25 results about "Strategy selection" patented technology

The invention provides a multi-intention processing method and device for a smart home, electronic equipment and a storage medium, and relates to the technical field of man-machine interaction. The smart home system can first judge whether intention conflicts exist among the different intentions, and if the intention conflicts exist, the smart home system can first calculate a first priority score corresponding to the intention map and a second priority score corresponding to the secondary intention; the method comprises the following steps: acquiring a first priority score and a second priority score, then performing strategy selection according to the first priority score and the second priority score, determining a conflict resolution strategy for the deliberate map and the secondary intention, and then executing a control instruction corresponding to the conflict resolution strategy, thereby improving intention understanding accuracy and response efficiency, and improving intention understanding experience. And the capability and decision of the system on complex intention conflicts are improved, the timeliness and flexibility of control instruction response are guaranteed, and the accuracy of system response is improved.

The invention provides a dynamic coding and decoding negotiation method based on link state quality. The dynamic coding and decoding negotiation method comprises the following steps: grouping reference station RTCM data and determining the priority of the RTCM data according to the message ID of the RTCM data; based on the determined first strategy selection mode, the RTCM data is subjected to check code generation, character coding and encryption in sequence, and the strategy selection mode of the base station is synchronized to a strategy selection module of the moving station; after the moving station receives the data, decryption, character decoding and verification are sequentially carried out on the encrypted data based on the information synchronized by the strategy selection module; meanwhile, the moving station counts the packet loss rate and the verification failure rate of the RTCM data in real time; and adjusting the strategy selection mode of the next group of data according to the quality of the current link, the system load and the priority of the RTCM data until all grouped data are sent. According to the invention, a real-time closed loop is provided, and the link quality is evaluated more comprehensively.

The application discloses a Text-to-SQL intelligent agent coordination system and method for perceiving potential intention of a user in a multi-round context. The system comprises a memory bank, a prompt strategy pool, a strategy recommender, an SQL generator, an SQL refiner and a strategy optimizer. The memory bank is used for storing system running data. The prompt strategy pool is used for storing prompt strategies. The strategy recommender is used for receiving user information and strategy information of the prompt strategies, identifying the current potential intention of the user based on the user information and the strategy information, and matching the optimal strategy from the prompt strategy pool according to the potential intention to constrain the generation of an SQL query statement. The SQL generator generates an SQL query statement based on the current natural language query of the user, the system running data and the optimal strategy. The SQL refiner is used for adjusting and iteratively updating the SQL query statement. The strategy optimizer is used for updating the prompt strategies. The application increases strategy selection before SQL generation, so that the SQL generator can generate an SQL statement meeting the user demand at one time with a higher probability, and the cost is reduced.

The present application relates to a kind of multi-stage self-defined combination strategy warehouse picking optimization method, device and medium, wherein, method includes collecting order data to be picked and preprocessing generation structured data set;Support user self-defined configuration wave division dimension and weight, complete order clustering and generate picking wave with unique identification;Divide the whole process of picking into four core stages, provide each stage strategy selection and combination interface, generate compatible multi-stage strategy combination scheme;Correlation scheme and wave task are executed in stages, real-time acquisition of field data and dynamically correct abnormal strategy;After completing picking, the core index is counted, report is generated and review output optimization suggestion, the present application solves the problem that traditional picking strategy is poor in adaptability, low efficiency, can be flexibly adapted to multiple warehouse scene, improve picking efficiency and accuracy, reduce resource consumption, and the practical value is remarkable.

The application discloses a planning calibration method and system for an information analysis intelligent assistant, and comprises the following steps: in each planning round, a candidate strategy set is generated by a planning agent, the candidate strategy set contains an information retrieval target, a retrieval keyword combination and an information processing flow; the candidate strategy set is evaluated by multiple agents with heterogeneous private memories, a strategy selection module based on the agents is constructed for information consistency, and a strategy with the most stable evaluation result is selected as an optimal strategy of the round; a difference between a self-selected strategy of the planning agent and the selected optimal strategy of the round is recorded, a calibration constraint module guided by consistency is constructed, and the difference is converted into a cognitive calibration constraint and stored in the memory of the planning agent; strategy selection optimization is realized in single-round planning by relying on the strategy selection module based on information consistency; and the cognitive calibration constraint generated by the calibration constraint module guided by consistency is combined, so that the constraint is transmitted between different rounds and gradually accumulated, and a constraint is formed on strategy generation of the planning agent; and the trained agent is applied to an open domain webpage retrieval task to generate accurate retrieval results.

