162 results about "Decision problem" patented technology

The invention relates to a distributed semantic and sentence meaning characteristic fusion-based character relation extraction method, and belongs to the field of natural language processing. The method comprises the steps of firstly performing training in a small amount of marked corpora and a large amount of unmarked corpora by utilizing statistic word frequency features and a Bootstrapping algorithm to obtain a relational feature dictionary; secondly constructing a triple instance of a statement through an element distance optimization rule, and constructing a triple feature space by fusing distributed semantic information and semantic information; and finally performing true-false binary decision on a triple, and obtaining a character relation type by utilizing a confidence degree maximization rule. According to the method, automatic generation of the feature relation dictionary is realized; a conventional relational multi-class problem is converted into a triple true-false binary decision problem, so that a conventional machine learning classification algorithm is better adapted; and by utilizing the distributed semantic information, the accuracy of relational classification is improved.

A risk classification technique that exploits the existing risk structure of the decision problem in order to produce risk categorizations for new candidates is described. The technique makes use of a set of candidates for which risk categories have already been assigned (in the case of insurance underwriting, for example, this would pertain to the premium class assigned to an application). Using this set of labeled candidates, the technique produces two subsets for each risk category: the Pareto-best subset and the Pareto-worst subset by using Dominance. These two subsets can be seen as representing the least risky and the most risky candidates within a given risk category. If there are a sufficient number of candidates in these two subsets, then the candidates in these two subsets can be seen as samples from the two hypothetical risk surfaces in the feature space that bound the risk category from above and below respectively. A new candidate is assigned a risk category by verifying if the candidate lies within these two bounding risk surfaces.

The invention discloses a heterogeneous network multi-attribute decision-making method based on a network analytic hierarchy process. According to the method, when computing network attributes weight, attributive interaction and feedback in a dynamic network are considered, influence on network selection of a target network is also considered, physical truths for decision problems of a heterogeneous network are well fit, and a low-delay and low-quivering network can be selected under a real-time voice service. Concrete steps comprise that weight is firstly calculated; attributive factors affecting the network selection and the target network are divided into a functional group, a cost group and a scheme group; mutual relation between intra-groups and between-group elements and mutual relation between groups are set up; judgment matrixes of pairwise comparison according to the analytic hierarchy process are set up, feature vectors are obtained, submatrixes are formed, and all the submatrixes form a non-weighted hypermatrix, and an ultimate hypermatrix and ANP weight are obtained through weighting and exponentiation operating of the hypermatrix; and then the ANP weight and normalization network parameters obtain network power functional values, the network power functional values are sorted, and the largest utility value is selected to serve as the target network.

Provided is a method of batching and scheduling for steelmaking production with plant-wide process consideration, including the steps of: establishing a mathematical model for quantitatively describing the decision problem of batching on steelmaking and continuous casting procedures; starting from the production capacity balance between parallel equipment of the same procedure, and material flow convergence between upstream and downstream procedures, establishing a model for the assignment and sequencing of batches on continuous casting equipment and the time dimension; integrating the batching plan and the production scheduling scheme, issuing the batching plan and the production scheduling scheme integrated to all production and manufacturing units at the steelmaking stage. The present invention improves product quality, increases the material yield, resource utilization rate and equipment operation efficiency, realizes load balance on parallel equipment and smooth material linkage between serial equipment, and reduces the material flow transportation jam, downstream equipment waiting time and inventory.

A subjective trust quantification evaluation method based on the cloud model theory which is applied under a network environment in a network activity based on interpersonal relationship is mainly used for solving the trust decision problem of a trust body during the interpersonal interactive process in the network environment which aims at profit or is nonprofit. The technical points of the invention are that: 1. comprehensive quantification processing and evaluating are carried out according to the historical subjective credit rating data of a trust object and aiming at the nondeterminacy of subjective data; 2. the comprehensive quantification processing and evaluating of the credit rating change of the object are carried out according to the subjective credit rating changing data of the trust object; 3. the trust decision is accomplished based on the quantification data obtained in the steps of 1 and 2. The subjective trust quantification evaluation method of thte invention can effectively solve the nondeterminacy problem of the subjective trust quantification evaluation and upgrades the successful rate of online trading or cooperation and the satisfaction of the trust body.

