244 results about "Random optimization" patented technology

The invention relates to a heterogeneous cloud wireless access network resource allocation method based on deep reinforcement learning, and belongs to the technical field of mobile communication. Themethod comprises the following steps: 1) taking queue stability as a constraint, combining congestion control, user association, subcarrier allocation and power allocation, and establishing a random optimization model for maximizing the total throughput of the network; 2) considering the complexity of the scheduling problem, the state space and the action space of the system are high-dimensional,and the DRL algorithm uses a neural network as a nonlinear approximation function to efficiently solve the problem of dimensionality disasters; and 3) aiming at the complexity and the dynamic variability of the wireless network environment, introducing a transfer learning algorithm, and utilizing the small sample learning characteristics of transfer learning to enable the DRL algorithm to obtain an optimal resource allocation strategy under the condition of a small number of samples. According to the method, the total throughput of the whole network can be maximized, and meanwhile, the requirement of service queue stability is met. And the method has a very high application value in a mobile communication system.

The present invention relates to the supply-demand side combination random scheduling technology, especially to an optimal scheduling method considering wind power output and demand response nondeterminacy. The method comprises the step 1, based on an extension probabilistic sequence operation theory, establishing a probability model representing excitation load response nondeterminacy and price load response nondeterminacy, considering influence of the demand response nondeterminacy on cost; the step 2, in response to wind power fluctuation and randomness as background, constructing the current optimal scheduling model of an electric power system considering the wind power output and demand response nondeterminacy at the same time based on a risk constraint and risk cost theory; and step 3, converting a random optimization problem as a linear convex optimization problem for solution through conversion of a target function and a constraint condition. The method provided by the invention performs modeling and analysis aiming at demand response nondeterminacy and performs fine description of the nondeterminacy in the demand response process so as to provide practical reference for the scheduling of the electric power system and improve the stability of the electric power system.

The invention discloses a cascade reservoir random optimization scheduling method based on deep Q learning. The method comprises the following steps: describing the reservoir diameter process of a reservoir; establishing a Markov decision process MDPS model; establishing a probability transfer matrix; establishing a cascade reservoir random optimization scheduling model; determining a constraint function of the model: introducing a deep neural network, extracting runoff state characteristics of the cascade reservoir, Meanwhile, realizing approximate representation and optimization of a targetvalue function of the scheduling model; applying reinforcement learning to reservoir random optimization scheduling; establishing a DQN model; and solving the cascade reservoir stochastic optimizationscheduling model by adopting a deep reinforcement learning algorithm. According to the cascade reservoir stochastic optimization scheduling method based on deep Q learning, cascade reservoir stochastic optimization scheduling is realized, so that the generator set is fully utilized in the scheduling period, the power demand and various constraint conditions are met, and the annual average power generation income is maximum.

The invention relates to a resource allocation method for a low-delay high-reliability service in mobile edge computing, and belongs to the technical field of mobile communication. According to the method, in a multi-MEC multi-user environment, a user task queue model and an MEC task queue model are described respectively, a theoretical model of mobile service provider network utility maximizationis established with the task queue overflow probability as the constraint, and joint allocation is performed on power resources, bandwidth resources and computing resources. Considering that a constraint condition in the optimization model comprises a limit constraint of a task queue overflow probability; a random optimization problem of time averaging is converted and decomposed into three sub-problems of single-time-slot solving through a Lyapunov optimization theory, including calculation resource allocation of users, bandwidth and power allocation and calculation resource allocation of MEC, and then the three sub-problems are solved respectively. According to the method, the average total income of the MSP is improved while the requirements of users for low time delay and high reliability are met.

Vehicle apparatus adjusts a vehicle powertrain of the vehicle in response to a speed setpoint. An optimizer selects a control policy to periodically generate speed adjustments for applying to the speed setpoint to operate at increased efficiency. The control policy is based on a value function providing an optimized solution for a cost model and a transition probability model. The transition probability model corresponds to a driving state defined according to a plurality of dimensions including a time-of-day dimension and a geographic region dimension. The transition probability model and the control policy have inputs based on road grade and speed. The optimizer collects road grade data during routine driving of the vehicle to construct a observed transition probability model and uses divergence between the observed transition probability model and a set of predetermined transition probability models to identify a control policy for use during the routine driving.

