49 results about "Global optimization problem" patented technology

A decision support system called Mounties that is designed for managing applications and resources using rule-based constraints in scalable mission-critical clustering environments. Mounties consists of four active service components: (1) a repository of resource proxy objects for modeling and manipulating the cluster configuration; (2) an event notification mechanism for monitoring and controlling interdependent and distributed resources; (3) a rule evaluation and decision processing mechanism; and (4) a global optimization service for providing decision making capabilities. The focus of this paper is on the design of the first three services that together connect and coordinate the distributed resources with the decision making component.

The invention discloses a method for matching and optimizing parameters of a mixed power locomotive with a fuel cell and a super capacitor. The method comprises: first, according to the index requirements of the power performance of the mixed power locomotive, determining the mixedness range of a system; then, building a multi-objective optimization function by using the power property of the whole locomotive, the cost of the whole locomotive, and the quality of a driving system under a certain working condition as optimization objects; solving the function by adopting a high-speed group intelligent optimization algorithm, and using the matching combination that the minimum value is used for the objective function as the optimum parameter matching result of the system of the locomotive; then, distributing load working conditions of various power sources by adopting a load distribution algorithm; building an objective function based on energy flow according to the distributed working conditions and the work efficiency of each part and based on the matched and optimized results of the parameters; optimizing the energy flow of the system; solving the global optimization problem with constraint conditions by adopting the high-speed group intelligent optimization algorithm. The method has the advantages that the consumption of hydrogen is reduced, the recycled service life of each power source is prolonged, and the performance of the whole locomotive is improved.

The invention discloses an image sequence category labeling method based on a mixed graph model. The method comprises a step of performing superpixel segmentation of an image sequence and characteristic description of superpixels; a step of performing nearest neighbor matching of inter-frame superpixels of a two-continuous-frame image; a step of using the mixed graph model to carry out global optimization modeling of the image sequence category labeling based on the spatial domain adjacency relation among superpixels of a single frame image and the time domain matching relation among superpiexels of a multi-frame image; and a step of using a linear method to solve a global optimization problem to obtain category labels of superpixles of a continuous multi-frame image. Compared with previous graph models, the mixed graph model created by the invention can describe the first-order and symmetric relation among superpixels in a single frame image and also the high-order and non-symmetric relation among superpixels in a two-continuous-frame image; the linear method is used for solution; and a category label which is better in consistency of time domain and higher in accuracy is effectively provided for each superpixel of an image sequence.

The invention discloses a multi-agent genetic clustering algorithm-based image segmentation method, which mainly solves the problems that the prior art is sensitive to an initial clustering center, has low convergence rate and is easily trapped in a local extremum. In the method, image clustering segmentation is converted into a global optimization problem. The method comprises the following steps of: firstly, extracting two-dimensional gray scale information of a neighborhood median and a neighborhood mean of pixel points of an image to be segmented to construct a new two-dimensional histogram; secondly, combining a multi-agent genetic algorithm (MAGA) with a fuzzy C-mean (FCM) clustering algorithm and obtaining an optimal clustering center and a membership degree matrix by using the global optimization capability of the MAGA; and finally, outputting clustering tags according to the maximum membership degree principle so as to realize image segmentation. The method has high anti-noise capability and high convergence rate, can improve the image segmentation quality and the stability of a segmentation result and can be used for extracting and identifying image targets.

The invention discloses a PHEV self-adaptive optimal energy management method based on path information, and the method comprises the steps: planning a driving path through a vehicle-mounted navigation system, and generating a prediction condition of a front path; establishing a travel mileage prediction strategy to predict the travel mileage of the user every day; generating a reference SOC basedon an SOC planning algorithm through the generated prediction data and the initial SOC; carrying out an APMP optimization algorithm, specifically, taking the minimum oil consumption as a global optimization target, introducing a collaborative state value, and converting a global optimization problem into a plurality of instantaneous optimization problems with Hammeton operators; optimizing the cooperative state initial value by adopting a genetic algorithm; solving an initial value of the cooperative state in the MAP by utilizing an interpolation method, and correcting the initial value of the cooperative state in real time according to the working condition information obtained by the vehicle navigation system and the reference SOC; and using a PMP optimization algorithm to carry out power distribution, transmitting the power to each execution component controller through a CAN bus, and completing whole vehicle control of the PHEV.