This invention relates to the field of wireless body area network (VBA) technology, and particularly to a method for selecting routing strategies for vehicular network (V2N) communication based on multiple communication state indicators. The method includes: S0, setting a strategy set; S1, obtaining communication demand features ( ) and network state feature vectors; S2, filtering to obtain a candidate strategy set P'; S3, constructing a strategy selection input description, associating each candidate strategy ∈ P' with a path optimization method type; S4, selecting the optimal strategy based on the calculated matching score of each candidate strategy; S5, generating a path control scheme; and S6, executing the path control scheme. This method can reduce the coupling between routing planning methods and specific implementations, improving the robustness and adaptability of V2N communication.

The application discloses a method for optimizing infrastructure network game strategy based on linear constraint, which comprises the following steps: obtaining the topology structure of the infrastructure network, determining the strategy set of the attack side and the defense side, and constructing a basic model of complex network game; determining an index representing network connectivity, calculating the income of the attack side and the defense side under various strategy profiles in the complex network game model, and thus obtaining an income matrix; adding a restriction condition to the strategy selection of the attack side and the defense side, and thus obtaining the strategy constraint set of the attack side and the defense side; converting the solution of the basic model of complex network game into the solution of a linear programming problem, obtaining a mixed Nash equilibrium solution, and thus obtaining the strategy optimization result of the attack side and the defense side. The method can greatly widen the application of complex network game research in practice by introducing the constraint matrix countermeasure into the infrastructure complex network game.

The invention relates to an analysis and decision-making method for channel selection of a double-channel closed-loop supply chain, and belongs to the technical field of supply chain management. According to the method, a demand and profit function is established based on a consumer net utility theory to construct a game model comprising three modes of manufacturer network direct sales (M strategy), retailer traditional retail (R strategy) and double-channel parallel (D strategy), and a reverse induction method is adopted to solve equilibrium decisions and optimal performance under each strategy. According to the method, the balance results of all strategies are compared and analyzed through the system, and key market parameters, namely acceptance of consumers on reprocessed products and acceptance of the consumers on channel transfer, and decisive influences on profits of all parties and optimal strategy selection, are revealed. According to the specific levels of the parameters, a scientific and quantitative decision basis is provided for a manufacturer to select an optimal channel strategy considering individual profit and system total benefit, and the adaptability and competitiveness of an enterprise in a complex market environment are effectively improved.

The invention discloses a port resource collaborative scheduling method, system and device and a medium, and the method comprises the steps: obtaining a strategy library and historical resource data of a target port; the strategy library comprises a plurality of different first solving strategies; performing example generation on the historical resource data to obtain a plurality of port examples; according to each first solving strategy in the strategy library, performing example solving on all the port examples to obtain a plurality of example training groups; and according to all the example training groups, performing parameter updating on the initialized strategy selection model to obtain the trained strategy selection model. According to the method, the target solving strategy is determined through the trained strategy selection model, and the to-be-solved resource example is solved, so that the efficiency and effect of port resource collaborative scheduling can be improved. The invention relates to the technical field of port scheduling.

The application discloses a kind of partition division strategy selection method and its related components based on two-way tracking, it is related to partition division field, obtains the train running time of track line in the same direction Same multiple one-way partition division strategies, respectively determine the train tracking interval time of these strategies in two driving directions, retain the one-way partition division strategy in the train tracking interval time of two directions is less than the preset tracking interval time on the direction pair, and delete other strategies, in the retained one-way partition division strategy, according to the specific partition quantity in one-way partition division strategy and train tracking interval time determine the final two-way partition division strategy. By selecting the strategy that train tracking interval time is relatively short in two directions in multiple partition strategies, a partition division strategy with higher regional carrying capacity and train operation efficiency can be selected under the premise of ensuring the safety of train operation.

