6 results about "Decision making methods" patented technology

The application discloses a kind of reasoning-action combined extension embodied intelligent task decision-making method and system, belong to artificial intelligence field, can solve the decision decoupling problem of reasoning process and action space.The method comprises: obtaining the historical reasoning and historical action of each historical moment selected in multiple historical moments;According to the environment information and task instruction of current moment, generate the candidate reasoning and multiple candidate actions of current moment;Candidate reasoning and each candidate action corresponding thereto form a candidate decision;According to historical reasoning and historical action, determine the comprehensive score of each candidate decision, and determine the running mode of embodied intelligent model in current moment;When running mode is exploration mode, randomly select a candidate decision from all candidate decisions as to-be-executed decision;When running mode is utilization mode, select the candidate decision with the highest comprehensive score and higher than preset score as to-be-executed decision.The application is used to determine embodied decision.

The invention discloses a large language model collaborative decision-making method and system based on reinforcement learning driving, and belongs to the technical field of strategic game decision-making, and the method comprises the steps: building a large language model strategy: converting a game state into a structured text prompt, and generating an executable micro-action sequence; constructing a tool strategy: generating a predefined macro action based on the micro action sequence; and collaborative decision-making: cooperatively scheduling the large language model strategy and the tool strategy by adopting a near-end strategy optimization algorithm. An action self-adaptive adjustment mechanism is introduced, and the execution length of the micro action and the macro action is dynamically adjusted according to the game stage; and a layered reward function is designed, and dense and effective training signals are provided in combination with the overall winning rate reward and the local performance reward. According to the method, the problems of semantic conflicts, resource conflicts and attention offset existing in tool calling of a large language model are effectively solved, and the decision-making efficiency, the strategy coordination and the training stability are improved.

This invention discloses a consensus decision-making method, system, and storage medium based on distributed feedback, belonging to the field of internet collaborative decision-making technology. The method constructs a fully implicit autonomous decision-making system, including: collecting feedback data from network nodes on a target object; weighting the feedback data based on the implicit real-time reputation weights of the nodes to generate a consensus quantification result; silently updating the system state of the target object according to the consensus quantification result; and dynamically and silently adjusting the implicit reputation weights of the nodes in the background according to the state update result. This invention severs the direct causal chain between the perceptible behavior of nodes and system feedback, encapsulating consensus decision-making as a background environmental law unknown to the nodes. This fundamentally reduces the motivation for strategic gaming and malicious attacks by nodes, thereby improving the system's anti-attack capability, stability, and decision quality. It achieves game-resistant, self-stabilizing, and sustainably optimizing group decision-making, effectively solving governance problems such as information overload and the proliferation of low-quality content in large-scale digital communities.

The application discloses an intelligent transaction decision-making method and system based on hierarchical multi-round confrontation debate, which is suitable for stock and other financial asset transaction scenarios. The method realizes efficient, low-delay and interpretable intelligent transaction decision-making by constructing a hierarchical multi-round confrontation debate mechanism, solving the limitations of existing technologies in deep reasoning, delay control and decision transparency. The core of the method is to build a complete process of multi-agent confrontation debate, dynamic reputation evaluation and transparent decision-making, divide the multi-role agent into the argument generation group and the refutation group, and execute R rounds of deep debate. In the debate process, the agent reputation is updated by the Bayesian reputation update, and the transaction signal is generated within milliseconds by combining the Soft-Borda dynamic voting mechanism. At the same time, through the transparent decision link of the multi-round debate log and the explanation vector, the requirements of the model interpretability and real-time audit of the financial supervision are met.

The invention provides a multi-agent decision-making method based on active reasoning. The method comprises the steps that a teammate active reasoning network is constructed, the teammate active reasoning network models a teammate active reasoning process from the three aspects of perception, belief and action by means of a main agent, and observation perception representation of the teammate agent to the environment, cognitive belief representation of the teammate agent to the environment and prediction action of the teammate agent to the next time step are obtained; the main agent screens the sensing representations of the modeled teammate agents, unreliable sensing representations and corresponding teammate agents are screened out, and the reserved teammate agents serve as collaboration objects; and the main agent fuses the belief representations of the selected collaborative objects, and makes a decision according to the fusion result and the historical trajectory representation of the main agent. According to the method, the modeling information is selected by modeling the active reasoning process of the teammate agents and combining a dual screening mechanism, and the task execution efficiency and success rate of the multi-agent system in a dynamic complex environment are remarkably improved by reasonably modeling teammate behaviors and screening high-credibility collaborators.