8 results about "Cost metric" patented technology

The invention provides a computing power scheduling method, device and equipment for large model heterogeneous reasoning and a medium, and belongs to the technical field of computing power scheduling, the method comprises the following steps: in response to a model deployment request, obtaining performance index data and cost index data of each candidate deployment unit in a heterogeneous computing resource pool; determining a dynamic weight of each candidate deployment unit, and calculating a comprehensive cost performance score of each candidate deployment unit based on the dynamic weight of each candidate deployment unit and the performance index data and the cost index data of each candidate deployment unit; and selecting the candidate deployment unit with the highest comprehensive cost performance score as a target deployment unit, and scheduling the target large model to heterogeneous computing resources corresponding to the target deployment unit for execution. According to the large model heterogeneous reasoning-oriented computing power scheduling method provided by the invention, the scheduling decision is made by comprehensively considering the dynamic weight, the performance and the cost of the candidate deployment unit, so that the real-time performance, the accuracy and the adaptability of computing power scheduling are improved.

The invention provides a data processing method and device, electronic equipment, a storage medium and a program product, relates to the technical field of information, and is used for improving the accuracy and reliability of data resource evaluation. The method comprises the steps that data resource index parameters of multiple dimensions of a basic cost index, a quality index, an application index and a resource relation index of a target entity are collected, the basic cost index is processed through a target resource model, and basic cost information of a first data resource of the target entity is obtained, on the basis of the basic cost information and the resource relation index, calculating resource dependence cost of the second data resource; calculating the value score of the first data resource based on the quality index and the application index through an analytic hierarchy process; determining the expected feedback rate corresponding to the value score interval in which the value score is located as the expected feedback rate of the first data resource; and based on the basic cost information, the resource dependence cost and the expected feedback rate of the first data resource, obtaining an evaluation result of the first data resource.

The application relates to the technical field of engineering cost supervision, and discloses a process execution supervision system and method suitable for engineering cost consultation projects. The method acquires a project whole-cycle data set through a distributed data acquisition interface, and the data set at least contains contract files and progress payment vouchers; the data set is subjected to structured feature analysis to generate a multi-dimensional project feature matrix containing time nodes, cost indexes and resource allocation amounts; a pre-constructed engineering cost benchmark parameter library is called, real-time deviation degrees of the multi-dimensional project feature matrix and the benchmark parameters are calculated based on a dynamic time warping algorithm; when the real-time deviation degrees exceed a preset threshold range, an abnormal procedure identification engine is automatically triggered to extract procedure feature vectors including a construction period delay rate, a cost overrun rate and a material loss rate from deviated procedures; the procedure feature vectors are input into a pre-trained compliance verification model to output an abnormal diagnosis report containing a violation type, a risk level and a responsible subject.

The invention discloses a low-carbon building scheme comparison and selection method and device, a storage medium and a computer program product, and relates to the technical field of low-carbon resource management, and the method comprises the steps: defining a building component in each candidate low-carbon building scheme as a first global adjustable parameter, associating the first global adjustable parameter with a first full life cycle carbon emission factor and a first cost factor; capturing a parameter adjustment instruction sent by a user, and calculating a carbon cost index based on the parameter adjustment instruction, the first full life cycle carbon emission factor and the first cost factor; obtaining an optimization target and a constraint condition set by a user, and setting a multi-target optimization algorithm according to the optimization target and the constraint condition; and inputting the carbon cost index into the multi-objective optimization algorithm for iterative calculation to obtain a Pareto optimal solution set as a low-carbon building scheme comparison and selection result. And full-life-cycle multi-objective collaborative optimization comparison and selection of the candidate building schemes are realized.

Techniques for generating, simulating, and optimizing one or more provider-specific cloud-based architectures from a provider-independent architecture definition are disclosed. An architecture generator maps provider-independent service definitions to provider-specific service components for one or more specific cloud service providers. An architecture simulator simulates execution of a set of operations on the provider-specific cloud-based architectures to determine one or more performance and cost metrics. An architecture optimizer varies one or more design choices or parameters of a provider-specific service component to suggest which variant is optimal with respect to an optimization objective.

The embodiment of the invention relates to a processing method and device for planning a short message retransmission strategy based on a network awareness state. The method comprises the following steps: carding a state type set S, defining a cost index C, and configuring a strategy parameter set and a strategy parameter value domain DP corresponding to the strategy parameter set; constructing an index prediction model and a strategy prediction model; constructing a first data set; training an index prediction model based on the first data set; optimizing the optimal strategy configuration corresponding to each state sample by taking the minimum cost index as an optimization target, and constructing a second data set based on all the state samples and the corresponding optimal strategies; training a strategy prediction model based on the second data set; and the strategy prediction model and the index prediction model dynamically plan the short message retransmission strategy of the server according to the real-time network state of the short message sending server and predict the cost index of the planning strategy. According to the invention, the configuration flexibility of the short message retransmission strategy can be improved, and the cost control can be optimized.

The invention relates to the technical field of edge intelligence and distributed resource optimization, in particular to a detachable large model reasoning service deployment method giving consideration to resource utilization efficiency and profit optimization, which comprises the following steps: initializing an edge network structure and large model reasoning chain parameters, including server computing power, video memory capacity, link bandwidth and unit cost thereof; each model sub-module is traversed in sequence, a local maximization profit factor of the model sub-module on each candidate server is calculated, and the factor comprehensively considers expected income and deployment cost, including calculation of resource lease and communication bandwidth cost; selecting the server with the highest profit factor to deploy the current sub-module, updating the network resource state, and gradually completing the deployment of the whole inference chain; outputting a module position, a communication path and a profit cost index; according to the method, profit maximization is achieved while efficient utilization of resources is guaranteed, and the economic benefits and the resource utilization rate of the system are remarkably improved.

The invention discloses a data labeling quality and cost collaborative optimization method for a multi-label scene. The method comprises the following steps: firstly, evaluating sample scarcity and ambiguity to generate a priority list; secondly, constructing a decision arm, and balancing budget distribution between exploration and utilization by adopting a multi-arm Huger machine algorithm; according to the distribution result, differential hierarchy verification is executed, and the higher the depth is, the more complex the verification round and the model combination is; periodically comparing the quality and cost indexes with a threshold value to generate a feedback signal, and driving the strategy to be dynamically adjusted; meanwhile, a sliding window counts historical data to construct a parameter library of difficulty levels, cost coefficients and quality weights, and the parameter library is synchronized to each link to realize adaptive optimization; and the invalid sample budget is automatically returned and preferentially redistributed to the high-revenue queue. According to the method, budget allocation is driven through quality-cost feedback, the verification depth is guided through cost-income, dynamic balance of labeling quality, resource efficiency and cost control is achieved, and the labeling cost performance in a multi-label scene is improved.