The invention relates to the technical field of air combat intelligent aid decision making, in particular to an aid decision making recommendation model adaptive method and system based on a Bayesian neural network, and is particularly suitable for man-machine collaborative decision making in a 2V2 air combat game scene. By designing a self-adaptive framework and an incremental learning algorithm, the model can be dynamically updated based on real-time data in the use process of a user, so that the recommendation accuracy and robustness are improved. The method comprises the following steps: constructing an incremental learning algorithm of a Bayesian neural network, analyzing the time complexity and the space complexity of the incremental learning algorithm, evaluating the real-time performance of the incremental learning algorithm in an online reasoning process, and researching the robustness performance under the condition of data uncertainty. The method can be widely applied to strategy selection in an intelligent recommendation system, an auxiliary decision-making platform and a multi-agent game.

The present application relates to the technical field of network security, in particular to a network attack and defense strategy selection method and system based on three-party evolutionary game, a three-party evolutionary game model is constructed by an attacker, a defender and a user as network attack and defense confrontation participants based on evolutionary game theory, the participants of the three-party evolutionary game model imitate the advantageous strategy in the evolutionary game and adjust their own strategy selection under the driving of game income; the stability of the three-party evolutionary game model is solved by using a three-party behavior strategy selection dynamic equation, the equilibrium point of the three-party evolutionary game model is obtained, and the best selection strategy of each subject of the participants is obtained according to the equilibrium point of the three-party evolutionary game model. The present application models and describes the real attack and defense scene more accurately by constructing a game model with the participation of the attacker, the defender and the user, and provides technical support for solving the network information security problem.

The invention discloses a new energy automobile charging mode analysis method based on an evolutionary game theory, and the method comprises the steps: firstly, determining a government, an automobile manufacturing enterprise and a consumer as game participants, and defining a strategy set and influence factor variables of the game participants; based on this, a payment matrix coupling benefits of three parties is constructed, and then a replication dynamic equation is established by taking the strategy selection probability of each party as a state variable; a system equilibrium point is obtained by solving an equation, stability analysis is carried out by using a Jacobian matrix eigenvalue method, and a stability condition for enabling the equilibrium point to become an evolution stability strategy point is deduced; and finally, based on the condition analysis, factor variables needing to be adjusted for realizing the expected stable strategy are analyzed, and a decision support result is generated. According to the method, the defect that the prior art lacks a dynamic view angle, a system framework and a quantitative model is overcome, the three-party strategy interaction evolution process can be quantitatively depicted, the stability condition is determined, and a scientific and quantitative basis is provided for industrial policy making and enterprise strategic decision making of a battery charging and replacing mode.

The present application belongs to the technical field of supply chain decision optimization, and particularly relates to a supply chain evolution game decision optimization method and system fusing multi-source data. The present application firstly calculates strategy distribution entropy based on the strategy selection proportion of each subject, then calculates external disturbance factors based on external environment indexes in the global trusted state data, secondly calculates comprehensive game state entropy based on the calculation results of strategy distribution entropy and the calculation results of external disturbance factors, and finally sets a game imbalance early warning threshold to pre-judge potential risks based on the calculation results of comprehensive game state entropy, and automatically generates a regulation and control strategy to intervene in the game process when the early warning is triggered, thereby guaranteeing the stable operation of the supply chain. This progressive calculation method makes the entropy value not only reflect the change of the internal game state, but also perceive the disturbance of the external environment, thereby realizing accurate risk measurement combining internal and external factors, and being more scientific and authentic than single-dimensional entropy value calculation.