A method and system are disclosed for improving a policy for a stochastic control problem, the stochastic control problem being characterized by a set of actions, a set of states, a reward structure as a function of states and actions, and a plurality of decision epochs, the method comprising using a sampling device obtaining data representative of sample configurations of a Boltzmann machine, obtaining initialization data and an initial policy for the stochastic control problem; assigning data representative of an initial weight and a bias of respectively each coupler and each node and the transverse field strength of the Boltzmann machine to the sampling device; until a stopping criterion is met generating a present-epoch state-action pair, amending data representative of none or at least one coupler and at least one bias, performing a sampling corresponding to the present-epoch state-action pair to obtain first sampling empirical means, obtaining an approximation of a value of a Q-function at the present-epoch state-action, obtaining a future-epoch state-action pair, wherein the state is obtained through a stochastic state process, and further wherein the obtaining of the action comprises performing a stochastic optimization test on the plurality of all state-action pairs comprising the future-epoch state and any possible action to thereby provide the action at the future-epoch and update the policy for the future-epoch state; amending data representative of none or at least one coupler and at least one bias, performing a sampling corresponding to the future-epoch state-action pair, obtaining an approximation of a value of the Q-function at the future-epoch state-action, updating each weight and each bias and providing the policy when the stopping criterion is met.

The invention discloses a multi-objective optimization method of a wind-powered pumped storage system. Based on a known current wind power prediction value and a determined load, the method comprises the steps of adopting an autoregressive moving average model to perform scene sampling and establishing a group of initial scenes based on the probability distribution of wind power output power prediction and prediction errors; processing the initial scenes by adopting scene reduction and obtaining an optimal wind power scene tree; and building a mathematical optimization model with low carbon and low cost as the goal, taking an on/off status of a generating set and the active power output allocation of the generating set at every period of time as a decision variable, adopting a second-order stochastic optimization model to optimize the optimal wind power scene tree, and outputting optimal scheduling of the generating set. The multi-objective optimization method of the wind-powered pumped storage system provided by the invention combines a scene tree creation and reduction technology to obtain an optimal wind power scene, based on second-order stochastic optimization, with the minimum cost and low carbon as the goal, an epsilon-constraint enhanced multi-objective algorithm and mixed integer linear programming are adopted, so that finally optimal scheduling of the generating set can be obtained, and the reliability and applicability are good.

The invention discloses a virtual power plant stratified random optimized dispatching method. First, a virtual power plant two-layer coordinated optimization dispatching model which comprises an upper layer virtual power plant layer and a lower layer micro power grid layer is established, wherein micro power grid optimized dispatching models in the lower layer are chance constraint models, in the models, an empirical distribution function is used for describing the power-out probability distribution of an uncontrollable micro power supply in an independent state, and according to the uncontrollable micro power supply, a histogram is combined to select a Copula function to establish a joint probability distribution model. Then, a sampling average approximation method and a KKT optimality condition are used for converting the virtual power plant two-layer coordinated optimization dispatching model to a single-layer deterministic model to be solved, and optimal dispatching is carried out on a virtual power plant. According to the method, coordinated operation of a plurality of micro power grids in the virtual power plant can be considered at the same time, probability distribution and Copula correlation analysis can be used for fully considering the influence of uncontrollable micro power supply randomness and correlation on optimized dispatching, and virtual power plant coordinated random optimized dispatching can be achieved.

The invention discloses a low-frequency oscillation control parameter optimization method of a grid-connected double-fed fan with virtual inertia belonging to the technical field of power system stability analysis. The method is realized by a processing system comprising a sequentially connected information acquisition module, a robust random optimization parameter adjustment calculation module and a parameter adjustment result output module, comprising three steps of information acquisition, parameter adjustment calculation and output parameter adjustment result optimization, and obtaining aset of control parameter policy variables; Judging whether the doubly-fed fan is the cause of inducing low frequency oscillation can guide the optimization of control parameters. After the parameter optimization strategy is applied, the power angle response level of the fan is effectively improved, the fan no longer transmits oscillation energy to the power network, and the low-frequency oscillation of the fan-connected system is restrained.

