CN-122022305-A - Construction technical scheme compiling self-adaptive optimization method and system based on multi-source data

CN122022305ACN 122022305 ACN122022305 ACN 122022305ACN-122022305-A

Abstract

The invention discloses a construction technical scheme compiling self-adaptive optimization method and system based on multi-source data, which relate to the field of construction management and comprise the steps of collecting various data of the whole construction period through a distributed terminal, and constructing a dynamic data resource pool after standardized processing; the method comprises the steps of utilizing a classification preprocessing model to clean data, extracting a construction scheme associated feature set through cosine similarity and deep learning, combining case reasoning and rule reasoning to generate an initial scheme, utilizing reinforcement learning to conduct multi-objective dynamic optimization, and finally establishing a real-time monitoring feedback mechanism through virtual simulation and test construction to verify the feasibility of the scheme, and automatically triggering scheme updating when actual data deviate from a threshold value. The method has the advantages that the method is driven by a multi-source data whole flow, and combines reinforcement learning optimization, virtual simulation verification and real-time deviation updating mechanisms through preprocessing, feature extraction and mixed reasoning generation schemes, so that the precise adaptation and dynamic optimization of the construction technical scheme are realized, and the engineering comprehensive benefit is effectively improved.

Inventors

DAI FENGHUA
LI CHAO
MA WEI

Assignees

中交第二公路工程局有限公司

Dates

Publication Date: 20260512
Application Date: 20260126

Claims (9)

