power flow values Z<1> and k corresponding to the 2<n> sample points; and obtaining origin moments of various orders of an optimal power flow Z. According to the multi-target probabilistically optimal power flow fuzzy modelling and solving method provided by the invention, the random distributed power supply output problem is solved by adopting a point estimation method; therefore, the probabilistically optimal power flow problem is converted into deterministic optimal power flow calculation and characteristic value statistics; various samples are calculated by adopting the multi-target optimal power flow modelling and solving method respectively; and thus, the statistical characteristic value of a probabilistically optimal power flow can be finally obtained.

The invention discloses a guideboard identification confrontation defense method based on a genetic algorithm. The method comprises the following steps: (1) constructing a training set, a test set anda confrontation set; (2) constructing a guideboard classifier, and training the guideboard classifier by using the training set and the test set to obtain a guideboard recognition model; (3) constructing a guideboard attack model according to a genetic algorithm, and generating an adversarial sample by utilizing the guideboard attack model; (4) correcting the adversarial sample, applying the adversarial sample in a physical scene, and collecting a physical image of the adversarial sample; (5) inputting the physical image of the confrontation sample into a guideboard identification model, andscreening to obtain a high-quality confrontation sample; (6) adding high-quality confrontation samples into the training set, and training the guideboard recognition model again by using the trainingset to realize confrontation defense optimization of the guideboard recognition model; and (7) identifying the guideboard image by using the guideboard identification model after confrontation defenseoptimization to realize confrontation defense of guideboard identification.

The invention discloses a planning operation joint optimization method and system for an electricity-gas hybrid system based on distributed multiple scenes. The method comprises the steps: collectingelectricity purchasing power, gas source point gas inflow and t-moment net load fluctuation power parameters in different scenes; establishing an electricity-gas hybrid system capacity optimization problem model comprising a capacity configuration layer sub-problem model and a system operation layer sub-problem model; decoupling an electricity-gas hybrid system planning operation joint optimization problem into a configuration layer sub-problem and operation layer sub-problems in n scenes by adopting an alternating direction multiplier method with an adaptive step length, and carrying out iterative solution; and obtaining an optimal configuration result and an operation strategy set of the electricity-gas hybrid system. According to the method, indexes such as the net load fluctuation ratio are adopted to evaluate the operation flexibility of the power distribution network, the flexibility influence is considered in the optimization target, and the multi-target optimization problem ofconsidering both economy and flexibility in the capacity optimization of the electricity-gas hybrid system is effectively solved.

The invention relates to a nuclear main pump resin-based composite material tilting pad water lubrication thrust bearing optimized design method. The method is characterized in that the geometric dimension of the tilting pad and the thickness of a cylindrical tile surface serve as design variables, the bearing capacity and power loss of a bearing serve as optimization problems of an objective function, an initial sample point of the design variables is determined through sampling, and an elastohydrodynamic lubrication model simulating the coupling effect of interface slippage, water film cavitation and tile surface deformation is established; a Reynolds equation corrected by an ultimate shear stress model is converted into a linear complementary problem through a parameter quadratic programming algorithm, and the linear complementary problem and a solid elastic deformation equation are subjected to iterative solution to obtain the sample point of the design variables and a target function value data set corresponding to the sample point; a Kriging agent model algorithm is used, and accordingly to an optimization point adding rule criterion and iteration ending conditions, multi-objective optimization of the inner diameter, outer diameter and warp angle of a tilting pad slippage of the thrust bearing, the thickness of the tile surface and the like is finally achieved. The methodhas the advantage that high bearing capacity and the low power loss of the large low-elasticity-modulus material tilting pad thrust bearing are achieved.

The invention relates to a genetic algorithm based antenna recognition network end-to-end service quality guaranteeing method, which adopts the wireless recognition network technology, and mainly comprises quality of service (QoS) index, intelligent decision and network reconfiguration in perception service running process. The QoS index in the perception service running process is used for evaluating the current service running state and for providing decision foundation to the intelligent decision process; the intelligent decision process adopts adjustable parameters in a network protocol stack as input, the genetic algorithm capable of solving multipurpose optimization problem is adopted, and an adjustable parameter value which can optimize the multiple multi QoS indexes is searched ina value space of each adjustable parameter; and the action process adopts the result of the intelligent decision as input, and the network is reconfigured according to the solved adjustable parametervalue so as to guarantee the service end-to-end QoS and optimize the network end-to-end performance.