The invention discloses a cognitive radio and edge calculation method based on an industrial wireless network. According to the method, a cognitive radio and edge calculation model architecture basedon the industrial wireless network is erected firstly. Secondly, system models are established, including a network model, a service model and a calculation model. Afterwards, distributed random optimization is described, including system state space, motion space and reward and optimization targets. Finally, optimal distribution motion minimizing the online calculation complexity is sought by useof dynamic random optimization and an augmented Markov decision process (MDP), wherein a state transition probability is calculated firstly, then a system reward is updated, and finally the augmentedMDP is reconstructed and joint resource management is performed. Through experimental tests, compared with a conventional method, the method can obtain higher system benefits, provide higher throughput and reduce system transmission delay under different spectrum arrival rates and numbers of CRECs.

Disclosed is an uncertainty optimization operating method for an alternating current and direct current microgrid comprising high-density intermittent energy source. According to the characteristics of the microgrid and on the basis of taking output uncertainty of the intermittent energy source into consideration, a wind energy and solar energy output fuzzy random model, a diesel generator fuel cost module and an energy storage cost module are established; by combination with the grid structure characteristics of the microgrid and the problem of high output fluctuation caused by access of a large amount of intermittent energy source, a fuzzy random optimization model capable of minimizing the comprehensive operating cost of the alternating current and direct current hybrid microgrid, and a real-time imbalance power adjusting model capable of minimizing adjusting expenditure are established; and the fuzzy random optimization model capable of minimizing the comprehensive operating cost of the alternating current and direct current hybrid microgrid is solved by a fuzzy random uncertainty alternating direction multiplier optimization algorithm so as to obtain an operating scheme of the alternating current and direct current hybrid microgrid. By virtue of the uncertainty optimization operating method, the accuracy of the dispatching plan of the microgrid comprising the high-density intermittent energy source can be effectively improved, imbalance power can be lowered, and imbalance power adjusting expenditure caused by day-head dispatching deviation can be reduced.

A method for hydraulic fracture imaging consistently integrates ground deformation data and microseismic data. A time sequence of hydraulic fracture images is computed from the combined data by first computing an orientation and location of the fracture. A size of the fracture (e.g., length, width, and opening) and detailed shape is then computed consistent with the computed orientation using a stochastic optimization procedure. This procedure may be performed by selecting a candidate model of the fracture consistent with the computed orientation, computing a predicted surface deformation from the candidate model (e.g., using elastic crack theory), computing a predicted seismic activity from the candidate model (e.g., using Dieterich seismicity rate theory), and comparing the predicted surface deformation and predicted seismic activity with the collected ground deformation data and earthquake locations derived from the collected seismic activity data.

The invention relates to the field of resource optimization in an Internet of Vehicles, in particular to an Internet-of-Vehicles resource optimization method based on a non-orthogonal multiple accesstechnology, which comprises the following steps of: when a vehicle task is processed in an NOMA-assisted vehicle edge computing system, taking minimization of total energy consumption of the vehicle edge computing system as a principle; determining an unloading and caching decision of the system, calculating and allocating caching resources, namely considering random flow arrival and queue stability of a vehicle user, and defining as a random optimization problem through joint optimization calculation of the unloading decision and the content caching decision and calculation and allocation ofthe caching resources; and utilizing a Lyapunov optimization theory to propose a dynamic joint calculation unloading, content caching and resource allocation algorithm for solving the problem, decoupling the algorithm into two independent sub-problems, and utilizing 0-1 integer programming and linear programming to solve the two sub-problems. According to the invention, the computing resources ofthe mobile edge computing server can be effectively processed, and the energy consumption of the system is reduced.

Systems and methods directed toward processing optimization problems using loss functions, wherein a loss function is decomposed into at least one stratum loss function, a loss is decreased for each stratum loss function to a predefined stratum loss threshold individually using gradient descent, and the overall loss is decreased to a predefined threshold for the loss function by appropriately ordering the processing of the strata and spending appropriate processing time in each stratum. Other embodiments and aspects are also described herein.

