71results about How to "Narrow solution space" patented technology

The invention discloses a multi-imaging satellite emergency task dynamic scheduling method. In order to solve the technical problem, the multi-imaging satellite emergency task dynamic scheduling method is invented to realize a scheduling target on the premise of satisfying the limitation. According to the technical scheme, the multi-imaging satellite emergency task dynamic scheduling method comprises the steps of: synthesizing adjacent target unit tasks located in one field range of a satellite sensor based on task classification; if the allocation of the synthesized tasks is failed, taking out the unit task with the maximum priority in the synthesized tasks through a restoration way for synthesis to therefore decompose the task into a synthesized task and an independent task with more imaging chances; meanwhile, establishing an available time slot in a new task insertion process; moving tasks in a waiting queue backwards; and obtaining a time slot satisfying with the implementation of newly inserted task when the waiting tasks are executable, thereby further improving a scheduling efficiency. By the adoption of the multi-imaging satellite emergency task dynamic scheduling method, tasks can be synthesized, the performing sequence and the performing time of the tasks are adjusted dynamically, imaging time of the satellite is reduced, and accomplishment rate and schedulability of the tasks are improved.

The invention discloses a complex structure variation detecting method based on a hybrid strategy. The method comprises the steps of collecting and performing statistics on inserting fragment length distribution and chain direction information of a read pair in double-end sequencing, determining a read pair contrasting chain direction, an inserting fragment length and the read pair with an abnormal chromosome through two-end read segment contrasting; identifying breakpoints by means of a hybrid strategy of a double-end mapping method, a local assembling method and a split read segment method;wherein the breakpoints are coordinates on a pair of reference genome and are adjacent in a sample and are separated in the reference genome; updating breakpoint position information according to a contrasting result, and changing the breakpoint interval which records the structure variation breakpoint information to an accurate position; wherein the structure variation breakpoint information comprises a structure variation type, a breakpoint starting position and a support read segment number; and recording the contrast quality of the read pair and the read pair number of the support breakpoints, thereby finishing accurate structure variation identification. The complex structure variation detecting method improves variation detecting precision and supplies a checkout method for complex structure variations.

The invention belongs to the field of electric power load monitoring and especially relates to a DFHSMM-based non-intrusion type electric power load monitoring method and system. A DFHMM model and an HSMM model are combined to form a DFHSMM model, an electric power load and electrical equipment included therein are modeled by use of the DFHSMM model, a novel method is provided for solving the DFHSMM model, and the work state of the electrical equipment is identified finally. The method and system have the following advantages: a stable-state power feature and a work state duration feature of the electrical equipment can be taken into consideration at the same time, the method and system are applied to occasions where non-modeled electrical equipment exists, and the practicality and the robustness are quite high; constraint conditions are increased by use of load events, the solution space of a load monitoring optimization model is substantially reduced, and the solving speed of the model is improved; and the electrical equipment is classified according to the power feature, the load monitoring optimization model is solved for each electrical equipment cluster, and the overall solving speed is improved.

The invention discloses a multi-pose pedestrian image synthesis algorithm based on a generative adversarial network. The multi-pose pedestrian image synthesis algorithm comprises the following steps:S1, obtaining a training data set and a test data set from a pedestrian re-identification task data set Market-1501; S2, constructing a generative adversarial network model through the training data set according to a preset method; S3, adding an attitude information latent code into the generative adversarial network model input by adopting a preset method; S4, constructing an objective functionof a generative adversarial network model based on the attitude information latent code, and synthesizing a multi-attitude pedestrian image by using the generative adversarial network model with the objective function; and S5, performing experimental result analysis according to the synthesized multi-pose pedestrian image. The multi-pose pedestrian image synthesis algorithm has the beneficial effects that the solution space of the generator is effectively reduced, so that the generative adversarial network training is more stable, and high-quality multi-pose pedestrian pictures can be generated.

