TWI821019B - Intelligence driven method and system for multi-factor optimization of schedules and resource recommendations for smart construction - Google Patents

Abstract

Techniques to generate a digitally optimized schedule for a construction activity to meet a construction objective(s) of a construction project are disclosed. An artificial intelligence system receives a plurality of input data sets that impact the construction project. Each of the plurality of input data sets is processed to achieve the construction objective(s). The artificial intelligence system processes the plurality of input data sets using a respective ensemble of machine learning models. The artificial intelligence system generates machine learning validated intermediate output data sets corresponding to each of the plurality of input data sets. The artificial intelligence system implements a supervisory machine learning model to generate an optimized schedule for the construction activity based on the machine learning validated intermediate output data sets and the construction objective(s).

Description

Smart drive for multi-factor schedule optimization and resource recommendations for smart construction Dynamic methods and systems

One of the technical fields to which the present invention belongs is artificial intelligence and machine learning. The invention particularly relates to the use of machine learning to optimize scheduling, resource allocation and work sequencing for Architecture, Engineering, and Construction (AEC) planning and execution. The present invention also relates to the use of machine learning intelligence, cognition, self-learning, and a combination of trainable systems and methods to optimize resource allocation, schedule generation and management of AEC.

Architecture, Engineering and Construction (AEC) planning utilizes multiple processes and implementation methods, including the generation and management of all or part of the representative construction design, related works, and the allocation and management of labor and materials. It involves the creation of a digital avatar of the construction design, as well as the simulation of various aspects of the construction plan, such as the construction schedule, including work packages, work orders, the order and timing of required materials, procurement plans, procurement timing and sources, etc. Other factors including labor, duration, dependence on ecosystem factors, construction area topology, weather patterns, and surrounding traffic conditions are also considered. In addition, cost parameters, schedules, understanding and compliance with regulatory processes, and environmental factors, etc., are important in AEC planning. It also occupies an important position in. AEC software covers the entire construction project from design concept to execution, including post-construction activities. This type of AEC software is used by organizations and individuals responsible for building, operating and maintaining a variety of physical infrastructure, from waterways to highways and ports, houses, apartments, schools, shops, office spaces, factories, commercial buildings and more.

From planning to design to construction, AEC software is required for every step in the process. Through the use of AEC software, users can understand the relationships between buildings, building materials and other systems in various situations and try to take these relationships into consideration in the decision-making process. However, the current AEC software is just an isolated framework. Once faced with multiple input factors from local or even global events, the software cannot adapt or make decisions instantly or near-instantly to take into account the dynamic nature of the construction case.

Typical systems in the AEC space rely on manual and rules-based approaches to generate context-specific results. However, these systems cannot understand input data or dynamic differences and may not be able to provide any meaningful insights or guidance for action. In addition, these systems have no tracking mechanism for follow-up actions, and there is no way to verify whether their guidance is beneficial to users.

This problem is even more severe in the AEC field because there are many factors that affect the construction schedule. Even if you know that they will occur, it is almost impossible to predict, plan and adjust before these factors are certain to occur or may occur to some extent. Realistic.

Therefore, it is necessary to propose technical solutions to the above needs and the inefficiency of existing technologies.

This application is based on Section 119(e) of the U.S. Patent Law (35 U.S.C.) and claims the U.S. Provisional Application No. 63/280,881 "Using Human-Machine Cognition for Intelligent Construction" filed on November 18, 2021 "Methods and Systems for Multi-Factor Schedule Optimization and Resource Recommendation", the entire content of which is incorporated into this specification by reference.

The following is a summary of some embodiments disclosed in the present invention to provide a basic understanding of various aspects of the disclosed invention. The following summary is not an extensive overview of the disclosure. It is intended to present key or critical elements of the disclosure or to delineate the scope of the disclosure. Its sole purpose is to present some embodiments of the present disclosure in a simplified form as a prelude to the subsequent more detailed description of the embodiments.

In an exemplary embodiment, an artificial intelligence system is disclosed that provides and dynamically implements a digital signature of activities involved in a construction project for optimizing schedules and resource allocation to achieve a set of one or more construction goals. Achieving construction goals contributes to cost optimization, human resource optimization, task optimization, material optimization and sustainability factors, etc. The artificial intelligence system is designed to receive a plurality of input data sets and derive a plurality of first intermediate data sets (which are cognitive, statistical and/or time-based) that may have a potential impact on the construction project. The system processes each of the plurality of input data sets to achieve one or more sets of construction goals. The artificial intelligence system is further configured to process the plurality of input data sets utilizing one or more ensembles of each of inferential and cognitive machine learning models. Individual machine learning model ensembles are determined based on associating objectives with an input data set, which may be an available, predicted, or inferred data set. The artificial intelligence system is configured to generate a machine knowledge-verified decision set (e.g., a first intermediate data set) and recommendations after processing the plurality of accumulated data, being trained/acquired, and using human intelligence for accuracy verification. The artificial intelligence system is further configured to implement a supervised machine learning model to generate optimization recommendations for construction activities based on accumulated knowledge, validated learning, datasets, and pairings to meet specified construction goals.

Such input data sets may include, but are not limited to, analyzed ongoing and ex post quality metrics. Collections of machine learning models can be related to: real-time climate and weather forecasts, space astronomy Weather patterns, microclimate factors; phase relationship analysis, structural analysis; quality analysis; inventory utilization and estimates; legal compliance; labor and efficiency indicators and global event impact indicators; supply chain; equipment and Internet of Things indicators; and labor effectiveness. System cognition may be a derivative of modeling human decision-making processes, inputs, and actions under multiple scenarios and mirror simulation results. System recommendations may be evaluated for resulting efficiency and calibrated to real-world conditions.

In one aspect, the present invention provides a computer-implemented method for multi-factor schedule optimization and resource recommendations for smart construction. The method includes: receiving a schedule request for a construction activity of a construction project, the schedule request specifying at least one construction goal; accessing a plurality of input data sets; and applying a plurality of machine learning models to the plurality of input data. Each of the collections is used to generate a plurality of intermediate data sets, each of which specifies factors that affect the construction project, wherein each machine learning model ensemble in the plurality of machine learning model ensembles will be integrated according to the respective machine learning model ensemble. The functional goal is to generate an intermediate data set; based on the intermediate data set, use a chart model to determine task data; based on the task data, generate a schedule of the construction activity, which schedule will achieve the at least one construction goal; and representing the schedule in response to the schedule request. In one embodiment, the schedule is represented through at least one of a graphical user interface, voice, and messaging.

In one embodiment, the method further includes: applying a stochastic model to the plurality of intermediate data sets to determine correlation data of the factors affecting the construction project; and applying the graph model to the correlation data and tenants data to generate combined data.

In one embodiment, the method further includes: receiving feedback data on issues related to the construction project; using the feedback data to modify the stochastic model; and updating the schedule of the construction project. In one embodiment, the feedback data is received as user input via a graphical user interface. In one embodiment, the feedback data is received by receiving a real-time data from a construction site. a camera feed; determining efficiency issue data related to the construction project from the real-time camera feed; and generating the feedback data from the efficiency issue data.

In one embodiment, generating a schedule includes: determining discrepancy data between actual construction and expected construction based on the schedule; and using the discrepancy data to improve the schedule.

In one embodiment, the plurality of machine learning model ensembles includes at least one of the following: a climate analysis module configured to analyze weather patterns from at least one input data set of the plurality of input data sets; a structural analysis module a group configured to analyze quality and design from at least one input data set of the plurality of input data sets; a quality analysis module configured to analyze progress from at least one input data set of the plurality of input data sets; an inventory analysis module configured to analyze inventory from at least one input data set of the plurality of input data sets; a method group configured to analyze inventory from at least one input data set of the plurality of input data sets based on a request Centralized analysis of on-time completion; a global event impact module configured to analyze global events from at least one input data set of the plurality of input data sets; a supply chain analysis module configured to analyze global events from the plurality of input data sets; At least one input data set of the data set analyzes supply and delivery; a device health and Internet of Things (IoT) indicator analysis module configured to analyze device demand and delivery from at least one input data set of the plurality of input data sets. sourcing factors; and a labor efficiency module configured to analyze labor availability and skills from at least one input data set of the plurality of input data sets.

100: Networked computer systems/systems

102:Internet

104:Client computer

106:Server computer

112:Model integration

114:Controller

116: Notifier

118:Monitor

120:Modifier

130:Data repository

132: Data feed database

134: Construction target database

136:Model configuration database

138:Training data database

140: Suggested database

202:Climate Analysis Module

204: Structural analysis module

206:Quality analysis module

208:Inventory analysis module

210:Legal Group

212:Global event impact module

214: Supply chain analysis module

216: Device health and IoT indicator analysis module

218:Labor Efficiency Module

220:Analytical Aggregator

222:Association Analyzer

224: Combined Analytical Knowledge and Semantic Graph Models

226:Schedule Suggestor

300:Flowchart

302:Feedback

350:Flowchart

352: Performance indicator analysis module

354: Contractor Execution Module

356:Schedule difference module

400:Method

402~410: Steps

500:Computer systems/computing devices/devices

502: Input and output subsystem

504: Processor

506:Memory

508: Read-only memory

510:Storage

512:Output device

514:Input device

516:Control device

518: Communication interface

520:Network link

522:Internet

524:Host

526:Internet Service Provider (ISP)

528:Internet

530:Server

600:Software system

602A~602N: Application

610:Operating system

615: Graphical user interface

620: Naked Hardware

630:Virtual Machine Monitor

Further advantages of the present invention will become apparent by referring to the detailed description of the preferred embodiments while considering the accompanying drawings.

Figure 1 shows an example of a networked computer system that can be used to implement various embodiments of the present invention.

Figure 2A shows an example of model integration according to an embodiment of the present invention.

Figure 2B shows an example of a controller according to an embodiment of the invention.

Figures 3A and 3B respectively show an example of an artificial intelligence platform flow chart according to an embodiment of the present invention.

Figure 4A shows an example of a method for generating a dynamic schedule according to an embodiment of the present invention.

Figure 4B shows an example of a schedule according to an embodiment of the present invention.

Figure 5 shows a block diagram of a computing device in which exemplary embodiments of the invention may be implemented.

Figure 6 shows a block diagram of a basic software system for controlling the operation of a computing device according to an embodiment of the present invention.

