21 results about "Planning approach" patented technology

The invention discloses an end-to-end planning method fusing mixed trajectory representation and curriculum reinforcement learning. The method comprises the following steps: constructing a discrete-continuous mixed representation end-to-end pre-training network; constructing a course strengthening fine tuning framework based on interactive deduction; and designing a hard and soft constraint coupled hierarchical course award mechanism. According to the method, a discrete intention and continuous residual error coupling mixed trajectory characterization mechanism is introduced, and on the basis that a driving intention is quickly locked by using discrete primitives, subgrid-level geometric correction is performed on a coarse-grained trajectory through parallel regression branches. According to the invention, on the premise of not increasing the complexity of the search space, accurate trajectory planning with both long-time-sequence intention consistency and dynamics smoothness is realized. According to the method, the reinforcement learning training efficiency is effectively improved, catastrophic forgetting of a long-tail risk scene is prevented, a safety boundary is established in a strategy planning decoder, and the robustness and decision-making ability of an automatic driving system under extreme working conditions are improved.

To provide a planning program, a planning method, and an information processing apparatus capable of suppressing local congestion in a feeding order into a work line.SOLUTION: A planning program that causes a computer configured to sequentially specify the order of processing a plurality of types of objects in a work line for processing the plurality of types of objects, to execute processing of: specifying, as a basic value, a ratio of the number of objects of each of the types to the total number of the plurality of types of objects; calculating, as an appearance probability, a value obtained by subtracting the number of objects of the type of which a processing order has been determined, from a product of a process number and the basic value for each of the respective types; and changing a frequency of the processing for each of the types in accordance with a selection status of each of the types, in selecting a type to be processed next using the calculated appearance probability.SELECTED DRAWING: Figure 4

To provide a planning program, a planning method, and an information processing apparatus capable of planning a feeding order to a work line while suppressing man-hours of a user.SOLUTION: A planning program that causes a computer to execute processing of: generating a plurality of priority order patterns with priority orders assigned thereto, for a plurality of work groups, when processes selected from the plurality of work groups are sequentially performed on each of a plurality of objects; specifying the order of processes in the work group for each of the priority orders according to the priority orders of the plurality of priority order patterns; and specifying the order of the plurality of objects corresponding to a combination of the specified processes.SELECTED DRAWING: Figure 6

The application provides a regional micro-grid planning control method and device based on period division, which comprises the following steps: obtaining power prediction data in a future first planning period of a region, determining micro-grid macro constraint parameters of the first planning period according to the power prediction data; constructing planning benchmark parameters of a second planning period based on the micro-grid macro constraint parameters and the power prediction data; determining operation control parameters of a third planning period based on the planning benchmark parameters; executing the operation control parameters of the third planning period, collecting operation data in execution and generating operation evaluation parameters; correcting the micro-grid macro constraint parameters of the first planning period and / or the planning benchmark parameters of the second planning period according to the operation evaluation parameters; and the planning method can adapt to regional power consumption changes.

A planning program according to the present invention causes a computer to execute a process for: accepting specification of a task; causing a large language model to iteratively perform, on the basis of the specified task and a set of functions which includes a first function that executes a process corresponding to a specific function and a second function that causes the large language model to select a function to be called from among candidates of permitted functions, planning in which subtasks are expanded to a hierarchy lower than the hierarchy of the specified task and the first function or the second function to be executed for the subtasks is selected; and outputting a tree-structure plan in which the subtasks are expanded for each hierarchy as a result of the iterative planning.

The application relates to the field of artificial intelligence and automatic planning, and in particular to a large-scale planning problem solving method based on general knowledge of a large language model, which comprises the following steps: parsing a PDDL domain file and an instance file to obtain an initial state, a target state and an action model; decomposing the target state into sub-targets in the form of directed binary tuples, constructing a directed acyclic dependency graph and topologically sorting to obtain an ordered sub-target sequence; recursively traversing the sequence to generate sub-problems and calling a planner to solve the sub-problems, and if the solving fails or times out, optimizing through two modes of the LLM; mode A selects actions from an executable action sequence to compress a search space, and mode B predicts an intermediate state to split an intermediate sub-problem; and finally, sequentially concatenating sub-plans to generate an overall plan; the application aims to solve the problems of large search overhead, difficulty in determining the order of sub-targets and unreliable LLM output of a traditional planning method.