The invention relates to a method for selecting schemes for power restoration of an active power distribution network based on the improved genetic algorithm. According to the method, interconnection switches and various DG resources are overall considered, and a comprehensive decision is made on power restoration of the active power distribution network; DG equivalence is carried out on standby interconnection circuits in the decision making process, true DGs integrated into the power distribution network are classified, the DGs, under VF control, with the frequency and voltage regulation capacity are designed as main power supplies, and the randomly-uncontrollable power supplies are designed as common DGs; then generalized islanding and power restoration of the power distribution network are researched based on the isolated island operation mode where master-slave control is adopted. In the processing of solving the issue through the genetic algorithm, gene coding is fully combined with physical characteristics of the active power distribution network, so that the solution space needing to be searched for is greatly reduced; besides, a non-feasible solution will not be produced in the intersection and mutation operation on the basis, so that the computation speed is effectively increased, and the demand for real-time power restoration is completely met.

The present invention discloses a magnetic target location method based on a magnetic sensor array coordinate system. The method comprises the steps of: establishing a magnetic sensor array coordinatesystem according to the position coordinates of each magnetic sensor, and after a magnetic target enters the magnetic sensor array coordinate system, collecting the magnetic field value of each magnetic sensor; according to the magnetic field value of each magnetic sensor, obtaining the magnetic field total amount value of each magnetic sensor, rejecting the magnetic field value and the positioncoordinates of the magnetic sensor with the maximum magnetic field total amount value, and obtaining remaining data; and employing the magnetic field value of each magnetic sensor to establish a target function, substituting the remaining data into the target function, employing a particle swarm optimization and an LM to perform optimization of the target function, and obtaining a location resultof a magnetic target. The method is rapid and effective, and can meet the timeliness and accuracy requirements of magnetic target location.

The invention discloses a distribution network fault recovery method based on an A* algorithm and a fireworks algorithm. The method comprises the steps of (1) reading original data of a distribution network and simplifying the distribution network, (2) converting a contact switch into a virtual distribution power generation device DG and classifying the DG, (3) carrying out first stage algorithm,defining an initial node and a solution node, and using the A* algorithm to carry out power supply recovery on an important load, (4) carrying out second stage algorithm, reading distribution networkdata after first stage optimization, and using the fireworks algorithm to carry out power supply recovery on a rest load, and (5) integrating the optimal programs of the two stages, and determining afinal fault recovery program. According to the method, a staged method is adopted, a globally optimal solution is ensured, and the computing speed is fast.

In a network topology design apparatus for designing a network topology connecting a plurality of points by communication lines so that a predetermined traffic volume is transmitted between the plurality of points, chromosomes wherein the network topology is represented in a matrix are created, and a genetic algorithm is applied to evolve the chromosomes in order to improve cost performance. Then by applying a local search method to the chromosomes which have been evolved by the genetic algorithm, to search the chromosomes in the neighborhood of the evolved chromosomes, a chromosome with a better cost performance is determined. In this way, genetic algorithms and local search methods are applied to design a network topology with excellent cost performance.

A method of performing tomographic imaging of a sample in a charged-particle microscope is disclosed. The method comprises the following steps: providing a beam of charged particles that propagate along a particle-optical axis; disposing the sample on a sample holder that can be tilted relative to the beam; in an imaging step, directing the beam through the sample so as to form and capture an image of the sample at an image detector; repeating the procedure at each of a series of sample tilts so as to acquire a corresponding set of images; and in a reconstruction step, mathematically processing images from the set so as to construct a composite image of the sample, and therefore, in the imaging step, a sequence of component images is captured at a corresponding sequence of focus settings. In the reconstruction step, for at least one member of the series of sample tilts, a plurality of members of the sequence of component images are used in the mathematical image processing. Accordingly, a 3D imaging cube rather than a 2D imaging sheet at a given sample tilt is obtained.

The invention provides a skeleton animation data processing method and device and communication equipment. The method comprises the steps: obtaining a unit direction vector of each skeleton of a human body according to movement capture data; inputting the unit direction vector of each skeleton of the human body into a target neural network model, and obtaining a first global rotation amount of each joint of the human body output by the target neural network model; according to the first global rotation amount of each joint of the human body and the three-dimensional coordinates of the root joint of the human body, obtaining the position information of the root joint of the virtual character and the second global rotation amount of each joint of the virtual character; wherein the three-dimensional coordinates of the human body root joint are extracted from movement capture data; driving a virtual character skeleton to move according to the position information of the virtual character root joint and the second global rotation amount of each joint of the virtual character; according to the method, a target neural network model determined based on deep learning is utilized to realize mapping from three-dimensional coordinates of a human skeleton to a global rotation amount.