The following embodiments are intended to assist those of ordinary skill in the art in making and using the systems and methods disclosed in the present invention. For purposes of explanation, specific details are provided herein to assist in the understanding of the invention. However, it will be apparent to one of ordinary skill in the art that these specific details are not required to practice the embodiments provided by the present invention. Application-specific embodiments are provided as representative examples only. Various modifications to the embodiments described herein will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other embodiments and applications without departing from the scope of the invention. The present invention is not limited to the disclosed embodiments, but the disclosed principles and features are to be interpreted in the broadest possible scope.

The embodiments of the present invention will be described in sections according to the following outline:

1.0 Overview

2.0 Structure Overview

3.0 Functional Overview

3.1 Model integration

3.2 Controller

3.2.1 Analysis aggregator and correlation analyzer

3.2.2 Chart model

3.2.3 Schedule Suggestor

3.2.3.1 Performance indicator analysis module

3.2.3.2 Schedule Difference Module

3.2.3.3 Contractor Execution Module

3.3 Notifier

3.4 Monitor

3.5 Modifiers

4.0 Process Example

5.0 Program Overview

6.0 Hardware Overview

7.0 Software Overview

8.0 Other aspects of the invention

1.0 Overview

The present invention discloses an artificial intelligence system that provides an electronically optimized schedule for a construction activity to meet the construction goals of a construction project. Examples of construction activities include: clearing, site preparation, excavation, foundation drilling, concrete pouring, carpentry, system modifications, etc. Examples of construction goals include: Time goals targets, budget/cost targets, quality targets, health targets, etc. The artificial intelligence system is configured to receive multiple input data sets that can affect the construction project. Each of the plurality of input data sets is processed to achieve construction goals. The artificial intelligence system is configured to process the plurality of input data sets using an ensemble of one or more different machine learning models. The ensemble of a machine learning model generates a machine learning validation data set (eg, a first intermediate data set). The artificial intelligence system is further configured to determine the task data set (eg, the second intermediate data set) from the one or more machine learning validation data sets. The artificial intelligence system is further configured to implement a supervised machine learning model to generate an optimized schedule for construction activities based on at least the task data set and construction goals.

In one embodiment, a construction target for one of the plurality of input data sets may be independent of each other of the plurality of input data sets.

In one embodiment, the artificial intelligence system is configured to receive feedback regarding an optimized schedule of construction activities. The feedback may be provided manually by the user to the artificial intelligence system through an input interface of a manual construction monitor (e.g., a graphical user interface), automatically provided by an automated construction monitor, or both.

In one embodiment, the artificial intelligence system is further configured to determine the difference between the actual construction progress and the expected construction progress based on the system recommendations and guidance, and formulate further schedules and resource allocations for the construction activities based on the determined differences. Improve details. A revised plan will take this difference into account and look ahead to the next scheduled number of weeks (e.g., six weeks) to review and see if any additional factors need to be considered given the current difference.

In one embodiment, the artificial intelligence system is further configured to implement an integration of closed-loop performance analysis machine learning models that can leverage new training data (including feedback) to self-generate Adjustment and/or self-training. The closed-loop performance analysis machine learning model may include one or all of the machine learning models described herein.

Other embodiments, aspects, and features will become apparent from the remainder of this disclosure.

2.0 Structure Overview

Figure 1 illustrates a networked computer system 100 that may be used to implement various embodiments of the present invention. For clear explanation of the example, Figure 1 shows a simplified schematic diagram, and other embodiments may include more, fewer or different components. Figure 1 and the other drawings, together with all disclosures and claims of the present invention, are intended to present, disclose and seek to protect a technical system and various technical methods, including specially programmed computers, which utilize A system design and instructions for a special purpose distributed computer programmed to perform the functions described. These components can perform functions that were not feasible in the past and provide a practical application of computing technology to solve schedule optimization, resource allocation and work sequencing issues related to AEC planning and execution. In this way, the present invention provides a technical solution to a technical problem; any interpretation that deems the disclosure or claims of the present invention to be ineligible for patents, such as treating the present invention as an abstract concept, mental process, or organization of human beings The activity methods or mathematical algorithms are not supported by this manual and are incorrect interpretations.

In some embodiments, the networked computer system 100 includes: a client computer 104, a server computer 106, and a data repository 130. The above components are directly or indirectly connected through a network 102. In one embodiment, the server computer 106 broadly represents one or more computers, such as one or more desktop computers, server computers, a server farm, a cloud computing platform (such as Amazon EC2, Google Cloud , container orchestration (Kubernetes, Docker, etc.)), or a parallel computing computer, a virtual computing instance in a public or private data center, and/or an instance based on a server application.

The server computer 106 includes one or more computer programs or sequences of program instructions written to implement artificial intelligence/machine learning algorithms to generate considerations, control functions, notifications regarding a construction project Information about functions, monitoring functions and modifying functions. The program or instruction sequence written in this way to implement the consideration data generation function may be referred to as a model integration 112 in this article; the program or instruction sequence written in this way to implement the control function may be referred to as an optimization project and Construction scheduler supervisory controller 114; a program or instruction sequence written in this manner to implement a notification function may be referred to herein as a notifier 116; a program or instruction sequence written in this manner to implement a monitoring function may be referred to herein as It is called a performance analysis and process monitor 118; the program or instruction sequence written in this way to implement the modification function may be called a modifier 120 herein.

Model integration 112 , controller 114 , notifier 116 , monitor 118 and/or modifier 120 may be part of an artificial intelligence (AI) platform implemented on server computer 106 . In some embodiments, one or more components of the server computer 106 may include a processor configured to execute instructions stored on non-transitory computer-readable media.

In some embodiments, the model integration 112 includes a plurality of modules, each module programmed to receive a data feed and determine a data feed based on relevant sections of the data feed related to the functional goals of the respective module. An intermediate data set containing considerations in a construction project. The data feed includes metadata or tags for the initial portion of the data feed that identify the segment and corresponding data type. Alternatively, the metadata or tags may be mixed in the data feed. For example, each data section may include metadata indicating the data type to which the data section belongs. If the data type is consistent with the functional goal of each relevant module, each relevant module will process the data section. For example, a microclimate analysis module will only process data sections of a data feed that are relevant to the functional goals of the microclimate analysis module (eg, weather analysis, prediction, determination, recommendations, etc.). In other words, this microclimate analysis The module identifies and processes data segments that have metadata indicating the type of weather data. If a data feed only includes weather data, the microclimate analysis module will process the entire data feed. If a data feed does not include any weather data, the microclimate analysis module will not process the data feed.

In some embodiments, the controller 114 is programmed to generate an optimized schedule for the construction project based on factors determined by the model ensemble 112 (eg, the first intermediate data set). The optimization schedule may be further optimized or updated based on feedback from monitor 118 .

In some embodiments, notifier 116 is programmed to communicate notifications to the user, such as an optimized schedule and other suggestions. For example, the notification may be presented in a user interface (eg, a graphical user interface) or by voice or text message. Notifier 116 may receive such notifications from controller 114 and data repository 130 .

In some embodiments, monitor 118 is programmed to receive feedback that can be used to make corrections and modifications at controller 114 (eg, at an analytics aggregator of controller 114). Example feedback may provide feedback on construction status, delays, and other topics and issues manually through an input interface (eg, a graphical user interface), or automatically through monitor 118 . Monitor 118 is also programmed to receive data feeds from one or more external sources, such as field sensors or video, and store the data feeds in data repository 130 .

In some embodiments, the modifier 120 is programmed to receive modification data to update existing artificial intelligence models in the system 100 and to add new artificial intelligence models to the system 100 . Modification data may be input by the user via an input interface (eg, graphical user interface).

In some embodiments, in order to comply with sound software engineering principles of modularization and separation of functions, the model integration 112 , the controller 114 , the notifier 116 , the monitor 118 and the modifier 120 are each implemented as logically independent programs. , program or library.

The computer-executable instructions described herein may be machine executable code of a CPU instruction set, and may have been based on original code written in Python, JAVA, C, C++, OBJECTIVE-C, or any other human-readable programming language or environment. Code is compiled and can be written independently or in conjunction with scripts such as JAVASCRIPT, other scripting languages, and other program source text. In another embodiment, the program instructions may also represent one or more files or source code projects that are digitally stored on a mass storage device (e.g., a non-volatile storage device) within the system shown in Figure 1 or an independent storage system. Random access memory or disk storage), executable instructions will be generated when compiling or interpreting, and when these executable instructions are executed, they will cause the computer to refer to these instructions to perform the functions described in this article or operate. In other words, Figure 1 may represent the way a programmer or software developer organizes and arranges source code for future compilation into an executable file or interpretation into bytecode or its equivalent for execution by the server computer 106 .

The server computer 106 may be coupled to a data repository 130 that includes a data feed database 132, a construction target database 134, a model configuration database 136, a training data database 138, and a recommendations Database 140.

In some embodiments, the data feed database 132 stores a plurality of data feeds collected from various sources (eg, an engineering website or AEC website, third-party paid or commercial databases, and real-time feeds such as RSS, etc.) . A data feed may include data segments related to real-time climate and weather forecast data, structural analysis data, in-construction and post-construction data, such as modular analysis of quality data, inventory utilization and forecast data, regulatory data, global event impact data , supply chain analysis data, equipment and IoT indicator analysis data, labor/efficiency data, and/or other data provided to the model integration 112 according to each construction goal. A data feed may include tenant information about other activities on the construction project, other construction projects, or both. Each data section may include a metadata indicating the data type to which the data section belongs. material. As described herein, real-time data, near-real-time data, and collated data are received by the monitor 118 and processed by various components of the server computer 106 .

For example, a data feed received by the monitor 118 is stored in the data feed database 132 and processed by different modules of the model integration 112. Each module generates a first intermediate data set through The controller 114 is provided within the artificial intelligence platform to generate dynamic schedules. For example, another data feed received by the monitor 118 may be stored in the data feed database 132 and used by the controller 114 (e.g., the schedule suggester 226 of the controller 114 shown in FIGS. 3A and 3B ) processing to determine the difference between actual construction and expected construction to further optimize the schedule. For example, further data feeds received by the monitor 118 may be stored in the data feed database 132 and analyzed by the controller 114 (e.g., the combined knowledge and semantic graph of the controller 114 shown in FIGS. 3A and 3B 224) processing to access the impact on factors related to construction projects or construction activities.