The invention provides a layout analysis method and system for automatically classifying test paper contents. The method comprises the following steps: acquiring an input document image; Extracting communicating parts of the document image to form an original communicating part set; Performing text and non-text classification on each communication component according to the communication components of the document image, and obtaining a first text communication component set and a non-text communication component set; Detecting and segmenting the character components of each communication component in the non-text communication component set to obtain the character components adhered to the communication components of the non-text classification, and adding the character components into the first text communication component set to obtain a second text communication component set; Classifying the printed characters and the handwritten characters for each communication component in thesecond text communication component set; And outputting a classification result of the document image content. By the adoption of the method, the classification problem of the elements is converted into a global optimization problem for solving the maximum joint probability of all the elements, and therefore the overall classification accuracy can be improved.

An approach for solving a global optimization problem is described. Specifically, one embodiment of the disclosure sets forth a method, which includes the steps of receiving a quantitative initial solution, generating a quantitative feasible solution, mapping the quantitative feasible solution to a qualitative feasible solution, determining whether to accept the qualitative feasible solution based on a first predetermined rule, wherein the qualitative feasible solution that is accepted is reverse mapped to the quantitative feasible solution, and transmitting a result of the determining step.

The invention provides a mobile sensing multi-user unloading optimization method based on mobile edge calculation based on limited calculation and wireless resources of a mobile edge calculation server, and finds an optimal unloading scheme to maximize the user utility in a system range by considering the mobility of users and task time delay. According to the method, a heuristic mobile perceptionsearch algorithm is provided to obtain an optimal unloading scheme, an original global optimization problem is converted into a plurality of local optimization problems, the local optimization problems are decomposed into sub-problems to be solved, and finally, the performance better than that of other technologies can be obtained under the condition of better calculation complexity.

The invention provides an orthogonal successive approximation method based on orthogonal experimental design and variable metric neighborhood search. The orthogonal successive approximation method comprises the following steps of setting an initialization parameter, and selecting a proper orthogonal table according to solved problem dimensions and discrete level numbers; conducting orthogonal experiments within a feasible region by beginning from an initial point x0, and calculating each experimental scheme through evaluation by adopting a penalty function method; selecting a new iteration point from the experimental schemes; if x1 is superior to x0, then allowing x0 to be equal to x1, and enlarging the step size in search to enhance global search at the same time; otherwise, shrinking search space to enhance local search; repeating the steps, and repeatedly iterating to successively approximate a global optimal solution until convergence conditions are met. The invention provides the orthogonal successive approximation method for solving a global optimization problem, has the advantages of simple principle, fewer calculating parameters, high convergence speed and the like and can be used for rapidly acquiring the optimal solution or the approximate solution of the global optimization problem.

The invention provides a variable cycle engine transition state optimization method based on a large-scale global optimization technology, and belongs to the technical field of aero-engine transitionstate control optimization. A transition state control optimization problem is converted into a large-scale global optimization problem; a competitive particle swarm algorithm is adopted to carry outoptimization search on a single-step increment sequence of a control input quantity, the defects that a traditional SQP algorithm depends on model characteristics, the later convergence speed is low and the traditional SQP algorithm is prone to falling into a local optimal solution are overcome, and certain rapidity and transportability are achieved.

The invention discloses a distribution calculation method for stable control on a global voltage of a power transmission / distribution grid. Aiming at a power transmission / distribution grid of a multi-stage scheduling center layered management structure, the method has a target of minimum total control cost of the power transmission / distribution grid, and boundary influence factors are introduced to decompose global optimization problems into sub-problems of stable control on voltages of a power transmission grid and different power distribution grids; by continuously switching voltages, equivalent value power and the boundary influence factors at boundary connecting points of the power transmission / distribution grid, distribution calculation of global control can be achieved; since the boundary influence factors need to be constructed by antithesis multipliers in different sub-problem optimization solutions, the sub-problems of stable control of the power transmission / distribution gridneed to be solved by using antithesis multiplier type optimization control algorithms. Equivalent value models of power transmission grids or power distribution grids do not need to be established, and stable voltage control distribution calculation of power transmission / distribution grids can be achieved by only exchanging a small amount of boundary node information.

The invention discloses an image segmentation method based on an organizational evolutionary cluster algorithm, for mainly solving the problems of sensitivity of an initial cluster center, slow convergence speed and proneness to falling into a local extreme value in the prior art. The method transforms image cluster segmentation into a global optimization issue. The realization steps comprises: first of all, combining an organizational evolutionary heredity algorithm (OEA) with a fuzzy C-means (FCM) cluster algorithm, at the same time, by use of pixel point space information, obtaining an optimal cluster center and a membership grade matrix through the global optimization capability of the OEA; and outputting a luster label according to a maximum membership grade principle so as to realize image segmentation. The advantages are as follows: the noise-immune capability is high, the convergence speed is high, the image segmentation quality and the stability of segmentation effects can be improved, and the method can be applied to the extraction and identification of an image target.