1. The construction technical scheme compiling self-adaptive optimization method based on the multi-source data is characterized by comprising the following steps of: The method comprises the steps of collecting multi-source heterogeneous data in a full life cycle of a construction project through a distributed data collection terminal, wherein the multi-source heterogeneous data comprises basic environment data, engineering design data, construction resource data, historical engineering data and real-time construction monitoring data, carrying out format standardized packaging on various collected original data, and storing the data in a dynamic data resource pool; Constructing a classification preprocessing model based on the data type difference, and carrying out layering processing on the data in the dynamic data resource pool to obtain a high-quality standardized data set; calculating the association degree among different data in a standardized data set through cosine similarity, acquiring weight coefficients of each data source, extracting core association features compiled by a construction scheme through a deep learning model, and forming a construction scheme association feature set after fusion processing; Constructing a construction scheme generation model based on a mixed reasoning mechanism combining case reasoning and rule reasoning, inputting a construction scheme association feature set, and generating an initial construction technical scheme by combining the core requirements of construction projects; constructing a construction technical scheme self-adaptive optimization model based on reinforcement learning theory, establishing a multi-objective optimization function, inputting real-time update data and associated feature sets, generating a plurality of groups of optimization candidate schemes through dynamic adjustment of optimization variables, and screening an optimal construction technical scheme by adopting a hierarchical analysis method; Constructing a virtual construction simulation platform, importing an optimal construction technical scheme, establishing a simulation model based on an actual environment and design data, simulating the whole construction process, comparing real-time acquisition data with simulation results by developing small-scale test construction, and evaluating scheme suitability; In the actual execution process, real-time comparison is carried out on the real-time monitoring data and preset parameters in the optimization scheme, and when the deviation value of the monitoring data and the preset parameters exceeds a threshold value, a scheme updating mechanism is triggered to generate an updated construction technical scheme.
2. The adaptive optimization method for construction technical scheme based on multi-source data according to claim 1, wherein the steps of collecting multi-source heterogeneous data in the whole life cycle of a construction project through a distributed data collection terminal, including basic environment data, engineering design data, construction resource data, historical engineering data and real-time construction monitoring data, and storing the collected various original data into a dynamic data resource pool after format standardized packaging comprise the steps of: Distributing multi-type acquisition terminals on key nodes, equipment working surfaces and surrounding environments of a construction area, constructing a global coverage distributed acquisition network, and synchronously acquiring data; Classifying and collecting multi-source heterogeneous data, including basic environment data, engineering design data, construction resource data, historical engineering data and real-time monitoring data; And carrying out format unified packaging on various original data, storing the data in a dynamic resource pool and setting a real-time update trigger mechanism.
3. The adaptive optimization method for constructing a construction technical scheme based on multi-source data according to claim 1, wherein the constructing a classification preprocessing model based on data type differences, performing hierarchical processing on data in a dynamic data resource pool, and obtaining a high-quality standardized data set specifically comprises: Constructing a classification preprocessing model, dividing a processing level based on data types, and establishing a data type-processing strategy mapping relation; Adopting a sliding window smoothing algorithm to eliminate random noise on the basic environment data, and complementing the missing data through trend fitting; carrying out grammar checking and logic consistency checking on engineering design data, removing redundant data and correcting error items; identifying and removing abnormal data from construction resource data by adopting an abnormal value detection algorithm, and unifying dimensions by normalization processing; extracting keywords and converting the keywords and the structures of the historical engineering data by adopting a natural language processing technology; and carrying out data downsampling and time sequence alignment treatment on the real-time construction monitoring data to obtain a high-quality standardized data set.
4. The adaptive optimization method for construction technical scheme programming based on multi-source data according to claim 1, wherein the calculating of the association degree between different data in a standardized data set through cosine similarity, obtaining weight coefficients of each data source, extracting core association features programmed by construction scheme through a deep learning model, and forming a construction scheme association feature set after fusion processing specifically comprises: carrying out data source classification marking on the standardized data set, and calculating the association strength among different data sources through cosine similarity; setting weights based on the association strength of historical engineering data and cases of the current project, performing stepwise adjustment on the weights of the real-time monitoring data based on the data updating frequency, and adopting fixed reference weights on static data to form a dynamic weight distribution scheme; feature level fusion is carried out on the data with different dimensions, and core associated features compiled by a construction scheme are extracted through a deep learning model; And performing redundancy removal processing on the fused characteristic data to form a structural construction scheme associated characteristic set.
5. The adaptive optimization method for constructing a construction technical scheme based on multi-source data according to claim 1, wherein the construction scheme generating model is constructed by a mixed reasoning mechanism based on combination of case reasoning and rule reasoning, the construction scheme association feature set is input, and the initial construction technical scheme is generated by combining the core requirements of construction projects, and the method specifically comprises the following steps: Constructing a construction scheme generation model, adopting a mixed reasoning mechanism combining case reasoning and rule reasoning, retrieving a case scheme with highest similarity from historical engineering data based on core association characteristics, and extracting key construction processes, procedure arrangement and parameter setting in the case; By combining the current construction specification, industry standard and project special requirements, an inference rule base is constructed, the case scheme obtained by searching is adaptively adjusted, the construction flow, the construction method of each procedure, the equipment type selection scheme, the material use plan and the personnel division rule are determined, and an initial construction technical scheme meeting the project basic requirements is generated.
6. The adaptive optimization method for construction technical scheme based on multi-source data according to claim 1, wherein the construction technical scheme adaptive optimization model is constructed based on reinforcement learning theory, a multi-objective optimization function is established, real-time update data and associated feature sets are input, a plurality of groups of optimization candidate schemes are generated through dynamic adjustment of optimization variables, and the method for screening the optimal construction technical scheme by using analytic hierarchy process specifically comprises: Constructing a construction technical scheme self-adaptive optimization model based on reinforcement learning theory, and constructing a multi-objective optimization function taking construction quality standard reaching rate, construction period completion efficiency, cost control precision and safety risk level as optimization objectives; Taking construction process parameters, resource allocation proportion and process connection time in the initial construction technical scheme as optimization variables; based on a model intelligent decision mechanism, dynamically adjusting an optimization variable by combining real-time data feedback to generate a plurality of groups of differential candidate schemes; And comprehensively evaluating the multiple groups of candidate schemes by using an analytic hierarchy process, screening out an optimal scheme, and if the optimal scheme does not meet a preset optimization target threshold, adjusting a model optimization strategy and repeating the iterative optimization process.
7. The adaptive optimization method for construction technical scheme programming based on multi-source data according to claim 1, wherein the constructing a virtual construction simulation platform, importing an optimal construction technical scheme, establishing a simulation model based on actual environment and design data, simulating the whole construction process, comparing real-time acquisition data with simulation results by developing small-scale trial construction, and evaluating scheme suitability specifically comprises: Constructing a virtual construction simulation platform, and constructing a simulation model consistent with an actual construction scene based on the collected basic environment and engineering design data; The optimized construction technical scheme is imported into a simulation model to simulate the whole construction process, and simulation results comprising construction period simulation results, quality standard nodes and resource consumption curves are output; Selecting project procedure segments as test construction areas, collecting construction data in real time, and comparing the construction data with simulation results dimension by dimension to obtain deviation causes; and constructing an evaluation system based on technical feasibility, economic rationality and safety controllability, generating a feasibility evaluation result, and if infeasible items exist, carrying out iterative optimization again.
8. The adaptive optimization method for construction technical scheme programming based on multi-source data according to claim 1, wherein in the actual execution process, real-time comparison is performed between real-time monitoring data and preset parameters in an optimization scheme, when deviation values of the monitoring data and the preset parameters exceed a threshold value, a scheme update mechanism is triggered, and the updated construction technical scheme is generated specifically including: In the actual execution process of the optimized construction technical scheme, synchronously acquiring the process completion progress, the component construction quality detection data, the material equipment consumption data and the field environment change data; Comparing the real-time monitoring data with preset parameters of the optimization scheme dimension by dimension, and triggering an updating process when the deviation value of the monitoring data and the preset parameters exceeds a threshold value by setting a deviation threshold value; Invoking the latest data in the dynamic data resource pool, multiplexing the feature fusion model, re-extracting the features associated with the deviation, and generating an updated feature set; and inputting the updated feature set into the self-adaptive optimization model, and carrying out local parameter correction or global flow optimization aiming at the deviation type to generate an updated construction technical scheme.
9. A construction technical scheme compiling self-adaptive optimization system based on multi-source data, which is used for realizing the construction technical scheme compiling self-adaptive optimization method based on the multi-source data as set forth in any one of claims 1 to 8, and is characterized by comprising the following steps: The data acquisition and management module acquires construction full-period multi-source heterogeneous data in real time, and stores the data into the dynamic data resource pool after standardized encapsulation; The data preprocessing module is used for carrying out layering processing on the heterogeneous data based on the classification model and outputting a standardized data set; the feature fusion and weight calculation module dynamically distributes data source weights by utilizing cosine similarity, extracts core associated features of the construction scheme through deep learning and generates a structured feature set; the scheme generation module is used for generating an initial construction scheme by combining case reasoning and rule reasoning; The reinforcement learning optimization module is used for constructing a multi-objective function by taking quality, construction period, cost and safety as targets, dynamically adjusting parameters to generate candidate schemes, and screening an optimal solution by an analytic hierarchy process; the virtual simulation module simulates the whole construction process, compares and evaluates the feasibility of the scheme through test construction data, and feeds back the optimization requirement; the updating and executing module monitors construction deviation in real time, triggers a scheme updating mechanism when the construction deviation exceeds a threshold value, and generates a correction scheme by combining the latest data; The processor is used for processing the calculation process of each formula and the construction calculation process of each model.