The invention discloses a deep strategy learning method facing complex missions in a large-scale environment. The deep strategy learning method uses a deep nerve network to describe a state variable sensed by an intelligent agent, constructs a strategy model having a deep recursion type model, uses a strategy searching study algorithm to search an optimal parameter, and trains a nerve network until convergence. The deep strategy learning method uses a highly abstract and distributed expression capability of the deep nerve network to express the state variable sensed by the intelligent agent and constructs a strategy model having the deep recursion type structure, and is a first complete and intensive learning scheme which can systematically solve a complex decision problem in the large-scale environment.

The invention provides a self-organizing cloud architecture and optimization method and system for edge computing. The method comprises: a model construction step: performing model construction on a hierarchical unmanned aerial vehicle system; a decentralized computing unloading algorithm step suitable for an infinite virtual machine resource situation: respectively performing game decision according to conditions met by an unmanned aerial vehicle terminal bandwidth of the unmanned aerial vehicle system so as to achieve Nash equilibrium; and a decentralized computing unloading algorithm step suitable for a finite virtual machine resource situation: designing a decentralized computing unloading algorithm to perform algorithm computing. The invention provides a layered decentralized unloading method to reduce the communication overhead while maintaining the energy efficiency; and according to the self-organizing cloud architecture and optimization method and system provided by the invention, the energy-saving computing unloading decision problem in the hierarchical unmanned aerial vehicle system is modeled as a non-cooperative game for research, the decentralized computing unloadingalgorithm is designed for the situations of infinite virtual machine resources and finite virtual machine resources, and it is proved that the algorithm can achieve the Nash equilibrium.

The invention discloses an automatic driving decision-making method and system based on partial observable migration reinforcement learning, and employs a scheme reuse method related to a scene, and achieves auxiliary solving of a driving problem under a strange road condition through a conventional scheme in a migration driving scheme database. In order to achieve good riding experience, reinforcement learning is used for solving a decision problem in the field of automatic driving. The system comprises a scene unit, a sensing unit, a decision-making unit, an action planning unit and a control unit. A new environment model is added to a virtual environment database to cope with increasingly complex driving scenarios; a convolution layer is added to the neural network to identify obstaclesaround the vehicle; the important historical information is memorized by adding a long-short-term memory unit into the neural network; a Q value is estimated more accurately by using a weighted depthdouble-Q network algorithm based on Boltzmann soft maximization; the probability that each driving scheme is selected is solved by using a maximum entropy Mellowmax algorithm.

In one embodiment, systems and methods are disclosed for a planning-driven framework for an autonomous driving vehicle (ADV) driving decision system. Driving decisions are classified into at least seven categories, including: conservative decision, aggressive decision, conservative parameters, aggressive parameters, early decision, late decision, and non-decision problem. Using the outputs of an ADV decision planning module, an ADV driving decision problem is identified, categorized, and diagnosed. A local driving decision improvement can be determined and executed in a short time frame on theADV. For a long term solution, if needed, the driving decision problem can be uploaded to an analytics server. The driving decision problems from a large plurality of ADVs can be aggregated and analyzed for improving the ADV decisions system for all ADVs.

A system and method for solving a decision problem having Boolean combinations of linear and non-linear operations includes translating the non-linear real operations using a COordinate Rotation DIgital Computer (CORDIC) method programmed on a computer device into linear operations maintaining a given accuracy. Linear and translated linear operations are combined into a formula. Satisfiablity of the formula is solved using a decision procedure for Boolean combinations of linear operations over integers and reals.

The present invention provides a resource receiving control method which is used in communication field. The method includes: calling region module based on operation strategy decision function processes resource strategy decision according with received operation request message and user subscription message, and sends resource request message to visiting region module based on operation strategy decision function; said visiting region module based on operation strategy decision function processes authentication check for said resource request message, and sends resource request answering message to said calling region module based on operation strategy decision function; said resource request answering message includes service quality message modified by visiting region module based on operation strategy decision function, or feedback acceptable service quality message in rejection; said calling region module based on operation strategy decision function makes ultimate decision; the invention also provides a system used in said resource receiving control method and solves cross-realm QoS strategy decision problem.