This invention provides a reactive planning method based on a large language model, comprising: designing replanning logic, including multi-step logic and three-hop logic, analyzing the impact of various factors on the completion of the task objective step by step, and updating the current plan in a chain; each step employs reasoning logic including cause analysis logic, conclusion derivation logic, and plan adjustment logic; designing replanning prompts; receiving user commands, generating an initial plan based on user instructions, storing the initial plan, and recording the current plan's progress information; detecting and capturing environmental change information through the robot, and iteratively updating the plan based on environmental change information, plan progress information, user commands, replanning prompts, and replanning logic. This invention also provides a reactive planning system, storage medium, and electronic device based on a large language model. Therefore, this invention enables low-cost, high-accuracy, and high-generalization robot reactive planning.

The invention relates to a planning method for planning a configuration structure and a procedure for a production unit for implementing a production process on a product in the production unit, wherein in a definition step 100 the product with the associated production information is defined, wherein in a selection step 200 a selection of selection substructures is made, wherein the selection substructures are suitable for use in the production process, and wherein the selection substructures form a set of selection substructures; wherein in a configuration step 300 a planning configuration structure of the production unit is generated from a plurality of target substructures, wherein the target substructures are formed from a subset of the selection substructures of the set of selection structures;wherein in a procedure generation step 400 a planning procedure with procedure steps for implementing the production process is formed with the planning configuration structure of the production unit; wherein in a coordination step 500 the planning procedure and the planning configuration structure are coordinated with each other in order to obtain the configuration structure and the procedure for the production unit for implementing the production process.

This invention relates to a two-stage robust planning method for demand-side capacity that considers both cost and flexibility resources, belonging to the field of power system operation and maintenance technology. By introducing an uncertainty adjustment parameter, this invention constructs a box-shaped uncertainty set with flexibly adjustable conservatism, effectively mitigating the inherent drawbacks of overly conservative traditional robust optimization methods. This allows the model to dynamically adjust the boundary of the uncertainty set according to the different risk tolerances of the actual system, thus achieving a refined trade-off between economy and reliability during the planning stage. The two-stage robust optimization model constructed in this invention, considering the electricity-carbon coupling mechanism, incorporates carbon trading costs as an important component of operating costs into the objective function, effectively responding to the needs of the national green and low-carbon energy strategy. The improved column generation and interior point hybrid algorithm designed in this invention proposes an adaptive scenario selection mechanism and a gradient-guided initialization strategy, significantly shortening the solution time while ensuring the global optimality of the solution.

This invention provides an autonomous planning method based on a large language model, applicable to the field of artificial intelligence. The method includes: using a large language model to perform reasoning analysis on target environment data and user instructions to obtain a set of executable actions for the robotic arm, state change rules for the robotic arm and environmental objects in the target environment data during the execution of the executable actions, and state definition information for the robotic arm and environmental objects at the start and end of executing user instructions; generating a sequence of actions to be executed by the robotic arm based on the set of executable actions, state change rules, and state definition information; and, during the execution of the sequence of actions to be executed by the robotic arm, for each action to be executed, repeatedly performing the following operations: determining the control instructions for the current execution stage based on the updated environment data and robotic arm state data obtained after the previous execution stage, so as to control the robotic arm to implement user instructions based on the control instructions.