The invention discloses an underwater wireless sensor network topology control method based on a genetic algorithm. The method comprises the following steps: a) constructing an underwater network model; b) establishing member node and cluster head node energy models and energy parameters; c) calculating the optimal cluster head number; d) determining an optimal cluster head by adopting a genetic algorithm, including the substeps: d-1) performing chromosome coding; d-2) selecting an initial population; d-3) constructing a fitness function; d-4) constructing a selection operator, a crossover operator and a mutation operator; d-5) calculating an individual fitness value; d-6) selecting an optimal cluster head; e) self-clustering and carrying out data communication. According to the underwaterwireless sensor network topology control method, the genetic algorithm is used in the cluster head selection stage, the multi-objective optimization problem can be solved, the energy and distance factors can be effectively balanced, the energy consumption of each node can be effectively balanced, and the overall energy consumption of the network is reduced, so the lifetime of the whole network isprolonged.

The invention discloses a reusable spacecraft surface impact damage feature recognition method. The method comprises the steps of representing a thermal image sequence of pixel points collected by a thermal infrared imager by a three-dimensional matrix; finding a temperature peak value in the three-dimensional matrix M; performing temperature division on the row of the frame number where the temperature peak point is located; performing temperature division on the column of the frame number where the temperature peak point is located; extracting transient thermal response in a blocking and step-by-step manner; adopting a mean shift algorithm to automatically classify the extracted typical transient thermal responses; adopting a decomposition multi-objective optimization algorithm based ongradient search weight adjustment to select representations of each type of transient thermal response to form a matrix Y; converting the three-dimensional matrix into a two-dimensional matrix, and performing linear change on the two-dimensional matrix by utilizing the matrix Y to obtain a two-dimensional image matrix R (x, y); and performing feature extraction on the two-dimensional image matrixR (x, y) by using a one-dimensional Ostu segmentation algorithm. According to the method, through the uniformly distributed solution sets, the difference and similarity are comprehensively considered,and the defect feature extraction accuracy is improved.

The invention discloses an underwater wireless sensor network security routing control method based on a quantum genetic algorithm. The method comprises the following steps: a) establishing an underwater clustering network model; b) establishing member node and cluster head node energy models, energy parameters and trust models; c) calculating the optimal cluster head number; d) determining an optimal cluster head by adopting a quantum genetic algorithm, including: d-1) chromosome coding; d-2) selecting an initial population; d-3) carrying out population measurement; (d-4) making fitness evaluation; (d-5) updating the population; (d-6) selecting an optimal cluster head; and e) carrying out self-clustering, and establishing routing and data communication. The quantum genetic algorithm is used in the cluster head selection stage, the multi-objective optimization problem can be solved, the average energy consumption of global nodes can be effectively averaged, the overall service life ofthe network is prolonged, the adaptability is achieved, node malicious behaviors are evaluated through trust values, the safety of the network is guaranteed, and the wide application value is achieved.

The invention discloses an improved distributed multi-agnet system for microgrid multi-target energy management. The system comprises at least one microgrid unit, a main container, and at least one Agent container. The main container comprises a basic Agent, an aggregation electrical Agent, and an Agent platform. The basic Agent is used for establishing a decision, negotiation, and executing a microgrid solution. The running time of the basic Agent is defined as a decision period. The decision period is asynchronously started by an unpredictable event. The finish time of the decision period is adjusted dynamically. The basic Agent also comprises a distribution unit. The microgrid unit is divided according to functions and at least one type of basic Agent is paired by the distribution unit. Each Agent container comprises a unit electrical Agent and at least one basic Agent. The unit electrical Agent is bound to a mircogrid unit and sends corresponding electrical information to the aggregation electrical Agent. The basic Agents communicate with one another via an Agent protocol. The system establishes communication among the Agents, decides self behaviors by using self information and local environmental information in order to finally achieve an overall optimization control purpose of the mircogrid.

The invention provides a topological optimization method based on a Matlab and CFX combined solving platform. The topological optimization method mainly includes (1), calling a CFX solver to perform fluid simulation calculation automatically by the Matlab; (2), setting up a dynamic link library to enable an application program to call functions not belonging to executable codes, solving the problem that the Fortran program cannot be recognized by the CFX directly and ensuring that source item information can be transferred to the CFX smoothly; (3), by the Matlab, calling a CFX postprocessing module to export calculation results in a specific format and from a specific patch, reading the calculation results and performing follow-up calculation. The topological optimization method has the advantages that two common softwares, the Matlab and the CFX, are combined together, the Matlab is used for topological optimization algorithm control, the CFX is used for flow field solving, an interface is set up, and the combined solving platform is established, so that mutual calling and data transfer between the Matlab and the CFX are achieved; the advantages of the two softwares are utilized,and optimization solving time can be shortened remarkably while multi-objective optimization can be realized.