An approach for solving a global optimization problem is described. Specifically, one embodiment of the disclosure sets forth a method, which includes the steps of receiving a quantitative initial solution, generating a quantitative feasible solution, mapping the quantitative feasible solution to a qualitative feasible solution, determining whether to accept the qualitative feasible solution based on a first predetermined rule, wherein the qualitative feasible solution that is accepted is reverse mapped to the quantitative feasible solution, and transmitting a result of the determining step.

The invention relates to a grape wine classification method based on Bayesian optimization and an electronic nose, and the method comprises the following steps: S1, employing a LightGBM algorithm, employing a Leaf-wise tree building method, finding a leaf with the maximum splitting gain from all current leaves each time during tree building, then splitting, and repeating the above steps; the LightGBM uses the maximum tree depth to prune the tree, and excessive fitting is avoided; S2, building a Bayesian optimization algorithm; S3, building a BO-LightGBM, and performing self-optimization adjustment on hyper-parameters of the LightGBM by using a Bayesian hyper-parameter optimization algorithm; enabling bayesian optimization to use a probability model to replace a complex optimization function, introducing the prior of a to-be-optimized target into the probability model, thus the model can effectively reduce unnecessary sampling. The Bayesian optimization method has the advantages that the Bayesian optimization method determines the optimization method of the next evaluation point by constructing the probability model of the function to be optimized and utilizing the probability model, the most advanced result is achieved on some global optimization problems, and the Bayesian optimization method is a better solution for hyper-parameter optimization.

The invention discloses a sparse representation and parameterization-based curved surface fitting method. The method comprises the following steps of inputting a model; segmenting the model, calculating initial parameterization coordinates of a local curved surface segment, optimizing a linear combination coefficient of sparse representation, and performing parameter optimization to obtain combined function optimization; solving a global optimization problem; and obtaining a curved surface fitting result. According to the sparse representation and parameterization-based curved surface fitting method, a simple monomial function is used as a basis function, good approximation for different geometric characteristics is realized through introduction of parameter optimization, and the input model can be a three-dimensional mesh model or point cloud; and the combined function optimization model is obtained in combination with the optimization of the linear combination coefficient of sparse representation and the parameter optimization, and the solving is performed in a cyclical iteration manner by introducing an auxiliary quantity.

The invention discloses a vector sector positioning method and a local optimization model prediction control method and device. By means of establishing a detailed prediction model, the output currentis controlled, and the neutral point potential of the direct current side and the suspension capacitor voltage of each phase can be controlled. Meanwhile, in allusion to the problems of large calculation amount and difficult hardware implementation when global optimization is carried out for traditional model prediction control, the triangular region where the reference voltage is located in an output voltage vector diagram is judged in advance by utilizing the vector sector positioning method provided by the invention, cost function judgment is carried out in a target triangular region, anda global optimization problem is converted into a local optimization problem in the reference region. According to the method, on the premise that the control over multiple output variables is achieved, the calculation load of the system is greatly reduced, calculation resources are saved, delay of algorithm control is reduced, and a five-level inverter controlled through the method has a good output effect.

The invention provides a distributed transmission and distribution cooperative reactive power optimization method and system based on Thevenin equivalent parameter identification. A global optimization problem of a power transmission network and a power distribution network is decomposed into sub-problems that each master system and each slave system independently carry out reactive power optimization. For the power transmission network, each power distribution network connected with the power transmission network is simplified into a PQ load, each power distribution network only transmits a complex power value at a root node to the power transmission network, and the complex power is the sum of net load of the power distribution network and complex power loss of the power. For the distribution network, the power transmission network is equivalent to Thevenin equivalent potential and equivalent impedance. The relative independence of an original power transmission and distribution network system and data is maintained. The state of the whole power transmission network system is represented only through the two parameters of the equivalent potential and the equivalent impedance, theinformation transmitted to the distribution network is less, and the communication traffic is reduced.