The invention relates to a resource allocation and unloading decision-making method based on multi-agent architecture reinforcement learning, and belongs to the technical field of mobile communication. According to the method, excitation constraints, energy constraints and network resource constraints are considered, wireless resource allocation, computing resource allocation and unloading decisions are jointly optimized, and a random optimization model for maximizing the QoE of a total user of a system is established and converted into an MDP problem. Secondly, according to the method, an original MDP problem is subjected to factorization, and a Markov game model is established; then, the method provides a centralized training and distributed execution mechanism based on an actor-evaluator algorithm. In the centralized training process, multiple agents obtain global information through cooperation, resource allocation and task unloading decision strategy optimization are achieved, andafter the training process is finished, all the agents independently conduct resource allocation and task unloading according to the current system state and strategy. According to the invention, theQoE of the user can be effectively improved, and the time delay and the energy consumption are reduced.

The invention discloses a method for establishing a comprehensive energy system random optimization model considering scene simulation. The method comprises the following steps: processing uncertaintyof wind-solar power supply output and thermoelectric load prediction by using a scene analysis technology in random optimization; carrying out latin hypercube sampling according to source load probability distribution to obtain various operation scenes in different time periods through simulation, then carrying out clustering reduction on the scenes through a Kmeans clustering algorithm, and therefore constructing a typical operation scene set used for system operation optimization. Based on operation scene set, the overall economy of system operation is taken as a target; meanwhile, the overall energy efficiency level and the new energy consumption capability of the system are considered, a distributed comprehensive energy system random optimization model is constructed, a random optimization problem is converted into a deterministic optimization problem in different operation scenes, operation strategies in different optimization periods are generated, the model complexity is simplified, and the economy and the safety stability of the system under the influence of uncertain factors are guaranteed.

The invention discloses a cascade reservoir ecological random optimization scheduling model and a solving method. The establishment of the optimization scheduling model comprises the following steps:firstly, establishing a cascade reservoir ecological random scheduling model target function with the minimum corrected annual flow deviation AAPFD index; secondly, describing constraint conditions ofthe cascade reservoir ecological random scheduling model, wherein the constraint conditions comprise a water level constraint, a discharge flow constraint, a reservoir capacity constraint and a guaranteed output constraint. The ecological random scheduling model is solved by utilizing an SARSA algorithm in reinforcement learning, so that the problem of dimensionality disasters caused by a randomdynamic programming algorithm can be solved. The method can improve the ecological demands on the premise of guaranteeing the output of the cascade reservoir hydropower station, and has higher guidingsignificance for the ecological utilization of reservoir water resources. Meanwhile, the proposed SARSA algorithm can well obtain the result of the cascade reservoir random ecological optimization scheduling model in a short time.

The invention discloses a pneumoelectric interconnection system optimization operation method considering wind power uncertainty. The method comprises the steps: building a pneumoelectric interconnection system random optimization operation model through employing the minimization of the total operation cost of a pneumoelectric interconnection system as a target and combining with a set related operation constraint; aiming at the nonlinear power flow equation constraint of a natural gas network, adopting a linearization method for processing; describing the uncertainty of wind power output byusing a scene analysis method, generating a large number of wind power output scenes through a Monte Carlo simulation method, then applying a scene reduction method based on synchronous back substitution elimination to reduce the scenes to obtain a small number of representative scenes, and then converting the random optimization problem of the pneumoelectric interconnection system considering thewind power uncertainty into a mixed integer linear programming problem of a plurality of deterministic scenes to be solved. Under the condition that the wind power output is uncertain, the economicalefficiency and safety of the operation of the pneumoelectric interconnection system and the absorption capacity of the wind power are effectively improved through optimal scheduling.