In some embodiments, construction target database 134 includes a plurality of construction targets. Each of the plurality of construction goals in the data feed database 132 is processed to achieve one or more of the plurality of construction goals in the construction goal database 134 . Construction goals are a collection of different user requirements, project requirements, regulatory requirements, technical requirements, etc. Construction objectives can be established before construction activities begin and can be adjusted during the construction phase to account for changing conditions. Construction objectives are defined for each construction project and construction phase. The data definition of construction goals will be used to define standardized construction goals. Data definitions include, for example, parameters for optimizing construction schedules to meet time targets, parameters for optimizing cost targets, parameters for optimizing carbon footprint targets, which are normalized to account for worker health, minimizing on-site workers, and minimize quality issues. One or more construction goals may be identified as part of a schedule request for construction activities for a construction project.

Please refer to the following scenario description. Consider a scenario where the construction goal is to keep costs within budget amounts. Data feeds from the construction phase feed into the data collected and generated in the system. The system checks data feeds against established goals and optimizes guidance to align with established goals. For example, if an incoming data feed indicates that the construction completion date may exceed the given end date, system analysis can indicate through the smart agent: add additional construction labor and purchase materials from nearby suppliers at a higher cost to transport them. Time is reduced to a minimum, which can help achieve the target completion date. Since the desired goal is to keep costs below the allocated budget, the system's recommendation will be to continue working on the current task at the current pace. The system's recommendations will confirm the construction target database 134 and all legal commitments before giving recommendations. In another different scenario, if the construction goal is to comply with the set construction completion date, the system suggestions may be: adding additional construction manpower, purchasing materials from nearby suppliers, etc. Details of system recommendations are discussed further below.

In some embodiments, the model configuration database 136 includes configuration data such as parameters, gradients, weights, biases, and/or other attributes that are required to enable the artificial intelligence model to operate after the training of the artificial intelligence model is completed.

In some embodiments, training data database 138 includes training data used to train one or more artificial intelligence models of system 100 . The training data database 138 is continuously updated with more training data obtained from the system 100 and/or external sources. The training data includes historical customer data as well as synthetic algorithm-generated data specifically designed to test the performance of the different artificial intelligence models described in this article. This comprehensive profile can be compiled to test multiple system performance factors. It may include false positive (false positive) and false negative (false negative) recommendation rates, model elasticity, and model recommendation accuracy indicators. An example of a training data set may include data about a contractor completing a task ahead of an estimated schedule. Another example of a training data set could include data related to quality issues in work performed by the contractor on the same task.

In some embodiments, the advice library 140 includes advice material, such as an optimization schedule generated by the controller 114 . An example of a construction schedule includes all tasks in order of priority, grouped tasks (also called work packages), and assigned resources to the tasks. Schedule optimization includes: the shortest path to achieve construction goals; optional work packages include: supplier recommendations (based on proximity, quality and cost) and contractor recommendations. As discussed in this article, schedule optimization is dynamic and takes into account current schedule progress, expected impedance, and the impact of quality issues and supply constraints.

The data repository 130 may include other databases that store data that can be used by the server computer 106 . Each database 132, 134, 136, 138, and 140 may be implemented using memory such as RAM, EEPROM, flash memory, hard drive, optical drive, solid state memory, or any type of memory suitable for database storage. memory.

Network 102 broadly represents one or more local area networks (LANs), wide area networks (WANs), metropolitan networks (MANs), global networks (such as public networks), or a combination of the foregoing. Each of the aforementioned networks may use or execute stored programs that implement Internet protocols according to standards such as the Open Systems Interconnection (OSI) multi-layer networking model, including but not limited to Transmission Control Protocol (TCP) or User Data Element Protocol (UDP), Internet Protocol (IP), Hypertext Transfer Protocol (HTTP), etc. All computers described herein can be configured to connect to the network 102, and the present disclosure assumes that all components of Figure 1 can be communicatively coupled to the network 102. The components shown in Figure 1 may also communicate with each other through direct communication lines (not shown for simplicity).

The server computer 106 can access the client computer 104 via the network 102 to request a schedule or a resource suggestion. Client computer 104 may include a desktop computer, laptop computer, tablet computer, smartphone, or any other type of computing device that allows access to server computer 106 . The group shown in Figure 1 The component parts are intended to represent one possible embodiment and are not intended to limit or limit the number of component parts for other embodiments.

3.0 Functional Overview

The server computer 106 includes a model integration 112, a controller 114, a notifier 116, a monitor 118, a modifier 120, and a data repository 130. The above components will operate with each other in non-traditional ways through programming according to usage requirements. , to generate schedules and resource recommendations.

3.1 Model integration

In one embodiment, the model integration 112 of Figure 1 includes a plurality of modules that communicate with each other. Each module in the plurality of modules may include an ensemble of one or more machine learning models.

The machine learning model may include appropriate classifiers and machine learning methods. Some machine learning algorithms include: (1) Multilayer Perceptron, Support Vector Machines, Bayesian learning, K-Nearest Neighbor as part of supervised learning ) or Naïve Bayes classifier; (2) Generative Adversarial Networks as part of semi-supervised learning; (3) Using autoencoders, Gaussian Mixture and K-means Unsupervised learning of K-means clustering algorithm; and (4) reinforcement learning (for example, using Q-learning algorithm, using temporal difference learning), and other suitable learning methods. Knowledge transfer is also used. In addition, in order to optimize the resources of the machine learning model, for small carbon footprint equipment, the binarization and quantification of the model can be performed. Each of the plurality of modules may implement one or more of the following methods: regression algorithms (e.g., ordinary least squares, logistic regression, stepwise regression, multivariate adaptive regression splines, local scatter Point smoothing estimation [locally estimated scatterplot smoothing], etc.); instance-based methods (e.g., K nearest neighbor, learning vector quantization, self-organizing map, etc.); regularization methods (e.g., ridge regression, least absolute compression and selection operator/Lasso algorithm, elastic network, etc.); decision tree learning methods (e.g., classification and regression Trees, ID3 [iterative dichotomiser 3] algorithm, C4.5, chi-square automatic interaction detection, multivariate adaptive cloud-shaped regression, gradient boosting machine, etc.); Bayesian learning methods (e.g., simple Bayesian, average single-dependence predictor, Bayesian belief network, etc.); kernel methods (such as support vector machines, radial basis functions, linear difference analysis, etc.); clustering methods (such as K-means clustering, expectation maximization, etc.); correlation rule learning algorithms (For example, Eclat algorithm, etc.); Artificial neural network model (for example, Perceptron method, back propagation method, self-organizing mapping method, learning vector quantization method, etc.); Deep learning algorithm (for example, restricted Boltzmann machine, deep belief network method, convolutional network method, stacked automatic decoder method, etc.); dimensionality reduction methods (such as principal component analysis, partial least squares regression, multidimensional scaling, etc.). Each processing portion of the system 100 may additionally utilize probabilistic, heuristic, deterministic, or other suitable methods for computational guidance, recommendations, machine learning, or combinations thereof. However, all suitable machine learning methods may additionally be incorporated into system 100. Furthermore, all suitable models (eg, machine learning, non-machine learning, etc.) may be used in the system 100 .

The model integration 112 receives input data from the data repository 130 . For example, the model ensemble 112 receives one or more data feeds from the data feed database 132 and predicts or otherwise generates a first intermediate data set for each data feed for further processing by the controller 114 .

As shown in Figure 2A, in some embodiments, the model integration 112 may include a climate analysis module 202, a structural analysis module 204, a quality analysis module 206, an inventory analysis module 208, a method scale Group 210, a global event impact module 212, a supply chain analysis module 214, an equipment health and IoT indicator analysis module 216, and a labor efficiency module 218. Each module of the model ensemble 112 may include an ensemble of one or more machine learning models.

The climate analysis module 202 of the model integration 112 receives one or more input data sets for analysis of local historical weather patterns, microclimate analysis, predictive weather analysis, stochastic modeling of events, establishment and supply storage, and supply availability. properties, correlation to shipping, correlation to working conditions, correlation to materials and material behavior (e.g. curing of weathered cement), as well as analysis of drainage, wind, sand, etc. The machine learning model of the climate analysis module 202 provides immediate climate factors as well as climate predictions for the next few weeks, and so on. Such input data sets may also include local historical weather/climate patterns during the construction phase. As described herein, the monitor 118 of Figure 1 assists the climate analysis module 202 in collecting relevant input data sets. The climate analysis module 202 can dynamically receive these input data and provide optimized suggestions for weather-related events.

The structural analysis module 204 of the model integration 112 receives one or more input data sets for analyzing quality and design considerations, structural integrity, etc. throughout the construction phase. For example, such data sets may be obtained through simulation and/or specific detection, etc. These input data sets are taken into consideration for construction updates and recommendations for any task redoing or rescheduling are provided as needed. Any impact on labor, cost, time, materials and schedule due to problems (such as quality issues) will also be further considered to optimize the schedule. The structural analysis module 204 will also provide information on structural condition, monitor work progress, receive visual input data of construction progress, evaluate automatic input data (e.g., from computer vision input) and determine task/work progress to predict/estimate A completion percentage (for example, approximately 40% or 60% complete, etc.). For example, although a user input may indicate that a task is completed, the structural analysis module 204 may determine the actual status of the task based on multiple input parameters and conclude that certain aspects of the task are not yet complete or need to be redone. . As described herein, the monitor 118 may assist the structural analysis module 204 in collecting relevant input data sets.

The quality analysis module 206 of the model integration 112 will receive one or more input data sets for user input data, computer imaging (e.g., video feed, still image capture, infrared, lidar technology, laser, etc.) Analyze progress, conduct quality assessment through manual and/or machine inspections to determine the cost of poor construction (material cost, time, labor, etc.), and allocate time and labor, etc. As described herein, the monitor 118 may assist the quality analysis module 206 in collecting relevant input data sets. The costs of shoddy construction can involve lost revenue (the inability to deliver construction to the agreed timetable and to the agreed quality), redo costs, additional labor and material costs, and potential legal and regulatory compliance issues.

The inventory analysis module 208 of the model integration 112 receives one or more input data sets for analyzing real-time inventory and making forecasts. The machine learning model of the inventory analysis module 208 determines how much inventory is needed based on available inventory, utilization, etc., to predict inventory demand. For example, a particular task is thought to require approximately 500 steel beams, but use may be slow, largely because construction for the task is still lagging behind. Therefore, the inventory analysis module 208 will analyze the input data to determine whether to recommend purchasing or defer purchasing. As described herein, the monitor 118 may assist the inventory analysis module 208 in collecting relevant input data sets.