A risk classification technique that exploits the existing risk structure of the decision problem in order to produce risk categorizations for new candidates is described. The technique makes use of a set of candidates for which risk categories have already been assigned (in the case of insurance underwriting, for example, this would pertain to the premium class assigned to an application). Using this set of labeled candidates, the technique produces two subsets for each risk category: the Pareto-best subset and the Pareto-worst subset by using Dominance. These two subsets can be seen as representing the least risky and the most risky candidates within a given risk category. If there are a sufficient number of candidates in these two subsets, then the candidates in these two subsets can be seen as samples from the two hypothetical risk surfaces in the feature space that bound the risk category from above and below respectively. A new candidate is assigned a risk category by verifying if the candidate lies within these two bounding risk surfaces.

An outlier detector that exploits the existing risk structure of the decision problem in order to discover risk assignments that are globally inconsistent is described. The technique works on a set of candidates for which risk categories have already been assigned. In the case of insurance underwriting, the invention pertains to the premium class assigned to an application. For this set of labeled candidates, the system finds all such pairs of applications belonging to different risk categories, which violate the principle of dominance. The invention matches the risk ordering of the applications with the ordering imposed by dominance and uses any mismatch during the process to identify applications that were potentially assigned incorrect risk categories.

The invention relates to a generator unit start-stop configuration method and system based on a depth deterministic policy algorithm. The method constructs a unit commitment start-stop allocation problem into an enhanced learning sequence decision problem. The method comprises the steps that 1) the output value vector of each unit at a previous historical moment is used as an intelligent body observation state and the input vector of a depth policy network; and 2) for a continuous time sequence unit start-stop optimization problem, the depth deterministic policy gradient algorithm is used to provide an optimization decision sequence and output the start-stop vector of the unit at the moment to realize the allocation of the start-stop status of the unit. Compared with the prior art, the method and system, which are provided by the invention, have the advantages that the depth deterministic policy learning method is used to directly acquire start-stop vectors at different moments; the method and system can adapt to the automatic allocation of the start-stop tables of unit clusters of different scales; and the system has better expandability.

The invention discloses an optimized operation method of an optical storage system considering the auxiliary service income of voltage regulation, which comprises the following steps: S1, a topological structure of the optical storage system and a flexible grid-connected operation mode are given; S2 integrates the power-on-grid revenue, auxiliary service revenue, power purchase cost, loss cost andoperation constraints, and constructs an optimal operation model aiming at maximizing the economic revenue of optical storage system; S3 To solve the continuous multi-stage global active and reactivepower bi-decision problem of optical storage system, a multi-dimensional dynamic programming algorithm is used to solve the optimal operation model of optical storage system. Compared with the traditional method, the optical storage system can actively respond to the voltage regulation auxiliary service price incentive, thus obtaining higher economic benefit, and effectively solving the problem that the voltage of the distribution network exceeds the limit.

The invention discloses a method for jointly determining resource allocation and offload ratios, which is applied to the technical field of in-vehicle wireless communications and used to solve the decision problem of communication resources, computing resource allocation and offload ratios faced by the macro base station for the VR service requested by vehicles in the wireless VR service scenariobased on vehicle networking edge calculation in the prior art. The invention firstly allocates a communication sub-channel for each vehicle, determines the optimal offload ratio and the optimal computing resource allocation at this moment according to the communication resource allocation, then compares with the even allocation of computing resources, further modifies the current computing resource allocation and offload ratio to make the task completion time always kept minimum, finally traverses the vehicle task completion time, if there is a channel that may be allocated, reallocates a newcommunication sub-channel for the vehicle with the longest completion time, and updates the computing resource allocation and offload ratio. The method provided by the invention may complete the VR service requested by vehicles in a short time.

The invention discloses a DDPG-based optimization method with distributed energy participating in distribution network voltage regulation. A deterministic depth deterministic strategy gradient algorithm is applied to distributed energy participating in distribution network voltage regulation. As the adopted DDPG algorithm is a model-friendly algorithm, the original distributed energy participatingin distribution network voltage regulation optimization strategy applies a Markov decision process method to be converted to a strategic decision problem, and the versatility of the algorithm is improved. The sample training stability is improved through applying a target network, easier convergence is achieved, and the method has the advantages of higher feasibility and lower operating cost.