Description

Construction technical scheme compiling self-adaptive optimization method and system based on multi-source data Technical Field The invention relates to the field of construction management, in particular to a construction technical scheme compiling self-adaptive optimization method and system based on multi-source data. Background With the development of modern engineering to complicacy and large-scale, the uncertainty of the construction environment is increased, and the fixed scheme is difficult to dynamically respond to the field change, so that the efficiency is low or the cost is hyperbranched. Meanwhile, the technology of BIM, internet of things, artificial intelligence and the like is mature, and conditions are provided for real-time data acquisition and intelligent analysis. The self-adaptive optimization method dynamically adjusts construction process, progress and resource allocation by integrating multi-source information and utilizing an algorithm, so that continuous optimization of the scheme in the implementation process is realized. The related method is difficult to realize accurate and efficient adaptation of the whole flow by compiling the construction technical scheme in the current market. Most methods still rely on manual experience driving, the data acquisition dimension is single and fragmented, and the distributed acquisition network and the standardized processing mechanism which lack global coverage cause uneven data quality, so that reliable support cannot be provided for scheme programming. The scheme generates multi-dependence single case or rule reasoning, fails to dynamically mine multi-source data core association characteristics, and has insufficient pertinence and scientificity. The optimization links are mainly static one-time adjustment, dynamic self-adaptive mechanisms based on reinforcement learning are lacked, multi-objective requirements such as quality, construction period, cost and the like are difficult to balance, and candidate scheme screening is lacked a system scientific evaluation system. Meanwhile, a closed loop updating mechanism of virtual simulation verification and real-time monitoring feedback is omitted, the problem of actual disconnection between a scheme and a site is outstanding, response is lagged when the construction environment changes or deviates, reworking, cost hyperbranched and safety risks are easy to cause, and the dynamic management requirement of complex engineering is difficult to adapt. Disclosure of Invention In order to perfect the existing method and system, a construction technical scheme based on multi-source data is provided for compiling a self-adaptive optimization method and system, the method is driven by the whole flow of the multi-source data, and the method combines reinforcement learning optimization, virtual simulation verification and real-time deviation updating mechanisms through intelligent preprocessing, feature extraction and mixed reasoning generation schemes, so that the accurate adaptation and dynamic optimization of the construction technical scheme are realized, and the engineering comprehensive benefit is effectively improved. In order to achieve the above purpose, the invention adopts the following technical scheme: The construction technical scheme compiling self-adaptive optimization method based on the multi-source data comprises the following steps: The method comprises the steps of collecting multi-source heterogeneous data in a full life cycle of a construction project through a distributed data collection terminal, wherein the multi-source heterogeneous data comprises basic environment data, engineering design data, construction resource data, historical engineering data and real-time construction monitoring data, carrying out format standardized packaging on various collected original data, and storing the data in a dynamic data resource pool; Constructing a classification preprocessing model based on the data type difference, and carrying out layering processing on the data in the dynamic data resource pool to obtain a high-quality standardized data set; calculating the association degree among different data in a standardized data set through cosine similarity, acquiring weight coefficients of each data source, extracting core association features compiled by a construction scheme through a deep learning model, and forming a construction scheme association feature set after fusion processing; Constructing a construction scheme generation model based on a mixed reasoning mechanism combining case reasoning and rule reasoning, inputting a construction scheme association feature set, and generating an initial construction technical scheme by combining the core requirements of construction projects; constructing a construction technical scheme self-adaptive optimization model based on reinforcement learning theory, establishing a multi-objective optimization function, inputting real-time update data and