The legal group 210 of the model integration 112 receives one or more input data sets for analyzing the completion schedule based on legal contracts, regulatory requirements, etc. For example, certain tasks may have contractual performance obligations. There may be penalties for delays, or there may be local or federal regulations that may affect construction, such as requirements regarding emission standards, flood control agreements, fire codes, etc. The legal team 210 will use natural language processing (NLP) algorithms to analyze various contracts, applicable regulations and other regulations to monitor construction progress and ensure compliance with regulations. As described herein, the monitor 118 may assist the legal team 210 in collecting relevant input data sets.

The global event impact module 212 of the model integration 112 will receive one or more input data sets for analyzing and tracking the impact of simple and complex global events, such as geopolitical events, supply Chain impacts, global weather, accidents at manufacturing plants, shipping route issues, market factors (such as steel price futures, cement futures, etc.) to plan procurement and schedule/task revisions as needed. The module will also monitor and analyze geopolitical events that may impact business, material sources, or supply chain factors affecting material procurement. The global event impact module 212 monitors such events and automatically re-adjusts, re-sequences and re-considers construction schedules and related activities accordingly. The global event impact module 212 analyzes simple and complex events and modifies the schedule accordingly. As described herein, the monitor 118 may assist the global event impact module 212 in collecting relevant input data sets.

The supply chain analysis module 214 of the model integration 112 will receive one or more input data sets for analyzing task maps (task/schedule lists), supplier analysis, estimated procurement time, historical performance analysis, time value analysis ( From ordering to delivery, such as cost, time, carbon emission considerations, etc.), and planning procurement based on current progress and forecast tasks, on-site/off-site storage factors, etc. This can also be a factor in impact analysis of global events. The supply chain analysis module 214 determines the best place to purchase materials and the best time to purchase them. These locations and times can be optimized to meet construction objectives such as cost, sustainability, environmental objectives and commercial objectives. As described herein, the monitor 118 may assist the supply chain analysis module 214 in collecting relevant input data sets.

The device health and IoT metrics analysis module 216 of the model integration 112 receives one or more input data sets to analyze: device demand and procurement factors (e.g., how long it takes to purchase the device, the source of the purchase, proximity analysis, etc.); equipment quality factors (e.g., field quality issues, operators needed, historical issue analysis, etc.); cost and carbon emissions considerations; IoT feeds for fuel consumption and equipment health (e.g., real-time equipment health analysis); Alternative path mapping (e.g., what to do if equipment fails, what are the sources of alternatives - should be considered when planning and selecting equipment suppliers/leasing), etc. The device health and IoT indicator analysis module 216 will analyze or monitor the provided signal Different equipment, for example, the diesel content in earth-moving equipment is low, so more diesel must be purchased to keep the equipment running; according to vibration or acoustic analysis, a machine is on the verge of failure, etc. The device health and IoT indicator analysis module 216 will analyze these input data and determine its impact on the schedule, or alert the user to take necessary remedial measures before the schedule is affected. As described herein, the monitor 118 may assist the device health and IoT metrics analysis module 216 in collecting relevant input data sets.

The labor efficiency module 218 of the model integration 112 will receive one or more input data sets for analysis: task map (task/schedule list); current labor availability and skills analysis, performance analysis (e.g., completion speed, quality indicators, etc.); historical trends (e.g., past efficiency, notes, comments, cost analysis of poor work quality, etc.); current work efficiency and outlook, labor health factors (e.g., work-related injuries, overtime, etc.); expected labor demand/geographical considerations ; Macro factors (for example, the availability period of the construction project); Micro factors (for example, recent demand/approaching deadline), etc. The labor efficiency module 218 will further determine efficiency indicators based on considerations such as labor issues, construction speed, quality issues, etc. The labor efficiency module 218 provides suggestions for adjusting tasks based on efficiency analysis. As described herein, the monitor 118 may assist the labor efficiency module 218 in collecting relevant input data sets.

The above external data sources, modules, models or programs are illustrated by examples. Different business needs may require the addition or modification of different modules or data sets for specific functions. Each mod works independently and is optimized for each mod which is encouraged to optimize its parts. However, the controller 114 communicates with each module and serves as a supervisory controller algorithm that combines all module output data (eg, the first intermediate data set) and optimizes the overall schedule to achieve the desired construction goals.

3.2 Controller

As shown in Figure 2B, in one embodiment, the optimization engineering and construction scheduler supervision controller 114 of Figure 1 includes an analysis aggregator 220, a correlation analyzer 222, a combined analysis knowledge and semantic graph model 224, and a schedule suggester 226.

3.2.1 Analysis aggregators and correlation analyzers

The analysis aggregator 220 analyzes information, such as the first intermediate data set from the model ensemble 112, to determine a subset of the first intermediate data set related to construction goals and construction activities, and uses self-learning multi-factor prediction to stochastically Analysis to correlate this subset of data with the current construction situation (e.g., ground truth). In some embodiments, analysis aggregator 220 determines from the first intermediate data set what data is relevant to the construction goal, and utilizes correlation analyzer 222 to make various output data of the machine learning model (e.g., what data analysis aggregator 220 determines to be relevant and The output data subset) is associated with the current construction situation.

3.2.2 Chart model

The correlation data input output by the correlation analyzer 222 is input to the combined analytical knowledge and semantic graph model 224 . The combined analytical knowledge and semantic graph model 224 utilizes image representations in the form of knowledge graphs and social graphs to generate weighted tasks. These knowledge and social graphs have multiple nodes that perform specific functions and analyze data sets, such as tenant data sets (including data from another activity on a construction project, and/or data from another construction project), to evaluate the impact of The impact of factors related to current construction projects or construction activities. For example, will a global event actually impact the schedule, and if so, how imminent is this impact, will it affect a single task or an entire work package or an entire construction site, etc.

3.2.3 Schedule Suggestor

The combined analytical knowledge and semantic graph model 224 provides analyzed output data to the schedule suggester 226, including weighted tasks (e.g., tasks with weighted values, the weighted values refer to indicates priority). The schedule suggester 226 will at least suggest a dynamic schedule from the weighted tasks as output data (eg, to the user).

In one embodiment, schedule suggester 226 may include a performance metric analysis module, a schedule variance module, and a contractor execution module; these modules individually or collectively implement a supervised machine learning model to Improve your schedule.

3.2.3.1 Performance indicator analysis module

Performance indicators use the output of related programs (for example, analysis aggregator 220, correlation analyzer 222, combined analysis knowledge and semantic graph model 224, etc.) to perform calculation classification, evaluation, and inference. Related programs include analysis aggregation, past and present Creation of data correlations and inferences from insights (knowledge accumulated for comparative analysis).

Algorithms used may include decision trees, decision diagrams, linear discriminant analysis and mathematical imputation. Mathematical interpolation may involve calculating a schedule's performance index, which is a measure of the historical and current schedule, and comparing it to the current task progress percentage and task severity. Performance indicators are an aspect of the system that help improve the efficiency of a construction process.

Supervised model evaluation can involve the synthesis and analysis of simulation and implementation scenarios, as well as the evaluation of different data and model characteristics. Feature weights can vary dimensionally to support model execution. When proposing multiple paths (schedules and tasks) to achieve an outcome, a model can highlight a desired performance outcome and use simulation to feed the decision diagram. Discriminant analysis improves predictions that take the current situation into account. Different feature weights will take into account factors such as seasonality, on-site workers, injury simulation, quality issues, supply constraints, etc.

3.2.3.2 Schedule Difference Module

A schedule variance analyzer utilizes a baseline schedule derived from comparable past schedules and manual input. Comparative metrics included in the baseline schedule include topography, size, value, Involving materials, environmental factors, etc. In order to find comparable past schedules, the system can use deep neural network algorithms, including K-nearest neighbor (K-NN) and convolutional neural networks (CNN), to build specific Confidential and interpretable model output to develop a baseline schedule.

Once a baseline schedule is established, the current schedule combination is continuously evaluated at frequent intervals to determine deviations from the baseline schedule. Expectations for differences can be factored into tolerance considerations. When tolerance levels are exceeded, the decision model ensemble attempts to develop course corrections to meet construction goals.

Model calculations may involve computational evaluation from different models in the ensemble. Algorithms utilized may include the traveling salesman problem (TSP), distance-dependent Hamiltonian cycles, and task dependency pairing.

In one embodiment, when the difference in current task progress significantly deviates from the baseline schedule and the model prediction cannot meet the expected project results, task reorganization will be performed and a decision diagram will be used to perform end-state simulation to verify that the expected project results are met. The best path to results.

3.2.3.3 Contractor Execution Module

The Contractor Execution Module is a recommendation system that evaluates the effectiveness of construction vendors and contractors to make vendor/contractor selection recommendations for desired tasks. Additionally, the recommendation system tracks current contractor output and progress metrics for consideration in its recommendations. When a desired contractor is paired with a set of tasks, the recommendation system recommends the best path for task execution based on ground-truth calculations of field metrics and schedule derivatives. Ground truth calculations are achieved by internally reviewing data inputs including observation reports, logs, quality analysis, visual analysis and speech-based annotations. The example models and algorithms can utilize latent semantic analysis (LSA), related topic modeling (correlated Topic Modeling (CTM), Named Entity Recognition (NER) and knowledge graph. Neural networks employing convolution and reinforcement learning can be used to handle image-based reasoning.

Generative Adversarial Network (GAN) can be used to evaluate schedule optimization and schedule suggestions for championship challenger scenarios. In essence, the data section and expected results of the current construction project will be considered together, and GAN will establish two premises: (1) If the difference and progress tolerance are within the acceptable range, continue the current plan and take efficiency surprises into account, and (2) change recommendations and reorganize task sets and take probabilities into account for future progress scenarios.

Both paths will be analyzed and recommendations made, and a confidence level will be provided for each recommendation. Recommendations made by the Contractor Execution Module may include: rearranging a set of tasks, substituting materials, substituting suppliers, substituting contractors, or a combination of the above.

3.3 Notifier

In one embodiment, the notifier 116 of FIG. 1 outputs warnings, notifications, and suggestions (eg, schedules generated by the schedule suggester 226 of the controller 114). For example, the output of suggestions can be achieved through an output interface (such as a graphical user interface), voice, multimedia messages, and/or other communication forms.