The invention discloses a parallel decision-making system and method for distributed data processing. The system comprises an initial logic node generation component, a logic node traversal component,a predetermined configuration cost calculation component and a parallel decision-making component. The initial logic node generation component receives task configuration data input by a user to generate an initial logic node topological graph for the distributed data processing system. The logical node traversal component traverses the initial logical node topological graph to obtain a predetermined configuration in the initial logical node topological graph. The predetermined configuration cost calculation component calculates the sum of the transmission cost and the calculation cost of each predetermined configuration. The predetermined configuration transformation component reduces the initial logic node and the connection edge according to the result of the predetermined configuration and merges the initial logic node and the connection edge, thereby obtaining a transformation result logic node topological graph after the initial logic node is reduced so as to reduce the solutionspace of the parallel decision. The parallel decision component obtains a parallel scheme with the total minimum cost through a local greedy strategy for the transformation result logic node topological graph.

The invention discloses a high-speed unmanned vehicle trajectory planning system and method based on multi-window sampling. According to the method, a multi-window model is used for representing a search space for motion planning of an unmanned vehicle, and a trajectory generation type model is constructed on the basis of human driving experience and a strategy of performing speed planning and path planning in sequence, so that generated candidate trajectories have relatively high dynamic obstacle processing capability and conforms to human driving habits; and finally, the trajectory with the minimum cost is selected as the optimal trajectory for planning by means of a cost function. The unmanned vehicle trajectory planning method is suitable for high-speed structured scenes mainly including dynamic obstacles, plays an important role in improving the traffic capacity of the unmanned vehicle in the actual environment, and can be widely applied to the fields of unmanned vehicle motion planning, advanced auxiliary driving technologies and the like.

The invention belongs to the technical field of power system power distribution network planning and operation, and particularly relates to a reconstruction method for a power distribution network with a distributed power supply after serious fault. The method for reconstructing the power distribution network with the distributed power supply after serious fault is suitable for rapid reconstruction of the power distribution network with the distributed power supply after the serious fault, and is convenient to model, high in simulation precision, high in calculation efficiency and strong in expansibility. The method comprises the following steps: 1, establishing an objective function and constraint condition model of a reconstruction problem after the power distribution network has the fault; 2, performing modeling analysis on the distributed power supply when power distribution network fault occurs; 3, adopting a spanning tree model based on the graph theory to solve an optimal methodof island division; 4, completing the weight dynamic division of a multi-objective function through a membership function, and normalizing a power distribution network reconstruction multi-objectiveoptimization problem into a single-objective optimization problem; and 5, solving the reconstruction optimization problem based on an empire competition algorithm to obtain an optimal reconstruction strategy after the serious fault of the power distribution network.

The invention discloses an order processing method and device, computer equipment and a storage medium, and belongs to the field of machine learning. The order processing method comprises the steps ofinputting feature data of at least one order in a target area into a recommendation model, and outputting a plurality of target transport capacity resources; according to the feature data, obtaininga plurality of predicted delivery paths for delivering the at least one order by the plurality of target transport capacity resources, and obtaining a plurality of target function increments corresponding to the plurality of predicted delivery paths; and processing the at least one order according to the plurality of target function increments. According to the order processing method, through therecommendation model, all transport capacity resources in the target area are screened for one round, so that the situation that all transport capacity resources in the target area are subjected to target function increment acquisition is avoided, and the calculation difficulty of finding an optimal solution is reduced, and the calculation amount and the processing duration of an order processingprocess are reduced, and an optimal scheduling effect is achieved.