To easily execute cast frame layout consolidated to a cast frame in a short time.SOLUTION: A plan creation apparatus (100) comprises: an initial set creation unit (34) creating a plurality of initial sets (C0) in such a manner that the breakdowns of an included virtual steel material (s) are changed; a dual problem solving unit (35) solving the dual optimum solution of dual problems of set partitioning problems; a promising set extraction unit (36) having a candidate set creation unit (361) creating candidate sets and a calculation unit (362) calculating contracted expenses (ηj) of the candidate sets, and extracting all the candidate sets in which the contracted expenses satisfy prescribed standards as promising sets (Ce); and an original problem solving unit (37) solving the optimum solution of the set partitioning problems based on the initial sets and the promising sets.SELECTED DRAWING: Figure 4

The invention provides a planning method and system for new facilities of an existing project, and the method comprises the steps: obtaining the function demands of the existing project, and determining the performance parameters of the corresponding new facilities; calling a preset standard flow of a newly-added facility in the existing project, and generating a preliminary construction scheme of the existing project and the newly-added facility; optimizing the preliminary construction scheme of the newly-added facilities of the existing project through a processing algorithm, and generating construction stages and corresponding construction contents of the existing project and the newly-added facilities; and generating a project progress plan based on the construction stages of the existing project and the newly-added facility and the construction logic relationship between the corresponding construction contents. According to the invention, by integrating the standard process of the newly added facilities in the existing project, combining the algorithm optimization to generate the construction stage and the corresponding construction content, and determining the progress plan based on the construction logic relationship, the problem that technical conflicts and project processing scheme generation are difficult to consider by depending on manual planning in the prior art is solved; and the problem of design modification in the implementation stage is solved.

The invention relates to a dialogue strategy planning method based on nested rollback strategy adaptation. The method comprises the following steps: formalizing a dialogue strategy planning task into a Markov decision process; initializing a strategy pi; the number of iterations and the hierarchy depth are set for controlling the recursion depth of the nested rollback strategy algorithm; dialogue strategy planning is conducted through a nested rollback strategy method, strategy optimization is achieved through multi-level exploration in the nested rollback strategy method, and the strategy weight is updated by simulating a dialogue path and selecting the optimal sequence on each level; and evaluating the quality of the final state of the simulated dialogue path through a reward function. In combination with the generation capability of a large language model, efficient dialogue strategy planning can be realized under a zero sample learning normal form. According to the method, through multi-level strategy optimization and dynamic weight adjustment, the strategy can be dynamically adjusted according to the dialogue progress, dialogue behavior selection is optimized, and therefore the response quality and task completion efficiency of a dialogue system are improved.

This planning program causes a computer to execute: a first step for arranging a plurality of target objects as processing entities for a plurality of processes, by means of forward arrangement in which the arrangements are planned in accordance with a predetermined order for the plurality of target objects, starting from a predetermined time point; a second step for arranging the plurality of target objects as processing entities for the plurality of processes, by means of backward arrangement in which the arrangements are planned with the time axis in the opposite direction, starting from a predetermined time point, in accordance with the order in which processing of each target object is completed in the final process in the arrangement obtained in the first step; a third step for left-aligning a processing start time point of each target object in each process in the arrangement obtained in the second step; and a fourth step for left-aligning the arrangement obtained in the third step to a predetermined time point.

The application relates to a power system multi-resource collaborative planning method based on attribution analysis, and belongs to the technical field of power system planning. The method comprises the following steps: constructing a target function based on the load demand of a power system and the output of each unit, taking the minimum mismatch degree between supply and demand as the target, constructing a planning model based on set constraints and the target function; solving the planning model based on multiple sets of boundary condition values of the planning model to obtain corresponding planning schemes; analyzing and obtaining quantitative representations of the influence degree of each boundary condition and each constraint on the planning scheme; adjusting the boundary condition values based on the quantitative representations and solving the planning model, and selecting the planning scheme with the lowest planning cost in the solving result as the optimal planning scheme. The application determines the key factors influencing the planning scheme through attribution analysis on power planning, and determines the optimal planning scheme by adjusting the key factors, so that the scientificity and economy of the planning scheme are further improved.