3.4 Monitor

In one embodiment, the monitor 118 of Figure 1 receives feedback regarding proposals and/or construction projects. Examples of feedback include manual or direct feedback input by a user, as well as automatic or indirect feedback determined by server computer 106 (eg, by monitor 118). Based on the feedback, the controller 114 generates a further optimized schedule (updated/revised optimization schedule) for the construction project. Examples of feedback include recommendations that were or were not implemented during construction.

In one embodiment, the monitor 118 assists modules of the model integration 112 in collecting input data sets (eg, data feeds) from one or more external sources, storing the input data sets in the data feed database 132 , and storing the input data sets in the data feed database 132 . The input data set is transferred from the data feed database 132 to the various components of the system 100 , such as the various modules of the model integration 112 or the controller 114 .

In one embodiment, monitor 118 also receives machine learning training data from a plurality of external data sources for model ensemble 112 . The plurality of external data sources, models or programs preferably include several different sources of input data sets that can be used to train the machine learning model of the system 100 . The trained machine learning modules in the plurality of modules (including climate analysis module 202, structural analysis module 204, quality analysis module 206, inventory analysis module 208, legal group 210, global event impact module 212. Supply chain analysis module 214, equipment health and IoT network indicator analysis module 216, labor efficiency module 218, and other modules) will provide the first intermediate data set to the controller 114 for use in dynamic schedules generate. As described herein, one or more of the modules of model ensemble 112 may include deep neural network-based models. Output data from these modules is provided to the controller 114 to generate an optimal schedule by allowing the controller 114 to continuously track and monitor multiple parameters of the model ensemble 112 .

As described herein, model ensemble 112, controller 114, and monitor 118 may implement an ensemble of closed-loop performance analysis machine learning models. In some embodiments, feedback received/generated at monitor 118 may be used to adjust and retrain the artificial intelligence model used by system 100.

3.5 Modifiers

In one embodiment, the modifier 120 in Figure 1 receives user input data and configures one or more machine learning models or adds new machine learning models by using a management console. The management console may serve as an input interface (eg, a graphical user interface) for communicating with or configuring one or more components of server computer 106 . The tube The management console may be implemented by modifier 120, or as one or more private or public computer servers, and/or as a component of a distributed computing system or cloud.

4.0 Process Example

Figure 3A shows an example flow diagram 300 of an artificial intelligence platform according to an embodiment.

As described herein, the artificial intelligence platform includes model integration 112, controller 114, notifier 116, monitor 118, and modifier 120. It should be noted that not all components of the artificial intelligence platform are shown in Figure 3A. In one embodiment, the artificial intelligence platform dynamically generates a schedule for AEC planning and execution. This artificial intelligence platform can use multi-factor predictive and stochastic analysis and machine learning models to dynamically generate schedules for AEC planning and execution. The artificial intelligence platform provides functions such as schedule composition, schedule dynamic monitoring, and schedule optimization to achieve specific construction goals.

Model integration 112 has one or more modules that include several different sources of input data sets. For example, a data set may be used as input data for training a machine learning model of the model ensemble 112 . Modules include climate analysis module 202, structural analysis module 204, quality analysis module 206, inventory analysis module 208, legal group 210, global event impact module 212, supply chain analysis module 214, equipment health and material Networked indicator analysis module 216, labor efficiency module 218, etc., these modules provide the first intermediate data set to the controller 114 for generating a dynamic schedule.

Each module of model ensemble 112 may be an ensemble of one or more models. For example, the climate analysis module 202 may not be a single model, but may include multiple models: one model may view the astronomical calendar, another model may view the current weather conditions or microclimate conditions, and yet another model may view the forecast or predicted weather conditions. Weather conditions, etc. Each model of the climate analysis module 202 will generate output data based on its inherent model (eg, the functional goal of the climate analysis module 202). For example, each model targets a specific function It works with functional goals and optimizes its output for its own focused functional goals. The output data from the models are then combined to form an input data provided to the schedule suggester 226 .

As another example, the inventory analysis module 208 may use a combination of visual input, enterprise resource planning (ERP) system analysis, construction speed and material utilization, etc., and generate an output based on analyzing these input data. Each input to the inventory analysis module 208 can be analyzed using its own ensemble of models with the appropriate weighting necessary to achieve the specific/respective construction goals. Within each module, multiple functions, factors, etc. can be received based on business and/or construction goals.

The artificial intelligence platform for dynamic schedule generation further receives dispersed, independent, and different first intermediate data sets from the modules as input data, and processes them via the analysis aggregator 220 . Each module of the model integration 112 has a defined set of requirements, such as input data and/or output data. The analytics aggregator 220 uses its self-learning multi-factor prediction and stochastic analysis to analyze and correlate information. The analysis aggregator 220 further uses a correlation analyzer 222 to correlate the various outputs of the modules and provide the correlation data to the combined analytical knowledge and semantic graph model 224. The combined analytical knowledge and semantic graph model 224 utilizes graphical representations in the form of knowledge graphs and social graphs. These graphs have multiple nodes and perform specific functions, and the data set is analyzed to assess the impact of other factors. For example, the combined analytical knowledge and semantic diagram model 224 may leverage different tenant knowledge (eg, knowledge from other construction activities, other construction projects) and apply such knowledge to current construction activities. The combined analytic knowledge and semantic graph model 224 will provide a second intermediate data set including weighted tasks as analytic output data to the schedule suggester 226 . The schedule suggester 226 will use the weighted tasks to generate an optimized schedule, which will be output to the user by the notifier 116 . The schedule can be output through a user interface, voice, text message and/or similar means.

Monitor 118 receives user or automated feedback 302 and forwards the feedback to analysis aggregator 220, thereby creating a closed performance analysis loop. The system 100 can take any business changes or new requirements into account. For example, feedback 302 may be generated as optimization recommendations for a specific construction activity based on accumulated knowledge and validated learning and data sets, and paired to meet construction goals. Users can manually input relevant feedback on a specific issue, problem, etc. into the system through the user interface using a manual construction monitor. Monitor 118 may also use an automated construction monitor to automatically calculate or otherwise determine construction status, delays, problems, local or global issues, schedules, and other conditions that may cause delays, halts, or disrupt construction goals or schedules. Efficiency issues and provide them as feedback 302 to further enhance, correct, or flag current issues.

For example, a camera can be used on a construction site to record daily work, construction speed and progress. The automated construction monitor intelligently matches camera recordings to one or more work plans, highlighting any issues and delays, as well as confirming any discrepancies with user interface-based input to determine whether to send Check whether there are any errors or incorrect information in the information fed into the activity tracker compared to actual work on the construction site.

The controller 114 may receive feedback 302 regarding an optimized schedule of construction activities based on the manual or automated feedback. As mentioned above, the artificial intelligence platform is configured to implement an integration of closed-loop performance analysis machine learning models to generate a further optimized construction activity schedule. In one example, users provide feedback to the artificial intelligence platform through a user interface. In another example, feedback is provided automatically by a construction monitor.

Figure 3B illustrates an example flowchart 350 of an artificial intelligence platform according to an embodiment of the present invention. The steps shown in Figure 3B are performed in a manner similar to the corresponding steps in the context of Figure 3A. Therefore, for the sake of brevity, the steps shown in Figure 3B will not be repeated.

In one embodiment, schedule suggester 226 provides suggestions, and monitor 118 evaluates whether the suggestions are relevant and effective in optimizing schedules, cost factors, etc. Schedule suggester 226 may include a gating mechanism to evaluate its own suggestions and its own effectiveness. For example, schedule suggester 226 may include a performance indicator analysis module 352, a contractor execution module 354, and a schedule variance module 356.

The performance index analysis module 352 checks and verifies whether the performance index is achieved. The contractor execution module 354 checks whether the user has performed the recommendations, for example, whether the task sequence provided by the system 100 has been performed/followed. The schedule discrepancy module 356 checks whether there are discrepancies between the schedule provided by the system and the actual field work. On-site data can be received from multiple sources and can include manual user input through mobile applications, as well as information feeds obtained through IoT device sensors, such as motion sensors, altimeter readings, pressure Sensors, temperature sensors, quantity measurement through object counting, camera feeds, motion sensors, etc. As data is generated and transmitted, the data is also received instantly and near-instantly, such as by the monitor 118 .

In one embodiment, the system 100 performs self-analysis, analyzes both the combined efficiency and the evaluation and efficiency of individual models, and performs self-learning from the feedback received. System 100 can also receive data for which the system may not have been trained and recalibrate itself. System 100 can propose new models or new versions of existing models and redeploy them to accommodate such learning.

The steps described herein are for the purpose of explaining the exemplary embodiments shown, and it is anticipated that ongoing technological developments will change the manner in which certain functions are performed. The examples herein are presented for illustrative purposes and are not intended to limit the invention. In addition, for the convenience of description, the boundaries of functional building blocks in this article are arbitrarily defined. However, other boundaries can be defined as long as the specified functions and their relationships are properly implemented. Based on the teachings contained herein, alternatives (including equivalents to those described herein, Extensions, changes, offsets, etc.) will be obvious to those with ordinary knowledge in the relevant technical field. Such alternatives are within the scope and spirit of the disclosed embodiments.

In summary, the embodiments proposed in this article enable the controller 114 to analyze each input data from a plurality of models/programs based on a specific stochastic and probabilistic model for predictive analysis (model integration), and combine each input data in an aggregator. The outputs of different models are aggregated and, after using portfolio analysis to correlate these outputs with business objectives, an optimal, cost-effective, time-bound and executable schedule is recommended.

The embodiments proposed in this article can also enable the controller to have a conversational nature (intelligent AI/machine learning digital assistant) and provide suggestions. Initially, the controller can evaluate improved efficiency based on the recommendations it provides, and over time can adopt stricter criteria for schedule compliance/variance.

5.0 Program Overview

Figure 4A illustrates an example method 400 of generating a dynamic schedule according to an embodiment of the present invention. Figure 4A can be used as a basis for encoding method 400 into one or more computer programs or other software components for execution by a server computer or as a hosted host.