The invention discloses an optimal maintenance decision method of power transmission and transformation equipment based on a Markov decision process. The method comprises that a state transfer relation diagram, of transfer relation among different states, of the power transmission and transformation equipment is established; a state maintenance model of the power transmission and transformation equipment is established according to the state transfer relation diagram of the power transmission and transformation equipment; the Markov decision process is used to solve the stable-state probabilities of different states of the power transmission and transformation equipment; a function relation between a maintenance strategy and pay corresponding to the maintenance strategy is established; a Markov based maintenance decision model of the power transmission and transformation equipment is established by taking maximization of certain function value of a pay sequence in the maintenance strategy as a sequence decision problem; and according to the stable-state probabilities of the different states of the power transmission and transformation equipment, a strategy iteration method is used to obtain an optimal maintenance strategy by solving. The method has the advantages that the Markov decision is used to make compromise between the maintenance cost and fault loss, the optimal maintenance decision is obtained, and reference is provided for maintenance deciding staff.

The present invention discloses a probability type asymmetrical encryption method based on public key certificate on the elliptic curve. It is a public key cryptographic system by proceeding from Diffie-Hellman decision problem on the elliptic curve and utilizing anti-collision miscollaneous function and public key certificate to make encryption and decryption. The system is made up on the elliptic curve, so that its calculation speed is quick, safety is high, it can use smaller key length to ensure higher safety. It can be extensively used in the fields of file encryption and computer network safety technology.

The invention discloses a gear hobbing method which is characterized in that during hobbing, optimization decision on gear hobbing process parameters comprises the following specific steps: (1) building a gear hobbing process ontology base; (2) expressing a gear hobbing process parameter decision target space; and (3) performing comprehensive optimization decision on the gear hobbing process parameters. The method disclosed by the invention has the advantages that by using the gear hobbing process domain ontology base stored in a database, the method can realize sharing and reuse of the knowledge about the domain of the gear hobbing, and not only can consider the gear hobbing process parameters as a system but also can solve new optimization decision problems of the gear hobbing process parameters by combining case-based reasoning with analytic hierarchy process during use.

The invention discloses a feedback type pond circulating water intelligent feeding system fusing machine vision and infrared detection technologies. The system has the realization principle that according to the ingestive behaviors of a fish school, a machine vision technology and an infrared sensing technology are cooperatively utilized, and a feeding decision is provided for the feeding system;a pond circulating water culture mode uses outdoor environment; due to natural environment change such as illumination, the visibility is low, so that fish school ingestive behavior feedback signals provided by machine vision are unstable, or due to low-temperature environment, the ingestive intensity is low, so that fish school ingestive behavior feedback signals provided by an infrared sensor are low; two kinds of decision modes are fused, so that the intelligent feeding decision problems under the pond circulating water culture mode can be better solved. The system has the advantages that the operation is simple and convenient; the intelligent level is high; the pond circulating water intelligent feeding can be effectively realized.

The invention discloses an ant colony algorithm-based firepower distribution method. The method comprises the steps of firstly building an air combat threat degree model and a firepower distribution decision model according to a fight situation of both sides; and secondly, in the aspect of model solving, performing algorithm improvement for deficiencies of a typical ant colony algorithm: improving the ant colony algorithm in combination with thoughts of typical ant colony system and max-min ant colony system algorithms, so that the improved ant colony algorithm is more reasonable in early evolution trend and higher in convergence speed, and can be better prevented from falling into local optimum. The improved ant colony algorithm proposed for firepower distribution not only can be used for firepower distribution of an air combat but also can be expected to be used for other combination optimization problems such as decision problems of firepower distribution and the like in an attack battle of ground tank groups and a maritime warship formation combat.

A method for generating a list of cases that are ranked by desirability (i.e. an ordinal ranking) based on attributes defined by a decision maker, and providing tools to assist the decision maker in statistically analyzing both the data inputted into a decision problem, the outcomes of the ranking process and the quality and consistency of inputs provided by raters.