The invention relates to a transient tripping control method for an out-going wind power base containing different types of draught fans. The transient tripping control method comprises the following steps of calculating a total tripping quantity required by a power system with transient stability constraints under a fault and a tripping quantity of a thermal power generation unit when the power system is in an optimized wind-fire proportion tripping condition; carrying out parameter initialization, wherein initial parameters comprise a tripping quantity reduction value of a double fed induction generator (DFIG) unit, a tripping quantity increase value of a squirrel cage induction generator (SCIG) unit, an initial tripping quantity of the DFIG unit and an initial tripping quantity of the SCIG unit in the power system during adjustment of a tripping scheme; building a double-target tripping model with the purpose of making the rate between a transient energy tripping index of the power system and line power oscillation damping maximum; carrying out fuzzy calculation on the double-target tripping model, and adding a weight coefficient to obtain a double-target weighted fuzzy transient tripping model; and adjusting the weight coefficient, carrying out calculation on the double-target weighted fuzzy transient tripping model to obtain an allocation result of the tripping quantity of the DFIG unit and the tripping quantity of the SCIG. The transient tripping control method can be applied to transient tripping control of the wind power base.

The invention discloses a park electricity distribution and sale planning operation method considering wind, light and storage. The method comprises the following steps: S1, establishing an investmentstrategy independent optimization model in which park electricity distribution and sale participate; S2, converting the optimization model into a mixed integer second-order cone programming model according to second-order cone relaxation, a bigM method and an auxiliary variable adding method; S3, determining a topological line structure, a reference capacity value, a planned age limit, a system load growth coefficient, typical load daily load data and wind-solar output data of the park power distribution network system; and S4, according to the parameters determined in the S3 and the planningmodel in the S2, performing calculation by utilizing Matlab to obtain planning capacities of wind, light and storage. According to the planning operation method, wind power, photovoltaic and energy storage are comprehensively considered, asset operation optimization configuration is effectively achieved, the optimal equilibrium solution can be rapidly achieved, the stability of the solution is good, and powerful support is provided for improving the operation capacity of companies in the power industry.

The invention discloses a reliable vehicle-mounted edge calculation unloading method based on reinforcement learning, which is used in scenes such as limited coverage of road side units (RSUs) caused by urban obstacles or insufficient calculation capability of the road side units, and a vehicle can perform reliable calculation unloading by means of an unmanned aerial vehicle (UAV). According to the method, a multi-objective optimization problem of maximizing system utility and minimizing power consumption is converted into two sub-problems of a power distribution problem and a calculation unloading problem, and task distribution and calculation are completed by the unmanned aerial vehicle and the road side unit together. According to the method, the chance constraint is created for the transmission power, the chance constraint is converted by using a Chebyshev inequality, the minimum transmission power is derived, and the reliability of task transmission is ensured. According to the method, the calculation unloading efficiency is improved through the improved deep reinforcement learning model, the predicted target Q value in the deep reinforcement learning model is adjusted by using the result of the multi-target optimization problem, and offline training and online updating of the edge server are realized.

The invention discloses an overload real-time system non-preemptive scheduling method based on MaxSAT optimal solution, which comprises the following steps: (S1) determining scheduling problem and coding task attributes into MaxSAT hard clauses; (S2) encoding the scheduling target as a MaxSAT clause; (S3) connecting the MaxSAT hard clause obtained from (S1) and the MaxSAT clause obtained from (S2)by using conjunction operator to obtain the MaxSAT problem; (S4) calculating the optimal solution of the MaxSAT problem by the MaxSAT solver, and obtaining the optimal scheduling scheme. By adoptingthe scheme, the invention solves the problems of difficulty in solving the optimal scheme and low efficiency existing in the overload real-time system, and finally improves the efficiency of the target task scheduling method, so that the method has strong expansibility and execution efficiency, but also has high practical values and popularization values.