In step 402, a plurality of input data sets affecting the construction project are received. For example, upon receiving a construction activity schedule request for a construction project, one or more of a plurality of data feeds is received at model integration 112 . The schedule request specifies one or more construction goals for the construction activity. For example, the schedule request could specify "Construct a 50-story building with XYZ design features in Midtown Manhattan" and "on time, on budget, consistent with quality and health considerations" as construction goals (e.g., building type, location, time frame, budget, quality and health considerations). The schedule request specifies a desired design feature identifier (eg, XYZ) that includes architectural information related to the building's appearance. Construction information includes tasks associated with the construction of the building. Monitor 118 may access multiple data feeds and store the plurality of data feeds in the data feed database 132 . Each of the plurality of input data feeds is processed with the construction objectives in mind. Each construction target may be specified or selected from a plurality of construction targets stored in the construction target database 134 .

In step 404, a plurality of input data sets are processed. For example, multiple data feeds are processed by various modules of model integration 112 . A module may include an ensemble of one or more machine learning models. In one embodiment, each individual module processes only the data segments in each data feed that are relevant to the functional goal of each module.

At step 406, a first intermediate data set is generated. For example, each module generates machine learning validated data sets corresponding to each of the plurality of input data sets. In some embodiments, each module of model integration 112 generates a first intermediate data set for each data feed.

At step 408, an optimized schedule is generated for a construction activity using a supervised machine learning model. For example, based on the first intermediate data set generated by each module of the model integration 112, the schedule proposer 226 of the controller 114 may implement a supervised machine learning model that generates an optimized schedule for the construction activities to achieve The construction goal(s) of a construction project. In some embodiments, the analysis aggregator 220 and the correlation analyzer 222 of the controller 114 may first perform analysis and correlation on the first intermediate data set. The data output from the correlation analyzer 222 may be further processed by the combined analytical knowledge and semantic graph model 224 to generate a second intermediate data set including weighted tasks for construction activities. The performance index analysis module 352, the schedule difference module 356, and the contractor execution module 354 jointly dynamically generate an optimized schedule from the weighted tasks. This optimized schedule will achieve construction goals.

At step 410, an optimized schedule is generated in response to the schedule request. The optimized schedule is generated based on weighted tasks. The optimized schedule is also based on performance indicators, task sequencing (e.g. recommendations) are followed and/or schedule variances are generated. The optimized schedule is output (visually and/or audibly) on a client device. Figure 4B illustrates an exemplary schedule.

The schedule can be presented interactively on a graphical user interface, such as in a browser or client application. The example schedule in Figure 4B shows different construction activities such as foundations, walls, roofs, etc. Each construction activity consists of one or more tasks, and each task includes a start date and duration. Construction activities are listed as a series of events. The sample schedule also includes a graphical timeline showing what each activity has to accomplish. Details of each task can be displayed, for example in an overlay, by activating (eg, moving with the mouse, double-clicking, etc.) the task in the timeline. Task details can include information about recommended contractors and vendors.

In one embodiment, feedback about the schedule may be received on monitor 118 to modify one or more subcomponents of controller 114, such as aggregator 220 of controller 114. An updated or further refined schedule can be generated. Controller 114 and monitor 118 implement an integration of closed-loop performance analysis machine learning models.

The advantage of the disclosed technology is that it can adapt to changes and make decisions on the fly to cope with the dynamic nature of a construction project. These techniques may be executed by processing logic, which may include hardware (circuitry, dedicated logic, etc.), software (e.g., running on a general computer system or a dedicated machine), or a combination of both. The processing logic may be included in any node or device (such as a core node, client device, controller, etc.), or in any other computing system or device. One of ordinary skill in the art will appreciate that the disclosed methods can be stored on an article of manufacture, such as a non-transitory computer-readable medium. In one embodiment, the manufactured object may include a computer program that can be accessed from a storage medium or any computer-readable device.

6.0 Hardware Overview

According to an embodiment of the present invention, the technology described herein is implemented by at least one computing device. These techniques may be implemented, in whole or in part, using a combination of at least one server computer and/or other computing device coupled with a network, such as a packet data network. A computing device may be a hardwired device to perform these techniques, or may include a digital electronic device, such as at least one application specific integrated circuit (ASIC) or field programmable gate array (FPGA) that is continuously programmed to Execution of these technologies may include at least one general-purpose hardware processor programmed to execute these technologies according to program instructions in firmware, memory, other storage, or a combination thereof. Such computing devices may also incorporate custom hardwired logic, ASICs or FPGAs and custom programming to implement the techniques described. The computing device may be a server computer, a workstation, a personal computer, a portable computer system, a handheld device, a mobile computing device, a wearable device, a body-mounted or implantable device, a smartphone, a smart appliance, an Internet device, Autonomous or semi-autonomous devices (such as robots or unmanned ground or air vehicles), any other electronic device that incorporates hardwiring and/or program logic to implement the techniques described herein, one or more virtual computers in a data center, or Example, and/or a network of server computers and/or personal computers.

Figure 5 is a block diagram illustrating an example computer system that can be used to implement an embodiment of the present invention. In the example of FIG. 5 , a computer system 500 and instructions are shown schematically (eg, as boxes and circles) that collectively operate to implement the disclosed techniques in hardware, software, or a combination of hardware and software. The level of detail presented is equivalent to that which one of ordinary skill in the art to which this invention belongs would normally convey relevant aspects of computer architecture and computer system implementations.

Computer system 500 includes an input/output (I/O) subsystem 502, which may include buses and/or other communication mechanisms for communicating information and/or instructions between the components of computer system 500 through electronic signal paths. I/O subsystem 502 may include an I/O controller, a memory controller and At least one input/output port. Electronic signal paths are represented schematically in the diagrams, for example as lines, one-way arrows or two-way arrows.

At least one hardware processor 504 is coupled to the input/output subsystem 502 to process information and instructions. Hardware processor 504 may include, for example, a general purpose microprocessor or microcontroller and/or a special purpose microprocessor, such as an embedded system, a graphics processing unit (GPU), a digital signal processor, or an ARM processor. Processor 504 may include an integrated arithmetic logic unit (ALU) or may be coupled to a separate arithmetic logic unit.

Computer system 500 includes one or more units of memory 506 (eg, main memory) coupled to input/output subsystem 502 for electronically digitally storing data and instructions to be executed by processor 504. Memory 506 may include volatile memory, such as various forms of random access memory (RAM) or other dynamic storage devices. Memory 506 may also be used to store temporary variables or other intermediate information while processor 504 executes instructions. When these instructions are stored in a non-transitory computer-readable medium and are accessible to the processor 504, they can transform the computer system 500 into a special-purpose machine that is customized to execute the instructions specified in the instructions. operation.

Computer system 500 further includes non-volatile memory, such as read-only memory (ROM) 508 or other static storage devices coupled to input/output subsystem 502 , for storing information and instructions provided to processor 504 . Read-only memory 508 may include various forms of programmable read-only memory (PROM), such as erasable programmable read-only memory (EPROM) or electronically erasable programmable read-only memory (EEPROM). . Persistent storage 510 may include various forms of non-volatile RAM (NVRAM), such as flash memory (FLASH), or solid-state storage devices, magnetic or optical disks, such as CD-ROM or DVD-ROM, and may be used with Input/output subsystem 502 is coupled to store information and instructions. Storage 510 is an example of a non-transitory computer-readable medium that can be used to store instructions and data that, when executed by processor 504, can cause computer-implemented methods to be performed to implement the techniques described herein.

Instructions in memory 506, read-only memory 508, or storage 510 may include one or more sets of instructions organized into modules, methods, objects, functions, routines, or call routines. These instructions may be organized into one or more computer programs, operating system services or applications (including mobile applications). These instructions may include: operating system and/or system software; one or more libraries that support multimedia, programming or other functions; data protocol instructions or stacks that execute TCP/IP, HTTP or other communication protocols; file format processing Commands that parse or render files encoded in HTML, XML, JPEG, MPEG, or PNG; user interface commands that render or interpret graphical user interfaces (GUIs), command line interfaces, or text user interfaces; applications Software such as office suites, Internet access applications, design and manufacturing applications, graphics applications, audio applications, software engineering applications, educational applications, games, or various other applications. These commands can execute web server, web application server, or web client functions. These instructions may or may not use Structured Query Language (SQL), object stores, graph databases, flat file systems, or other data storage formats, and are organized into presentation layers, applications layer and data storage layer, such as a relational database system.

Computer system 500 may be coupled to at least one output device 512 through input/output subsystem 502 . In one embodiment, the output device 512 is a digital computer monitor. Examples of displays that may be used in various embodiments include touch screen displays, light emitting diode (LED) displays, liquid crystal displays (LCD), or electronic paper displays. Computer system 500 may include other types of output devices 512 as alternatives to or in addition to display devices. Examples of other output devices 512 include printers, receipt printers, plotters, Projector, sound card or display card, speaker, buzzer or piezoelectric device or other sound device, lamp or LED or LCD indicator, tactile device, actuator or servo device.

At least one input device 514 is coupled to the input/output subsystem 502 to transmit signals, data, command selections, or gestures to the processor 504 . Examples of input devices 514 include touch screens, microphones, still image and video digital cameras, alphanumeric and other keys, keypads, keyboards, graphics tablets, image scanners, joysticks, clocks, switches, buttons, dials, slideshows Chips and/or various types of sensors, such as force sensors, motion sensors, thermal sensors, accelerometers, gyroscopes and inertial measurement unit sensors and/or various types of transceivers, such as Wireless transceivers (such as cellular or Wi-Fi, RF or infrared transceivers) and Global Positioning System (GPS) transceivers.

Another type of input device is a control device 516, which can perform cursor control or other automatic control functions, such as navigating through a graphical interface on a display screen, providing alternative or additional input functions. The control device 516 may be a touchpad, a mouse, a trackball, or a cursor direction key, and is used to transmit direction information and command selections to the processor 504 and control cursor movement on the display. The control device may have degrees of freedom in at least two axes, namely a first axis (eg x) and a second axis (eg y), so that the device can indicate a position on a plane. Another type of input device is a wired, wireless, or optical control device such as a joystick, wand, console, steering wheel, pedals, shift mechanism, or other type of control device. Input device 514 may include a combination of multiple different input devices, such as a camera and a depth sensor.

In another embodiment, computer system 500 may include an Internet of Things (IoT) device in which one or more of output device 512, input device 514, and control device 516 are omitted. Alternatively, in such embodiments, the input device 514 may include one or more cameras, motion detectors, thermometers, microphones, seismometers, other sensors or detectors, measurement devices, or encoders, and the output device 512 may include special Purpose displays, such as single-line LED or LCD displays, one or more pilot lights, display panels, gauges, valves, solenoids, actuators or servos.