The invention discloses a multi-dimensional continuous optimization variable global optimization method based on reinforcement learning, wherein the method comprises the steps of establishing a reinforcement learning environment; selecting a specified number of optimization variables in the specified optimization variable set by using a reinforcement learning method, and then performing optimization on values of the optimization variables by using a continuous optimization variable optimization algorithm in a sequence optimization strategy; and optimizing an overall process and a constraint introduction method. According to the method, for the global optimization problem of the multi-dimensional continuous optimization variables, the purpose of intelligent optimization is achieved; the limitation of a traditional global optimization method on the number of the optimization variables can be broken through; and wide application of the artificial intelligence technology in the aspect of optimization becomes possible. The method can be applied to industrial design, manufacturing and processing, control optimization, investment decision, system engineering and other occasions with large-scale design variables; benefited from the strong intelligent combination optimization capability of deep reinforcement learning, the method also has a good global optimization effect on a system with a complex coupling relationship among variables.

A system that solves a global inequality constrained optimization problem specified by a function ƒ and a set of inequality constraints pi(x)≦0(i=1, . . . , m), wherein ƒ and pi are scalar functions of a vector x=(x1, x2, x3, . . . xn). The system performs an interval inequality constrained global optimization process to compute guaranteed bounds on a globally minimum value of the function ƒ(x) subject to the set of inequality constraints. The system applies term consistency and box consistency to a set of relations associated with the global inequality constrained optimization problem over a subbox X, and excludes any portion of the subbox X that violates the set of relations. The system also performs an interval Newton step on the subbox X to produce a resulting subbox Y. The system integrates the sub-parts of the process with branch tests designed to increase the overall speed of the process.

The invention provides a joint resource allocation method and an area orchestrator, and belongs to the field of wireless communication. The method comprises the following steps: S1) customizing a network slice according to a service category; S2) generating a corresponding global optimization problem, and decomposing the global optimization problem into a plurality of sub-problems; S3) calculatingsub-problem results; S4) calculating a resource allocation result of the network slice, and updating an auxiliary variable and a dual variable; S5) judging whether the global optimization problem meets an iteration stopping criterion or not according to a preset rule, if so, turning to the step S6), and if not, repeatedly executing the step S3) and the step S5) according to the updated auxiliaryvariable and dual variable until the global optimization problem meets the iteration stopping criterion; and S6) performing resource allocation. Global optimization is realized through coordination ofthe area orchestrator, so that each type of service is ensured to have good delay performance, and the overall delay of the system can be minimized.

One embodiment of the present invention provides a system that solves an unconstrained interval global optimization problem specified by a function ƒ, wherein ƒ is a scalar function of a vector x=(x1, x2, x3, . . . xn). The system operates by receiving a representation of the function ƒ, and then performing an interval global optimization process to compute guaranteed bounds on a globally minimum value ƒ* of the function ƒ(x) and the location or locations x* of the global minimum. While performing the interval global optimization process, the system deletes all of part of a subbox X for which ƒ(x)>ƒ_bar, wherein ƒ_bar is the least upper bound on ƒ* that has been so far found. This is called the “ƒ_bar test”. The system applies term consistency to the ƒ_bar test over the subbox X to increase that portion of the subbox X that can be proved to violate the ƒ_bar test.

A method based on CTU level rate-distortion optimization for rate control in video coding which can effectively improve the perceptual rate-distortion performance and coding efficiency is provided. Firstly, a perceptual rate-distortion model is established using a divisive normalization framework, which characterizes the relationship between local visual quality and coding bits. Subsequently, the established perceptual rate-distortion model is applied to overall distortion optimization which is transformed into a global optimization problem and solved with convex optimization algorithms to obtain optimal CTU level coding bit allocation.

The invention discloses a cluster resource scheduling method and device, equipment and a storage medium, and relates to the technical field of data analysis. The method comprises the steps: updating historical data corresponding to each application in an application cluster, screening out candidate historical data matched with the current resource state of each application from the historical data, updating model parameters of a small disturbance linear statistical model of each application based on the candidate historical data; determining the deviation between the current resource utilization rate of each application and an expected value, and sorting each application according to the deviation of each application to obtain the capacity adjustment priority of the plurality of applications; selecting a target application of which the capacity is to be adjusted from the plurality of applications according to the capacity adjustment priority; and carrying out capacity adjustment on the target application, and returning to the operation of updating the historical data corresponding to each application except the target application in the application cluster. According to the scheme, the solving difficulty of the global optimization problem can be reduced, and the rationality, accuracy and scheduling effect of resource scheduling are improved.