The invention discloses a multi-objective optimization configuration method and system for a wind-light-hydrogen storage system and a storage medium, and the method comprises the steps: constructing an objective function with the lowest total cost and the minimum system load power shortage rate, taking the power and capacity limitation of the system as constraint conditions, and taking the operation number of each device in the system as a decision variable; performing population initialization on the wind-light-hydrogen storage system model; performing non-dominated sorting on the initial population through a non-dominated genetic algorithm with an elitist strategy, and obtaining a first generation of offspring population through selection, crossover and mutation operations; from the second generation, combining the parent population with the child population, performing non-dominated sorting on the combined population, performing crowding degree calculation on individuals in each non-dominated layer, and selecting the individuals to form a new parent population according to a non-dominated relationship and the crowding degree of the individuals; performing selection, crossover and mutation operation on the new parent population; and outputting a capacity optimization configuration scheme when the maximum number of iterations is reached. According to the invention, the problem of optimal configuration in a complex scene can be solved.

The invention discloses a preemptive scheduling method of an overload real-time system based on a MaxSAT optimal solution. The method comprises the following steps: (S1) determining a scheduling problem and encoding a task attribute as a MaxSAT hard clause; (S2) encoding the scheduling target as a MaxSAT clause; (3) connecting the MaxSAT hard clause obtained from (S1) and the MaxSAT clause obtained from (S2) by using conjunction operator to obtain the MaxSAT problem; (S4) calculating optimal solution of the MaxSAT problem i by the MaxSAT solver, and the optimal scheduling scheme is obtained. The invention improves the efficiency of task preemptive scheduling in an overloaded real-time system through the scheme, so that the invention not only has strong expandability and execution efficiency, but also has high practical value and popularization value.

The invention discloses a multi-objective optimization method and system for dynamically using a neural network, and relates to the technical field of multi-objective optimization algorithms, and the method comprises the steps: obtaining a decision variable of a first-generation individual of a genetic algorithm, and putting the obtained decision variable into a simulator to obtain a corresponding performance index; training the neural network model; converting a performance index corresponding to a first generation of individuals of the genetic algorithm into a fitness function, obtaining a dominating relationship between a feasible solution set and a non-feasible solution set based on a rapid non-dominating sorting algorithm, and then selecting a next generation of parent individuals from non-dominating leading edges of the feasible solution set and the non-feasible solution set by using an accessibility algorithm; obtaining a decision variable of the current generation individual, and putting the decision variable into a simulator to obtain a performance index; the feasible parent individuals and the infeasible parent individuals of the next generation are selected through a reachability algorithm, and the number of the individuals generated by each operator of the next generation is dynamically allocated. According to the method, the problem of multi-objective optimization can be solved under the condition that physical simulation is relatively complex and computing resources are limited.

The invention discloses an online multi-target resource allocation method combining a bipartite graph matching method and a constraint solver. The online multi-target resource allocation method specifically comprises the following steps: reading a resource holder set and a resource requester set; enabling the newly arrived entity set and the overdue entity set to update the current available entity set holding resource entity set and the request resource entity set; if all newly arrived request resource entities can be directly allocated to nearby held resource entities by using a greedy method, and when the overdue entities do not need to be optimally allocated again at the moment, directly performing on-site allocation; if the number of the overdue entities exceeds a set threshold value, obtaining a maximum allocation score by adopting a bipartite graph matching method; and performing multi-objective optimization by using a constraint solver to obtain a distribution result.

The invention discloses an improved distributed multi-agent system that can be used for micro-grid multi-target energy management, including: at least one micro-grid unit; a main container, which includes a basic Agent, an aggregated electrical Agent and an Agent platform; the basic Agent is used for Making decisions, negotiating and executing microgrid unit solutions, its running time is defined as the decision cycle, which starts asynchronously with unpredictable events, and the completion time is dynamically adjusted; the basic Agent also includes an allocation unit, and the microgrid units are divided according to functions , and the allocation unit is paired with at least one type of basic Agent; at least one Agent container, each Agent container includes a unit electrical Agent and at least one basic Agent, the unit electrical Agent is bound to a microgrid unit, and sends the corresponding electrical information to Aggregate electrical Agent; each basic Agent communicates with Agent protocol. Establish communication among agents, make decisions on their own behavior based on their own information and local environmental information, and finally achieve the overall optimization control goal of the microgrid.