When the computer system 500 is a mobile computing device, the input device 514 may include a Global Positioning System (GPS) receiver coupled to a GPS module capable of triangulating positioning with a plurality of GPS satellites for the computer. System 500 determines and generates geolocation or location data, such as latitude and longitude values. Output devices 512 may include hardware, software, firmware, and interfaces for generating location reporting packets, notifications, pulse or heartbeat signals, or other repetitive transmission of data that specifically identifies the location of computer system 500, either alone or with other applications. The program-specific data is combined and sent to the host 524 or server 530.

Computer system 500 may implement the techniques described herein using custom hard-wired logic, at least one ASIC or FPGA, firmware and/or program instructions or logic that are loaded and used in or combined with the computer system. When executed, it causes the computer system or programming to operate the computer system as a special purpose machine. According to one embodiment, the techniques described herein are performed by computer system 500 in response to processor 504 executing at least one sequence of at least one instruction contained in main memory 506 . The instructions may be read into main memory 506 from another storage medium (eg, storage 510). Execution of sequences of instructions contained in main memory 506 causes processor 504 to perform the process steps described herein. In alternative embodiments, hardwired circuitry may be used in place of or in combination with software instructions.

The term "storage media" as used herein refers to any non-transitory medium that stores data and/or instructions that cause a machine to operate in a specific manner. Such storage media may include non-volatile media and/or volatile media. Non-volatile media include, for example, optical disks or magnetic disks, such as storage 510 . Volatile media includes dynamic memory, such as memory 506. Common forms of storage media include, for example, hard drives, solid state drives, flash drives, magnetic data storage media, any optical or physical data storage media, memory chips, or the like.

Storage media is different from, but can be used with, transmission media. Transmission media participates in data transfer between storage media. For example, transmission media include coaxial cables, copper wires, and fiber optics, including the wires that make up the busbars of the input/output subsystem 502 . Transmission media can also take the form of sound or light waves, such as those produced during radio and infrared data communications.

The process of transmitting at least one sequence of at least one instruction to processor 504 for execution may include different forms of media. For example, the instructions may initially be loaded on a remote computer's disk or solid state drive. The remote computer can load the instructions into its dynamic memory and transmit the instructions through a communications link (such as fiber optic, coaxial cable, or telephone line using a modem). The modem or router local to the computer system 500 can receive the data on the communication link and convert the data into a format that can be read by the computer system 500 . For example, a receiver (such as a radio frequency antenna or infrared detector) can receive the data carried in the wireless or optical signal, and appropriate circuitry can provide the data to the input/output subsystem 502, such as by placing the data on a bus. . The input/output subsystem 502 transfers the data to the memory 506, and the processor 504 retrieves the instructions from the memory 506 and executes them. Instructions received by memory 506 may be stored on storage 510 before or after execution by processor 504, as appropriate.

The computer system 500 also includes a communication interface 518 coupled to a bus on the input/output subsystem 502 . Communication interface 518 provides two-way data communication with network links 520 that are directly or indirectly connected to at least one communication network, such as network 522 or a public or private cloud on the Internet. For example, the communication interface 518 can be an Ethernet interface, an Integrated Services Digital Network (ISDN) card, a cable modem, a satellite modem, or provide a data communication connection to a corresponding communication line type (such as an Ethernet cable, Any type of metal cable, fiber optic line, or telephone line) modem. Network 522 may broadly represent a local area network (LAN), a wide area network (WAN), a campus network, the Internet, or a combination of any of the foregoing. Communication interface 518 may include a local network card to provide data communication connection to a phase A local area network that contains either a cellular radiotelephone interface that is wired to send or receive cellular data in accordance with the cellular radiotelephone wireless network standard, or a satellite radio interface that is wired to be connected in accordance with Satellite wireless network standard for sending or receiving digital data. In any such implementation, communication interface 518 sends and receives electrical, electromagnetic, or optical signals through signal paths that carry digital data streams representing various types of information.

Network link 520 typically provides electrical, electromagnetic, or optical data communications to other data devices directly or through at least one network, using technologies such as satellite, cellular, Wi-Fi, or Bluetooth technology. For example, network link 520 may provide a connection to host 524 through network 522.

Additionally, network link 520 may provide connections to other computing devices through network 522 provided through Internet devices and/or computers operated by Internet service providers (ISPs) 526 . ISP 526 provides data communication services through the global packet data communication network (represented as the Internet 528). Server 530 may be coupled to Internet 528. Server 530 broadly represents any computer, data center, virtual machine or virtual computing instance (with or without a hypervisor), or computer running a containerized programming system, such as DOCKER or KUBERNETES. Server 530 may represent an electronic digital service that is implemented using more than one computer or instance and is implemented by transmitting web service requests, Uniform Resource Locator (URL) strings with parameters in HTTP payloads, API calls, Application service calls or other service calls to access and use. Computer system 500 and servers 530 may form components of a distributed computing system that includes other computers, processing clusters, server farms, or other computer organizations that cooperate to perform tasks or execute applications or services. Server 530 may include one or more sets of instructions organized into modules, methods, objects, functions, routines, or call procedures. These instructions may be organized into one or more computer programs, operating system services or applications (including mobile applications). These instructions may include: operating system and/or system software; support for multimedia, programming or other functions one or more libraries; data protocol commands or stacks that execute TCP/IP, HTTP, or other protocols; file format processing commands to parse or render files encoded in HTML, XML, JPEG, MPEG, or PNG; use User interface commands used to present or interpret a graphical user interface (GUI), command line interface, or text user interface; application software, such as office suites, Internet access applications, design and manufacturing applications, graphics Applications, audio applications, software engineering applications, educational applications, games, or various other applications. Server 530 may include a web application server that uses (or does not use) Structured Query Language (SQL), object store, graph database, flat file system, or other data storage The form carries the presentation layer, application layer and data storage layer, such as a relational database system.

Computer system 500 can send messages and receive data and instructions, including program code, through the network, network link 520, and communication interface 518. In an example using the Internet, server 530 may transmit the requested application code through Internet 528, ISP 526, local area network 522, and communication interface 518. The received code may be executed by processor 504 upon receipt and/or stored in memory 510 or other non-volatile storage device for subsequent execution.

The execution of instructions described in this section may execute a process in the form of an executing instance of a computer program consisting of the program code and its current activities. Depending on the operating system (OS), a process can consist of multiple threads that execute instructions simultaneously. In this case, a computer program is a passive collection of instructions, and a process may be the actual execution of those instructions. Several processes may be associated with the same program; for example, opening several instances of the same program often means that more than one process is being executed. Multitasking may be performed to allow multiple processes to share processor 504. Although each processor 504 or processor core performs only one task at a time, the computer system 500 can be programmed to multi-task, allowing each processor to switch between executing tasks without having to wait for each task. over become. In one embodiment, switching may be performed when a task performs an input/output operation, when the task indicates that it can be switched, or when a hardware interrupt occurs. Can implement time-sharing to quickly respond to interactive user applications, and can provide the appearance of multiple processes executing simultaneously by quickly executing context exchanges. In one embodiment, for security and reliability considerations, the operating system can prevent direct communication between independent processes and provide strictly mediated and controlled inter-process communication functions.

7.0 Software Overview

Figure 6 is a block diagram of a basic software system 600 that can be used to control the operation of the computing device 500. The software system 600 and its components, including the connection methods, relationships and functions thereof, are only examples and are not intended to limit the execution of the exemplary embodiments. Other software systems suitable for implementing the exemplary embodiments may have different components, including components with different connections, relationships, and functions.

The software system 600 is used to guide the operation of the computing device 500 . Software system 600 may be stored on system memory (RAM) 506 and fixed storage (such as hard disk or flash memory) 510 and includes a kernel or operating system (OS) 610.

The operating system 610 manages low-level aspects of computer operations, including managing program execution, memory allocation, file input and output (I/O), and device input and output. One or more applications, represented by 602A, 602B, 602C...602N (collectively, applications 602), may be "loaded" (e.g., transferred from fixed storage 510 to memory 506) for use by software system 600 implement. Applications or other software used on software system 600 may also be stored as a downloadable set of computer-executable instructions, such as from a network location (such as a network server, application store or other online service) and installation.

The software system 600 includes a graphical user interface (GUI) 615 for receiving user instructions and data in a graphical (eg, "click" or "touch gesture") manner. It can also be reversed by the software system 600 performs actions on the aforementioned input according to instructions from the operating system 610 and/or the application 602. The user interface 615 is also used to display operation results from the operating system 610 and the application 602. The user can provide additional input or terminate the session (eg, log out) on the system or application.

Operating system 610 may execute directly on bare hardware 620 (eg, processor 504) of device 500. In addition, a hypervisor or virtual machine monitor (VMM) 630 can also be installed between the bare hardware 620 and the operating system 610. In this configuration, virtual machine monitor 630 may serve as a software "buffer" or virtualization layer between operating system 610 and bare hardware 620 of device 500 .

Virtual machine monitor 630 instantiates and runs one or more virtual machine instances ("guests"). Each guest machine includes a "guest" operating system (eg, operating system 610), and one or more applications (eg, application 602) that are designed to execute on the guest operating system. The virtual machine monitor 630 uses a virtual operating platform to display the guest operating system and manage the execution of the guest operating system.

In some examples, virtual machine monitor 630 may allow the guest operating system to run as if running directly on bare hardware 620 of device 500 . In such instances, the same version of the guest operating system configured to execute directly on bare hardware 620 may also execute on virtual machine monitor 630 without modification or reconfiguration. In other words, in some instances, virtual machine monitor 630 may provide full hardware and CPU virtualization technology for the guest operating system.

In other examples, the guest operating system may be specifically designed or configured to execute efficiently on virtual machine monitor 630. In such instances, the guest operating system "knows" that it is executing on a virtual machine monitor. In other words, in some instances, virtual machine monitor 630 may provide para-virtualization technology for the guest operating system.

The above description of basic computer hardware and software is presented for the purpose of illustrating the basic computer components that may be used to implement the exemplary embodiments. However, the exemplary embodiments are not necessarily limited to any particular computing environment or computing device configuration. On the contrary, after reading the above disclosure, a person with ordinary skill in the art to which the present invention belongs should understand that any type of system architecture or processing environment can be used to implement the features and functions described in the exemplary embodiments as long as it can support them. Examples.