The invention provides a port and wharf capacity planning method and system based on full-link intelligent scheduling, and the method and system construct a full-link collaborative double-layer planning model, and achieve the purpose of combining simulation modeling with intelligent decision. The model adopts a double-layer structure, and the upper layer adopts a multi-target intelligent optimization algorithm and is used for calculating investment cost, throughput and resource utilization rate. And the lower layer generates an optimal scheduling scheme based on dynamic constraint planning. And global decision optimization is realized through bidirectional data interaction between the upper layer and the lower layer. The method has the advantages that simulation modeling and intelligent decision making are combined, and the scientific preciseness, the economic efficiency and the adaptability of the planning process are remarkably improved. According to the method, the challenges of demand uncertainty, large peak-valley demand difference, strong resource relevance and system complexity faced by a traditional planning method are solved. The method can effectively deal with the operation characteristic of coexistence of the wharf operation peak value and normal operation of the oil logistics support base in the middle sea.

To improve the accuracy of a manufacturing plan when one or more types of shape steel are cut out from a plurality of billets.SOLUTION: A planning device (100) comprises: an information acquisition unit (31) configured to acquire order information of a steel material; a condition setting unit (33) configured to set a constraint condition indicating a constraint on manufacturing a billet and the steel material; a planning unit (34) configured to make, based on the order information and the constraint condition, a manufacturing plan including an arrangement pattern in which a type and a quantity of a steel material to be cut out from each billet are defined and a quantity for each arrangement pattern; and an output unit (2) configured to output the manufacturing plan thus made. The constraint condition includes a billet shape condition that defines a relation between a state of a roller used for rolling and a shape of a billet to be manufactured using the roller.SELECTED DRAWING: Figure 7

To make it possible to make individual processing plans for a plurality of objects including an object processed through a plurality of plants.SOLUTION: A plan creation device (100) includes: an acquisition unit (11) which acquires, for a plurality of objects manufactured through one or a plurality of processing processes, order information indicating relation between the objects, kinds, and order quantities, kind and process information indicating relation between the kinds and the processing processes and processing facilities required for manufacturing the kinds, capability information indicating processing capability of the processing facilities for each kind, and facility change information indicating a replaceable processing process or processing facility; a generation unit (12) which generates a plurality of candidates of processing plans based on the order information, the kind and process information, the capability information, and the facility change information indicating the replaceable processing process or the processing facility; and a determination unit (14) which determines the processing plan from among the plurality of candidates.SELECTED DRAWING: Figure 2

This invention proposes an event planning method and system based on a large-scale model thinking chain, belonging to the field of event planning technology based on large-scale models. The method includes: constructing a planning chain from collected event planning data to obtain a planning chain reasoning trajectory; training a pre-set large language model through knowledge distillation based on the planning chain reasoning trajectory to obtain a chain-like planning large language model; and generating event plans through reasoning using the chain-like planning large language model. This invention has the advantages of simultaneously considering the effectiveness, efficiency, scenario generalization ability, and interpretability of event planning.

The invention relates to the field of artificial intelligence and automatic planning, in particular to a large-scale planning problem solving method based on general knowledge of a large language model, and the method comprises the steps: analyzing a PDDL domain file and an instance file, and obtaining an initial state, a target state and an action model; the target state is disassembled into sub-targets in a directed two-tuple form, a directed acyclic dependency graph is constructed, topological sorting is carried out, and an ordered sub-target sequence is obtained; recursively traversing a sequence to generate a sub-problem, calling a planner to solve, and optimizing through LLM two modes if the sub-problem fails or is overtime; the mode A selects actions from an executable action sequence to compress a search space, and the mode B predicts an intermediate state to split an intermediate sub-problem; and finally connecting the sub-plans in series in sequence to generate an overall plan. The invention aims to solve the problems that the traditional planning method is high in search overhead, the sub-target sequence is difficult to determine and the LLM output is unreliable.