The invention discloses a garment deformation method based on input human body posture real-time generation. The scheme relates to main algorithms including a skeleton driven sensitivity analysis algorithm, a hybrid garment algorithm based on sensitivity and a random optimization scheme based on a greedy algorithm so as to construct a garment instance database. The algorithm flow is listed as follows: after one input posture is given, skin calculation and processing are performed on a skeleton according to the skeleton driven analysis method so that a human body skin model is obtained and subdivided into multiple single areas; garment samples similar to the posture of each area of the skin model are inquired in the constructed garment database; deformation processing is performed on each garment sample; then fitting is performed on the garment sample by using the hybrid garment algorithm so that an initialized garment deformation effect is obtained; and the penetration and friction effect processing is performed on the initialized garment deformation sample so that the final garment deformation effect is obtained. With application of the method, the garment deformation simulation process can be optimized and the lifelike garment deformation effect can be obtained.

The invention relates to a light storage charging tower random optimization scheduling method considering peak load shifting of a power distribution network. The method comprises the following steps of: 1, establishing a tower day-ahead multi-objective optimization scheduling model by taking the minimum daily operation cost of the tower and the peak clipping and valley filling of the power distribution network as objectives; 2, evaluating the influence of peak load shifting by adopting the minimum variance of the curve of the electricity purchase quantity of each time period of the tower, andconverting multi-objective optimization into a single-objective optimization problem through an economic conversion coefficient; 3, considering the influence of the output uncertainty of the photovoltaic unit in the tower on day-ahead scheduling, describing by adopting a representative scene, and constructing a random optimization scheduling model by taking the tower day-ahead optimization scheduling and representative scene adjustment as the sum of first-stage decision and second-stage decision, day-ahead scheduling cost and real-time adjustment expected operation cost as targets; and 4, testing the actual light storage charging tower, and solving the model. According to the method, the probability optimal scheduling strategy under the scene that the light storage charging tower participates in peak load shifting of the power distribution network can be effectively obtained, economic, efficient and reliable operation of the tower is guaranteed, and peak load shifting of the regional power distribution network are achieved.

The invention discloses a comprehensive energy system optimization method, which comprises the following steps of obtaining system parameters of a comprehensive energy system, and constructing a jointoptimization model based on two-stage random optimization by utilizing the system parameters, wherein the system parameters comprise equipment parameters and line parameters during planning and source load parameters during operation; deconstraining the constraint optimization problem in the joint optimization model to obtain a target joint optimization model of the unconstrained optimization problem; carrying out iterative solution on the target joint optimization model by utilizing an accelerated gradient descent algorithm to obtain system optimization parameters; and utilizing the system optimization parameters to optimize the planning and operation of the comprehensive energy system. According to the method, rapid, effective, real-time and dynamic planning optimization and operation optimization can be carried out on the comprehensive energy system. The invention further discloses a comprehensive energy system optimization device and equipment and a readable storage medium which have corresponding technical effects.

The invention discloses an industrial edge network system architecture, which comprises an application layer, a control layer and a field layer, and the application layer is divided into a plurality of industrial virtual fragmentation networks; the control layer mainly comprises a software defined network (SDN) controller; and the field layer comprises an edge layer and a field subnet. The invention also discloses an industrial edge network system resource scheduling method, which comprises the following steps of: dividing a virtual sub-network into different NOMA clusters, describing a relationship between an AoI over-limit probability and a queue overflow probability, converting an AoI constraint into a queue overflow probability constraint, establishing a theoretical model of system power consumption minimization, and performing joint allocation on the bandwidth resources, the power resources and the computing resources; and secondly, converting and decomposing a random optimizationproblem of time averaging into two sub-problems solved by a single time slot, and respectively solving the two sub-problems. According to the method, the industrial virtual fragmentation network andthe virtual sub-network are established, so that better resource allocation is facilitated, and full space-time monitoring and cooperative control of key parameters in the industrial process are realized.

The invention relates to an observability constraint-based random planet landing track optimizing method, and belongs to the technical field of the deep space probe navigation and guidance. According to the method, a deep space landing guidance control task based on the monocular vision navigation is used as the background, the dual problem between the effective control and the reliable estimation is considered, the system uncertainty is used as partial cost and is introduced into a quadratic performance index through an expansion-state space, so that a random optimization feedback control law can be provided by adopting a linear quadratic type control technique, an expansion-state space description model of a established probe system is substituted, the real-time optimization of a planet landing track is realized. The dynamic planning and the burden in calculation based on a searching method are avoided, the problem of observability shortage in the landing process is effectively overcome, and the navigation estimation performance of the system is improved, so that the overall performance during the planet probe navigation and guidance control is ensured, and the final target that a planet is reliably landed is achieved.