8.0 Other aspects of the invention

Although some of the figures mentioned in the above description include flowcharts with steps shown in sequence, the steps can be performed in any order and are not limited to the order shown in the flowcharts. In addition, some steps may be optionally performed, may be performed multiple times, and/or may be performed by different components. All steps, operations, and functions in the flowchart shown are intended to represent operations performed in each embodiment using programs in a dedicated computer or a general computer. In other words, each flowchart in the disclosure, combined with the associated text, serves as a guide, plan, or specification of all or part of an algorithm for programming a computer to perform the described function. As is generally known, the technical field to which the present invention belongs requires a high technical level to be implemented. Therefore, the information conveyed in the flowcharts and related text shown is based on the exchange of programs, algorithms and implementation methods by persons skilled in the art. Presented with the fullness and level of detail that can be expected.

In the foregoing description, exemplary embodiments of the invention have been described with reference to numerous specific details. However, these details may vary from implementation to implementation depending on the requirements of the particular implementation. Therefore, the exemplary embodiments of the present invention are to be regarded in an illustrative rather than a restrictive sense.

100: Networked computer systems/systems

102:Internet

104:Client computer

106:Server computer

112:Model integration

114:Controller

116: Notifier

118:Monitor

120:Modifier

130:Data repository

132: Data feed database

134: Construction target database

136:Model configuration database

138:Training data database

140: Suggested database

Claims (20)

A computer-implemented method includes: receiving a schedule request for a construction activity of a construction project from a user device through a server device, the schedule request specifying at least one construction goal; accessing a plurality of data through the server device input data sets, wherein the plurality of input data sets include data sets from a data feed database that includes a plurality of data feeds from various sources, and wherein each of the data feeds includes a data segment , the data segments include metadata indicating a data type, and wherein the plurality of input data sets includes at least one camera feed, wherein the at least one camera feed includes tags identifying all segments of the at least one camera feed, or metadata; use new training data to train a plurality of neural network machine learning model ensembles, the new training data includes data related to the construction project; transfer the trained plurality of neural network machines through the server device An ensemble of learning models is applied to each of the plurality of input data sets to generate a plurality of intermediate data sets that each specify factors affecting the construction project, wherein the trained plurality of neural networks Each trained neural network machine learning model ensemble in the neural network machine learning model ensemble generates an intermediate data set based on the functional goals of the respective neural network machine learning model ensemble, the functional goals derived from the functional goals of the respective neural network machine learning model ensemble. A specific segment among all segments associated with the functional goal of the network machine learning model integration; through the server device, based on the plurality of intermediate data sets, at least one knowledge and semantic graph model is used to determine the task data, one of which analyzes The aggregator correlates various outputs of each neural network machine learning model ensemble using a correlation analyzer to generate correlation data that is used as input to the at least one knowledge and semantic graph model; Generate a schedule of the construction activity through the server device and from the task data, the schedule will achieve the at least one construction goal and comply with a tolerance associated with a baseline schedule; and through the server device based on The schedule is informed from the at least one construction target identified in the schedule request. For example, the computer implementation method of claim 1 further includes: applying a stochastic model to the plurality of intermediate data sets to determine relevant data of the factors affecting the construction project; and applying the at least one knowledge and semantic graph model The associated data and tenant data are used to generate the task data. For example, the computer implementation method of claim 2 further includes: receiving feedback data on issues related to the construction project; using the feedback data to modify the random model; and updating the schedule of the construction project. As claimed in claim 3, the computer implementation method, wherein the receiving feedback data includes: receiving a real-time camera feed from a construction site; determining efficiency problem data related to the construction project from the real-time camera feed; and obtaining efficiency problem data from the real-time camera feed. Generate this feedback information. The computer-implemented method of claim 3, wherein the feedback data is received as user input via a graphical user interface. For example, the computer implementation method of claim 1, wherein generating the schedule includes: determining the difference data between actual construction and expected construction based on the schedule; and using the difference data to improve the schedule. The computer implementation method of claim 1, wherein the schedule is represented by at least one of a graphical user interface, voice and message. The computer-implemented method of claim 1, wherein the ensemble of trained neural network machine learning models includes at least one of the following: a climate analysis module configured to derive from at least one input of the plurality of input data sets A data set analyzes weather patterns; a structural analysis module configured to analyze quality and design from at least one input data set of the plurality of input data sets; a quality analysis module configured to analyze quality and design from the plurality of input data sets A centralized analysis of progress from at least one input data set of the plurality of input data sets; an inventory analysis module configured to analyze inventory from at least one input data set of the plurality of input data sets; a method group configured to, based on the requirements, from the At least one input data set of a plurality of input data sets analyzes completion on time; a global event impact module configured to analyze global events from at least one input data set of the plurality of input data sets; a supply chain analysis module , which is configured to analyze supply and delivery from at least one input data set of the plurality of input data sets; an equipment health and Internet of Things (IoT) indicator analysis module configured to analyze supply and delivery from the plurality of input data sets. at least one input data set analyzing equipment demand and purchasing factors; and a labor efficiency module configured to analyze labor availability and skills from at least one input data set of the plurality of input data sets. A non-transitory computer-readable storage medium storing one or more programmed instructions that, when executed by one or more computing devices, result in: one received from a user device through a server device A schedule request for a construction activity of a construction project, the schedule request specifying at least one construction goal; accessing a plurality of input data sets through the server device, wherein the plurality of input data sets include data from a data feed database a data set, the data feed database includes a plurality of data feeds from various sources, and wherein each of said data feeds includes a data section, said data section includes metadata indicating a data type, and wherein said plurality of an input data set including at least one camera feed, wherein the at least one camera feed includes tags or metadata identifying all segments of the at least one camera feed; training a plurality of neural network machine learning model ensembles using the new training data, The new training data includes data related to the construction project; a plurality of trained machine learning models are integrated and applied to each of the plurality of input data sets through the server device to generate a plurality of intermediate data sets , each of the plurality of intermediate data sets indicates factors affecting the construction project, wherein each of the trained neural network machine learning model ensembles is based on its respective neural network machine learning model ensemble. A network machine learning model integrates functional targets to generate an intermediate data set from a specific segment among all segments associated with the functional targets integrated by the respective neural network machine learning models; through the The server device uses at least one knowledge and semantic graph model to determine the task data based on the plurality of intermediate data sets, wherein an analysis aggregator uses a correlation analyzer to correlate the various outputs of each neural network machine learning model integration , to generate associated data that is used as input to the at least one knowledge and semantic graph model; Generate a schedule of the construction activity through the server device and from the task data, the schedule will achieve the at least one construction goal and comply with a tolerance associated with a baseline schedule; and through the server device based on The schedule is informed from the at least one construction target identified in the schedule request. The non-transitory computer-readable storage medium of claim 9, wherein one or more programmed instructions are stored therein, which, when executed by one or more computing devices, further results in: applying a stochastic model to the A plurality of intermediate data sets are used to determine related data of the factors affecting the construction project; and the at least one knowledge and semantic graph model is applied to the related data and tenant data to generate the task data. If the non-transitory computer-readable storage medium of claim 10 stores one or more programmed instructions, when the instructions are executed by one or more computing devices, it further results in: receiving questions related to the construction project feedback data; use the feedback data to modify the random model; and update the schedule of the construction project. The non-transitory computer-readable storage medium of claim 11, wherein the receiving feedback data includes: receiving a real-time camera feed from a construction site; determining efficiency issue data related to the construction project from the real-time camera feed; and The feedback data is generated from the efficiency problem data. The non-transitory computer-readable storage medium of claim 11, wherein the feedback data is received as user input via a graphical user interface. For example, the non-transitory computer-readable storage medium of claim 9, wherein generating the schedule includes: determining the difference data between actual construction and expected construction based on the schedule; and using the difference data to improve the schedule. The non-transitory computer-readable storage medium of claim 9, wherein the schedule is represented by at least one of a graphical user interface, voice and message. A computing system, which includes: one or more computer systems, which includes one or more hardware processors and storage media; instructions stored in the storage medium, when the computing system executes the instructions, will cause the Computing system execution: receiving a schedule request for a construction activity for a construction project from a user device through the server device, the schedule request specifying at least one construction goal; accessing a plurality of input data sets through the server device , wherein the plurality of input data sets includes data sets from a data feed database that includes a plurality of data feeds from various sources, and wherein each of the data feeds includes a data segment, the data The segments include metadata indicating a data type, and wherein the plurality of input data sets includes at least one camera feed, wherein the at least one camera feed includes tags or metadata identifying all segments of the at least one camera feed; using Use new training data to train a plurality of neural network machine learning model integrations. The new training data includes data related to the construction project; integrate and apply the trained plurality of neural network machine learning models through the server device. In each of the plurality of input data sets, a plurality of intermediate data sets are generated, each of the plurality of intermediate data sets specifies factors affecting the construction project, wherein the trained plurality of gods Each trained neural network machine learning model ensemble in the network machine learning model ensemble generates an intermediate data set based on the functional goals of the respective neural network machine learning model ensemble, the functional goals coming from the respective neural network machine learning model ensemble. The neural network machine learning model integrates a specific segment among all segments associated with the functional goal; through the server device, based on the plurality of intermediate data sets, at least one knowledge and semantic graph model is used to determine the task data, wherein An analysis aggregator correlates the various outputs of each neural network machine learning model ensemble using a correlation analyzer to generate correlation data that is used as input to the at least one knowledge and semantic graph model; by The server device generates a schedule of the construction activity from the task data, the schedule will achieve the at least one construction goal and comply with a tolerance associated with a baseline schedule; and through the server device, based on the The schedule is informed by the at least one construction target identified in the schedule request. Such as the computing system of claim 16, wherein the instructions, when executed by the computing system, will cause the computing system to further perform: apply a random model to the plurality of intermediate data sets to determine the factors affecting the construction project. associated data of factors; and applying the at least one knowledge and semantic graph model to the associated data and tenant data to generate the task data. For example, the computing system of claim 17, wherein the instructions, when executed by the computing system, will cause the computing system to further perform: receive feedback data on issues related to the construction project; use the feedback data to modify the stochastic model; and Update the schedule for this construction project. The computing system of claim 18, wherein receiving the feedback data includes: receiving a real-time camera feed from a construction site; determining efficiency issue data related to the construction project from the real-time camera feed; and from the efficiency issue data. The feedback data is generated from the data. The computing system of claim 16, wherein generating the schedule includes: determining difference data between actual construction and expected construction based on the schedule; and using the difference data to improve the schedule.