The invention discloses an NSGA (Non-dominated Sorting Genetic Algorithm) based pipeline multi-target layout optimization method. Three targets of the pipeline length, the number of bends and installability are subjected to modeling, individual fitness values are evaluated with a non-dominated sorting method, and a niche technology is introduced to obtain a uniformly distributed Pareto solution set; and the Pareto solution set of pipeline layout schemes is solved by applying an NSGA. The method is substantially different from a conventional pipeline single-target layout and linear weighting method; and with the method, the problem in pipeline multi-target optimization can be substantially solved, and a decision maker can select different Pareto solution layout schemes in the solution set according to engineering needs.

The invention discloses a hydrothermal power economic dispatch method and system based on a hybrid rice algorithm. The method includes: determining the power generation range of a hydropower plant and the power generation range of a thermal power plant; randomly selecting any value within the power generation range of a hydropower plant as a hydropower plant Any value within the range of the initial power generation and the power generation of the thermal power plant is taken as the initial power generation of the thermal power plant; the first objective function and the second objective function are constructed; the optimal value of the first objective function is solved to obtain the first optimal value; Solve the optimal value of the second objective function to obtain the second optimal value; determine the corresponding optimal power generation of thermal power plants and optimal power generation of hydropower plants; determine according to the optimal power generation of thermal power plants and optimal power generation of hydropower plants The optimal allocation ratio; determine the optimal scheduling scheme according to the optimal allocation ratio. The method or system of the present invention is used to perform multi-objective reactive power optimization based on the hybrid rice algorithm for hydrothermal power dispatching, thereby realizing optimal dispatching of hydrothermal power economy.

The invention discloses a method and system for selecting the capacity and installation position of a static var generator for a power distribution network, and the method comprises the steps: enabling the capacity Qi of reactive compensation equipment arranged at each node position of the power distribution network to form a chromosome gene, and enabling each chromosome to represent a scheme forthe site selection and capacity configuration of the reactive compensation equipment; forming a phenotype of the gene by adopting a plurality of target functions representing the operation effect andthe economic index of the reactive power compensation scheme, and taking the phenotype as a basis for gene selection; initializing a population and population algebra of the gene, and giving the maximum number of iterations; and if the current iteration is completed, outputting an optimization result, selecting a scheme with the minimum voltage square deviation, and determining the capacity and the installation position of the reactive power compensation equipment. According to the method and the system, the multi-objective optimization problem of reactive power compensation configuration canbe quickly and efficiently solved, and the method and the system have important engineering value and wide expansibility.

The invention relates to a nuclear main pump resin-based composite material tilting pad water lubrication thrust bearing optimized design method. The method is characterized in that the geometric dimension of the tilting pad and the thickness of a cylindrical tile surface serve as design variables, the bearing capacity and power loss of a bearing serve as optimization problems of an objective function, an initial sample point of the design variables is determined through sampling, and an elastohydrodynamic lubrication model simulating the coupling effect of interface slippage, water film cavitation and tile surface deformation is established; a Reynolds equation corrected by an ultimate shear stress model is converted into a linear complementary problem through a parameter quadratic programming algorithm, and the linear complementary problem and a solid elastic deformation equation are subjected to iterative solution to obtain the sample point of the design variables and a target function value data set corresponding to the sample point; a Kriging agent model algorithm is used, and accordingly to an optimization point adding rule criterion and iteration ending conditions, multi-objective optimization of the inner diameter, outer diameter and warp angle of a tilting pad slippage of the thrust bearing, the thickness of the tile surface and the like is finally achieved. The methodhas the advantage that high bearing capacity and the low power loss of the large low-elasticity-modulus material tilting pad thrust bearing are achieved.

The invention discloses a distributed photovoltaic power supply medium and long term locating and sizing method based on an NSGA-III algorithm. The method comprises the following steps: constructing a distributed photovoltaic site selection optimization model by taking probability voltage sensitivity and network loss as objective functions; and constructing a distributed photovoltaic site selection capacity optimization model by using a voltage out-of-limit risk, annual network loss and a photovoltaic installed capacity objective function. In order to solve the problem that an existing optimization algorithm is poor in effect when facing a multi-objective optimization problem, an NSGA-III algorithm is adopted to solve an optimization model, and a Pareto solution set of the multi-objective optimization problem is generated. Besides, in consideration of the influence of medium and long term load change on an optimization result, the method carries out rolling solution on the optimization model by taking a year as a unit, and provides a decision basis for the future photovoltaic installed capacity expansion of the region.