The invention discloses a multi-symbol joint channel estimation method in high speed mobile environment, wherein the method is based on an OFDM wireless communication system, and a channel estimation method based on pilot frequency is used. The multi-symbol joint channel estimation method in high speed mobile environment comprises the following steps of: 1) modelling for a time frequency double selective channel through a complex exponential-basis expansion model; 2) calculating the number J of joint estimation OFDM symbols, meeting that the channel corresponding to J continuous OFDM symbols has joint sparse property; 3) for the J continuous OFDM symbols, designing a sparse pilot frequency mode, and deriving a channel estimation model; 4) obtaining optimal pilot frequency position distribution through a discrete random optimization algorithm; 5) reconstructing a sparse coefficient through a block synchronization orthogonal matching pursuit (BSOMP); 6) recovering a channel tapping coefficient according to a coefficient reconstructed by a BSOMP algorithm; 7) performing piecewise linear smoothing for the channel tapping coefficient obtained through estimation. The multi-symbol joint channel estimation method in high speed mobile environment can effectively resist time frequency double selective fading, and can improve channel estimation accuracy and spectrum efficiency.

The invention relates to an unmanned aerial vehicle inspection path planning method based on random optimization. The method comprises the steps of S1 modeling all unmanned aerial vehicles, starting nodes of tasks, target nodes of the tasks and charging stations; S2 modeling a first constraint condition; S3 modeling a first optimization target; S4 randomly optimizing and solving the first optimization target to obtain a solution of an unmanned aerial vehicle target allocation task; S5 modeling the single unmanned aerial vehicle and the starting node, the target node and the charging station ofthe target task corresponding to the single unmanned aerial vehicle; S6 modeling a second constraint condition, wherein the second constraint condition comprises an environment constraint, a maximumflight distance constraint, a maximum turning angle constraint and a turning adjustment distance constraint, and the maximum flight distance constraint considers flight distance recovery of the unmanned aerial vehicle after the unmanned aerial vehicle is charged at the charging station; S7 modeling a second optimization target, wherein the second optimization target comprises a path voyage cost function; and S8 randomly optimizing and solving the second optimization target to obtain an optimal solution of unmanned aerial vehicle inspection path planning. According to the invention, more efficient task execution of the unmanned aerial vehicle is ensured.

Methods and systems for online domain adaptation for multi-object tracking. Video of an area of interest can be captured with an image-capturing unit. The video (e.g., video images) can be analyzed with a pre-trained object detector utilizing online domain adaptation including convex multi-task learning and an associated self-tuning stochastic optimization procedure to jointly adapt online all trackers associated with the pre-trained object detector and a pre-trained category-level model from the trackers in order to efficiently track a plurality of objects in the video captured by the image-capturing unit.

An integrated optimization and control technique performs process control and optimization using stochastic optimization similar to the manner in which biological immune systems work, and thus without the use of historical process models that must be created prior to placing the control and optimization routine in operation within a plant. In particular, an integrated optimization and control technique collects various indications of process control states during the on-line operation of the process, and stores these process control states within a memory. During steady-state operation of the process, the integrated optimization and control technique attempts to optimize the process operation by developing a series of sets of process control inputs to be provided to the process, wherein the series of process control inputs may be developed from the stored process control states using an objective function that defines a particular optimality criteria to be used in optimizing the operation of the process. Moreover, the integrated optimization and control technique may respond to a significant change in the current process state by determining a new set of process control inputs to be provide to the process based on one or more of the stored process control states. In this case, the optimization and control technique compares the disturbance inputs of the current process control operating condition, after one or more of these disturbance inputs experiences a significant change, to the disturbance inputs of at least some of the stored process control states to determine the one or more of the stored process control states that is/are closest to the new process operating condition. The integrated optimization and control technique then develops the new set of control inputs to be delivered to the process based on the control inputs associated with the one or more stored process control states determined to be closest to the new process operating condition.