The application relates to a kind of energy storage cost dredging evolutionary game method, comprising: (1) collecting the corresponding information of target area power supply side, grid side, user side main body, and establishing the evolutionary game matrix for energy storage cost dredging;(2) based on the evolutionary game matrix, calculate the expected income of power supply side main body strategy selection and its replication dynamic equation;(3) based on the evolutionary game matrix, calculate the expected income of grid side main body strategy selection and its replication dynamic equation;(4) based on the evolutionary game matrix, calculate the expected income of user side main body strategy selection and its replication dynamic equation;(5) carry out evolutionary game, analyze the scene of evolutionary game;(6) combined with the policy that may be taken for the result of each evolutionary game, consider the change of user demand response willingness, simulate the probability change of user willingness by Markov chain, and return to step (1) to continue evolutionary game.The method is beneficial to accurately analyze the possibility of each main body bearing energy storage cost.

The invention provides an official document writing optimization method and device, electronic equipment and a storage medium, and relates to the technical field of artificial intelligence, and the method comprises the steps: extracting a state feature of a current writing task from writing demand information of a user; inputting the state features into a strategy selection model, obtaining a writing strategy output by the strategy selection model, and generating a first document according to the writing strategy; generating and fusing different types of reward signals according to the interactive behavior data of the user for the first document of the official document to obtain a comprehensive reward signal; and updating parameters of the strategy selection model by using the comprehensive reward signal to complete online optimization of the writing strategy. The writing strategy is dynamically selected through the state characteristics of the writing task, and the parameters of the strategy selection model are optimized in real time based on user interaction feedback, so that the problems that an existing official document writing system is fixed and rigid in strategy and cannot adapt to a dynamic scene are effectively solved; dynamic matching of strategy selection and scene requirements and optimization of strategy performance are realized.

The invention discloses a self-adaptive multi-strategy fusion large language model training optimization method. The method comprises the following steps of 1, constructing a multi-strategy data processing framework; 2, designing an intelligent strategy selection algorithm; 3, adopting a LoRA parameter high-efficiency fine tuning technology; 4, adopting a gradient optimization strategy; 5, processing the training data based on the selected optimal data processing strategy; according to the adaptive multi-strategy fusion large language model training optimization method based on data feature perception provided by the invention, a traditional training mode depending on artificial experience is changed fundamentally; the method has the technical advantages of automatic strategy selection, remarkably improved training efficiency, optimized resource utilization rate, high universality and expandability and high parameter efficiency.

The invention provides a multi-strategy engine application method and system of an intelligent agent, and the method comprises the steps: a strategy evaluation model of a strategy layer receives request feature data of a task request, input by a user, of the intelligent agent, outputs an evaluation result, and a strategy selection model selects a target strategy engine from more than two strategy engines based on the evaluation result, the target strategy engine generates a calling instruction; the execution layer receives the calling instruction, executes tool calling operation and returns an operation result to the strategy layer, and the strategy layer generates task reply information according to the operation result and returns the task reply information to the user through the interface layer. According to the method, the intelligent body can adapt to different tasks and is suitable for a proper target strategy engine, and the coordination of solving complex and simple tasks at the same time is improved.

The application discloses a production line scheduling method based on data and model hybrid driving, first, according to the resource position and layout, the number and type of operation orders, random noise disturbance and other elements in the flexible production line, a flexible production line scheduling model under random disturbance is constructed; then, according to the data and model hybrid driving framework, the scheduling model is solved; in the data and model hybrid driving framework, the data driving refers to the multi-agent reinforcement learning algorithm based on value decomposition, the model driving refers to the heuristic dispatching rule and experience knowledge, and the hybrid driving of the two mainly embodies three aspects: one is that the scheduling rule is used for guiding the agent strategy selection, two is that the reward function is designed based on the scheduling scene knowledge, and three is that the historical experience and data are supervised to train the noise detection and denoising module; finally, in the disturbance environment, the production line scheduling is carried out by adopting the data and model hybrid driving method, and the application is used for solving the flexible production scheduling under random disturbance.