The invention provides a multi-access edge computing task unloading method based on a boxing problem. According to the method, a user terminal and edge servers are regarded as task containers, and tasks are regarded as articles, so that a task unloading decision problem in edge calculation is converted into a boxing problem, the number of the edge servers started in a network is minimized througha heuristic method, and a task unloading decision is solved. The method comprises the following steps that firstly, calculating the loading capacity of each edge server and the ratio of the input datasize of each terminal task to resources needing to be calculated; forming two queues according to the capacity loading capacity and the task ratio from large to small; and finally, sequentially taking out the tasks in the task queue, configuring the tasks to an edge server with the maximum capacity and residual computing resources in the container queue, and repeating the operation until the taskqueue is empty. The method can be suitable for the task processing process of a multi-terminal and multi-task multi-access edge computing network, an appropriate task unloading scheme can be clearlyformulated, the computing energy consumption is minimized while the task time delay requirement is met, and the cost is saved.

The invention provides a decision support system for managing ecological construction in a county on one hand. The decision support system comprises a human-machine interactive interface module and a background control module, wherein the background control module comprises a decision module and a knowledge base module; the decision module comprises a synthesis body module, a simulation body module and a evaluation body module which are mutual coupling operation; the synthesis body module is used for allowing a user to select a decision problem, input a strategy variable value and query and store feedback contents and further driving the operation process of a model group of the simulation body; the simulation body module is used for carrying out analogue computation on parameters required for the decision problem and outputting parameter result to an evaluation body module; the evaluation body module is used for screening an evaluation result and an optimal strategy and feeding the evaluation result and the optimal strategy to the synthesis body module; and the evaluation result and the optimal strategy are output by the human-machine interactive interface module. The decision support system provided by the invention comprises a mathematical statistic module and a space analysis module as basic tool modules; and the invention provides a series of solutions for evaluating, simulating and predicting biological problems.

The invention provides a method for intelligently analyzing decision problems, relating to a method for analyzing decision problems expressed in Chinese natural languages. The method has the following beneficial effects: by adopting the method, various management decision problems which are input into a computer and expressed in Chinese natural languages can be analyzed and understood, the methods for solving the problems are found out and the computer is utilized to solve the problems so that the decision makers make decision schemes according to the results, thus improving the degree of automation and efficiency of decision problem analysis.

The invention discloses a load recovery two-stage optimization method considering wind power connection and used for researching a load recovery decision problem with wind power participation. Firstly, a single time-stepping load recovery problem is subjected to two-stage decoupling, wherein in the first stage, by taking the sum of a weighting load and a wind power output recovery value as the maximum target, a load recovery 0-1 planning model is established; next, in the second stage, by taking the shortest recovery time consumption as the target and by taking the decision variable obtained from the first stage as the initial condition, an alternating current power flow-based nonlinear planning model is established; and finally, each time stepping is solved in sequence to obtain a recovery decision scheme in the full process. By virtue of different time stepping decision idea, an effective recovery decision scheme can be obtained, wind power sequential access is realized and load recovery time consumption is shortened, so that the technical problems in load recovery decision in conditions of relatively large system scale and wind power system participation can be solved.

The invention relates to the computer and communication technical field, in particular to a decision method and a decision system for electronic equipment service condition. Whether facial features of a human face are contained in an acquired input image is detected, and whether the acquired input image is moved is detected, and the service condition of electronic equipment is judged. The decision method and the decision system solve the error decision problem of the service condition of the electronic equipment.

The present invention discloses a decision optimization method of multi-land seed selection based on combination optimization. Aiming at a multi-land seed selection decision problem, a multi-land grain seed selection decision model based on a combination optimization theory method is constructed to achieve decision optimization of multi-land seed selection. The method comprises the steps of: constructing a training sample set; obtaining key factors which influence a yield; constructing a neural network model and performing training; constructing a training test set; obtaining a yield prediction value and a variance of a parcel through the trained neural network model; constructing a multi-land grain seed selection decision optimization model based on combination optimization; and employinga decomposition algorithm to solve an optimal seed selection proportion to obtain the optimal varieties of the seeds and a usage proportion. The decision optimization method of multi-land seed selection based on combination optimization can perform analysis and optimization of grain plantation and seed selection, and can guide concrete land agriculture plantation and agriculture plantation planning; and moreover, the decision optimization method provides technical support for macroscopic agriculture policy or regional sales and a stocking strategy so as to have great economic and social values.