The invention discloses an ultra-high density electrical method electrode planning method for eliminating polarization effect interference. In the method, the influence of the polarization effect in the ultra-high-density electrical acquisition process is considered. By optimizing the selection order of an ultra-high density electrical method power supply electrode and a measuring electrode, the interference of the polarization effect in the engineering is reduced, and the data acquisition quality is improved. In the method, the optimization process of electrode arrangement is transformed intoa global optimization problem, and individual diversity is maintained and the optimization quality of an algorithm is improved through global cross-update and two local mutation operations. The method can be optimized according to the electrode arrangement characteristics of the ultra-high density electrical method, is simple to implement, has convenience in calculation, and has high practicability in engineering.

The invention requests to protect a multi-user computing unloading resource optimization decision method based on D2D communication, and belongs to the technical field of mobile communication. The method comprises the following steps: 1, establishing a data communication model based on D2D communication; 2, respectively establishing calculation overhead models including time and energy overhead in the task transmission stage and the task execution stage; 3, establishing a global optimization problem for minimizing the total calculation overhead of all users of the whole system; 4, establishing a preference sequence of the bilateral users by taking the calculated total overhead as a sorting basis; 5, based on the established preference sequence, obtaining a resource optimization decision of multi-user D2D calculation unloading by using a stable matching algorithm. The calculation unloading method based on D2D communication is beneficial to reducing unloading time delay and energy overhead, a resource optimization decision of calculation unloading is obtained by using a stable matching algorithm, compared with a random matching method, the total calculation overhead of the system can be effectively reduced, and performance very close to the optimal exhaustion search method can be obtained with low calculation complexity.

The invention provides a membrane computing frame-based spectral clustering algorithm, which comprises the steps of constructing a similarity graph G; generating a laplacian matrix; figuring out feature vectors corresponding to the first k minimum feature values of the laplacian matrix, and constructing a feature vector space; clustering the feature vectors in the feature vector space by using themembrane clustering algorithm; realizing the membrane clustering purpose by adopting a tissue type P system, wherein the tissue type P system comprises q cells, each cell comprises m objects, and each cell is used for transferring an optimal object thereof to the environment by utilizing the transferring rule; updating the optimal object corresponding to the environment; adopting a PSO speed-displacement model as an evolutionary rule, and adopting the optimal object in the environment as an obtained optimal solution after the shutdown according to a preset shutdown condition. The membrane computing can be used for processing global optimization problems. Therefore, the membrane computing frame-based spectral clustering algorithm not only enriches the type of the clustering algorithm, butalso optimizes the effect of spectral clustering. The application field of membrane computing is expanded.

The invention provides a vehicle driving path and driving speed collaborative planning method and system and a storage medium. According to the method, a global optimization problem of a coupling path and speed is built based on driver demands, traffic network information and vehicle longitudinal dynamic characteristics; a multivariable collaborative optimization method for path, speed and power transmission system control based on a genetic algorithm is provided, so that an economical driving path and a reference vehicle speed can be provided for a driver; in addition, the running path and the reference vehicle speed are tracked through a hybrid power system real-time optimization control strategy combining a global power distribution strategy and rolling linear quadratic tracking control, and therefore economical self-adaptive cruise control is achieved. The method can effectively improve the fuel economy of the vehicle and reduce the transportation cost and pollution.

The present invention proposes a layout analysis method and system for automatic classification of test paper content. The method includes: obtaining an input document image; extracting connected components of the document image to form an original set of connected components; The connected components are classified into text and non-text, and the first set of text connected components and the set of non-text connected components are obtained; for each connected component in the set of non-text connected components, the text components are detected and segmented, and the connected components of the non-text classification are obtained. The text components in the connected components, and add this component to the first text connected component set to obtain the second text connected component set; for each connected component in the second text connected component set, carry out the classification of printed characters and handwritten characters ; Output the classification result of the document image content. By adopting the method of the invention, the classification problem of elements is transformed into a global optimization problem for solving the maximum joint probability of all elements, so that the overall classification accuracy rate can be improved.

The present invention relates to an underwater multi-target positioning method, a terminal device and a storage medium. The method includes: S1: constructing a dynamic factor graph for multi-target real-time cooperative positioning; S2: changing the dynamic factor graph over time Each node evolves and updates in real time; S3: Obtain the real-time position of each target according to the dynamic factor graph. The present invention can decompose complex global optimization problems into multiple simple local optimization problems for distributed solution, has the characteristics of low algorithm complexity, etc., and takes into account the distributed characteristics of the network, and the algorithm can be made according to the change of the network state Appropriate adjustment has obvious advantages in solving the problem of underwater multi-target real-time cooperative positioning.