The application relates to the technical field of intelligent conversation, and discloses an artificial intelligence-based conversation marketing strategy optimization method, which comprises the following steps: collecting user voice, text, facial expression and physiological signals in real time through a multi-modal sensing component of a terminal device, constructing a dynamic emotion map and extracting a multi-modal feature vector; using a cross-modal information processing module to align the multi-modal data through a contrast learning algorithm, generating a causal relationship description of user behavior and strategy and a strategy risk score; cooperatively training a global strategy model through differential privacy and homomorphic encryption technology; generating an emotion-adaptive conversation script in combination with the dynamic emotion map and the causal model, and optimizing strategy selection through a reinforcement learning algorithm; and dynamically updating the emotion map, the risk score and the global model through a closed-loop feedback mechanism, so as to form a real-time optimized strategy generation system. The application can improve the practicability of artificial intelligence for real-time business.

The technical scheme of the invention discloses a master-slave game regulation and control strategy method oriented to vehicle network interaction. The master-slave game regulation and control strategy method for vehicle-network interaction disclosed by the invention comprehensively considers the comprehensive benefits of a power grid side, an electric vehicle master side and an operator side, guides the realization of a global target by relying on a dynamic electricity price mechanism and an integral incentive policy, tracks the strategy selection of both parties of the vehicle network, ensures that the discharge electricity price is adaptive to a user response range, and improves the user experience. Finally, a Nash equilibrium point is formed to maximize the revenue of both parties.

The invention discloses a big language model dialogue method, a big language model construction method, a big language model dialogue device, a big language model dialogue system and big language model construction equipment. According to the big language model dialogue method, a deep thinking part of a model is trained through an inference type big language model, so that the deep thinking part can carry out detailed inference analysis on risk problems, and the inference analysis comprises analysis of user intentions and strategy selection. And analyzing the user problem to analyze a risk type, a user risk level, a risk intention and a demand intention in the user problem. According to the requirements and the safety criteria of the user, a safety reply strategy and a requirement meeting strategy are selected, the safety reply strategy comprises refusal or forward guidance of the risk intention, and the requirement meeting strategy comprises meeting of the requirements of the user within a legal range. By adopting the processing mode, the problem of excessive rejection of the large language model can be effectively solved, and risk migration caused by rejection type interaction is avoided.

This invention belongs to the field of organizational strategic management and can help organizations achieve informatization and digitalization in the strategic selection phase. To address the problems of lack of systems, inability to integrate data, and low level of process specialization in the strategic selection phase, this invention proposes a strategic selection system and implementation method integrating LLM and BI. It includes strategic selection modules (overall strategy module, business unit strategy module, blue ocean strategy module, functional strategy module, etc.), APIs (Application Programming Interfaces) with mainstream domestic and international LLM (Large Language Model) systems, table and dashboard modules, APIs (Application Programming Interfaces) with mainstream domestic and international BI (Business Intelligence) systems, and implementation methods.

The application discloses an advertisement material automatic fission method and system based on artificial intelligence, and relates to the technical field of digital advertisement. The method comprises the following steps: automatically identifying a hit advertisement material through real-time monitoring of performance data, and performing multi-modal feature extraction on the hit advertisement material for in-depth analysis; a central dispatching engine dynamically selects a fission strategy combination from a strategy library based on the analysis result, and calls a generative AI model to generate new derivative materials; the new materials are put into test, performance data feedback is collected to the engine, a closed-loop optimization is formed to update the strategy selection logic. The application aims to solve the problems of low advertisement creative fission efficiency, content homogenization and unstable effect.

The application discloses a kind of based on variational bayes inverse reinforcement learning strategy emotional support dialogue system, its characteristics are to adopt including: emotional support strategy selection module, dialogue common sense analysis reasoning module and reply generation module are constructed strategy emotional support dialogue system, the emotional support strategy selection module is by dialogue state encoder and variational bayes inverse reinforcement learning strategy selection module, according to current dialogue history information and the strategy used to carry out strategy selection;The dialogue common sense analysis reasoning module generates and predicts the common sense information analysis relevant to current dialogue;The reply generation module will generate content and dialogue history generation reply.The application compared with prior art has enhanced the information richness and accuracy of generated reply, effectively improves the strategy selection accuracy and user's experience in the strategy emotional support dialogue scene, with good use prospect and commercial development value.