The invention discloses a flow shop production scheduling method for distributed precast concrete components, which comprises the following steps: collecting production data, and establishing a mathematical programming model; a factory and a workpiece being subjected to integrated coding, and separating the factory from the factory by 0; generating an initial solution; performing local search on the initial solution, and updating a local optimal solution; judging whether a local search stop condition is met or not, and if so, turning to the next step; otherwise, carrying out the next local search; disturbing the local optimal solution generated after the completion to obtain a new solution; performing iterative local search on the generated new solution, judging whether conditions are metor not, and updating a local optimal solution; judging whether a termination condition is met or not, if so, entering the next step, and otherwise, carrying out the next iteration; and decoding the obtained optimal solution to obtain a scheduling scheme. The method is high in applicability and high in scheduling scheme quality, the scheduling scheme satisfying enterprises can be obtained within reasonable time, the inventory or delay punishment cost is reduced, and the production efficiency is improved.

For an easily implementable method for model predictive control of a DC/DC converter, and a corresponding controller, with which the optimization problem of the model predictive control can also be solved sufficiently quickly with large prediction horizons, the optimization problem is divided into two optimization problems by a model predictive output variable control and a model predictive choke current control being implemented in the control unit (10), wherein the strands of the multi-phase DC/DC converter (12) for the output variable control are combined into a single strand and a time-discrete state space model is produced therefrom and the output variable control predicts the input voltage (uv,k+1) of the next sampling step (k+1) for this single strand on the basis of a first cost function (Jv) of the optimization problem of the output variable control, said input voltage being given to the choke current control as a setpoint and the choke current control therefrom determining the necessary switch positions of the switches (S1, S2, S3, S4, S5, S6) of the strands of the multi-phase DC/DC converter (12) for the next sampling step (k+1) on the basis of a second cost function (Ji) of the optimization problem of the choke current control.

The invention provides a network slice optimization processing method and system, and the method comprises the steps: constructing a heterogeneous Markov decision process model, and constructing a hierarchical deep reinforcement learning architecture; constructing a joint optimization decision model based on a heterogeneous Markov decision process model and a hierarchical deep reinforcement learning architecture; wherein the joint optimization decision model comprises an upper-layer sub-model and a lower-layer sub-model, the upper-layer sub-model is used for executing an inter-slice resource allocation process, and the lower-layer sub-model is used for executing an intra-slice task scheduling process; and inputting network resource information and business request data obtained in real time into the joint optimization decision model to obtain a network slice optimal strategy solution output by the joint optimization decision model. The resources allocation and task scheduling of the network slice can be jointly optimized, so that the potential business capability of the system can be mined, the available resources of the system are fully mobilized to reduce the network delay of the user, and the user experience is improved.

The invention discloses a network life time prolonging method for an underwater wireless sensor network, which comprises the following: 1, a base station irregular periodic motion conversion method, which is to convert the irregular base station periodic moving path in a three-dimensional environment into the combination of the z-axis circular motions of base stations and ensures that the life time of the network does not reduce after conversion; 2, a base station circular motion plane low node eliminating principle, which is to ensure that nodes below a motion plane can always find a corresponding point with the largest load on the motion plane when the base stations perform the circular motion around the z axis; 3, a base station secondary large circular motion eliminating principle, which ensures that when the base stations move in the same plane, the radius is smaller when the life time of the network is shorter; and 4, a base station motion policy, base on which multi-operator spectrum server (MOSS) base stations perform periodic circular motion at the outmost periphery of a monitored area first downwards in turn from the largest circle to the second largest circle till the bottommost part and then upwards till the largest circle in a period.

The invention discloses a cold standby reliability redundancy distribution method and system based on a swarm intelligence algorithm, and aims to solve the problem of RRAP cold standby reliability redundancy distribution so as to optimize the system reliability of a cold standby redundancy system. The cold standby RRAP is also a model of mixed integer optimization programming; and the problem is difficult to solve by using a traditional method. According to the invention, in the cold standby redundancy system, a new solution structure is adopted; the number of variables is reduced by half; thepreprocessing of each variable boundary is found to reduce the solution space; a new SSO updating mechanism is adopted to be combined with elite selection to balance utilization and exploration and new local search so as to improve newly discovered gBest and solve the cold standby RRAP; and the optimal or high-quality solution of the RRAP of the cold standby redundancy system is obtained, the solution quality, the solution feasibility and unused resources of the final optimal solution are greatly reduced, the cold standby redundancy of the server and the like is greatly reduced, and the server operation efficiency of the cold standby redundancy system is improved.