Building equipment intelligent scheduling method based on AI algorithm

Abstract

This invention provides an AI-based intelligent scheduling method for construction equipment, belonging to the field of construction scheduling technology. This AI-based intelligent scheduling method constructs a three-layer intelligent scheduling framework: macro, meso, and micro. The macro layer uses an improved time series prediction algorithm to deeply integrate the construction cycle and external events, achieving accurate and forward-looking resource planning. The meso layer employs a combined auction algorithm to quantify equipment synergy, optimize task allocation, and significantly improve the efficiency of multi-equipment joint operations. The micro layer utilizes a distributed negotiation mechanism to simulate swarm behavior, endowing the system with strong self-organization and dynamic adaptability. Supplemented by a resilient scheduling algorithm, it continuously monitors system health and automatically recovers. Through dynamic weighted multi-layer decision-making, it ensures the system maintains stable and efficient operation under various disturbances.

Description

Technical Field

[0001] This invention relates to the field of building scheduling technology, specifically to an intelligent scheduling method for building equipment based on AI algorithms. Background Technology

[0002] Construction equipment refers to the general term for all kinds of machinery, tools, and facilities used in the construction process, including tower cranes, excavators, bulldozers, concrete mixers, construction elevators, cranes, road rollers, etc. They provide necessary power support and technical guarantees for construction and are an indispensable and important part of modern construction. Construction equipment scheduling refers to the rational allocation, coordination, and management of the usage time, location, and tasks of various equipment based on the construction plan, project progress, and site requirements to ensure the efficient utilization of equipment resources. The main reasons for construction equipment scheduling include: improving construction efficiency by avoiding equipment idleness or conflicts through scientific scheduling, thus shortening the construction period; reducing project costs by minimizing equipment rental or purchase expenses and avoiding resource waste due to improper equipment scheduling; ensuring construction safety by rationally arranging equipment operating time and work areas to reduce the risk of accidents caused by equipment overlap or overuse; optimizing resource allocation by dynamically adjusting equipment usage according to the needs of different project stages to ensure equipment support for key processes; and improving project quality by ensuring the continuity and stability of construction processes through appropriate equipment scheduling, thereby avoiding quality problems caused by insufficient or improper equipment.

[0003] Current technologies for scheduling construction equipment largely rely on centralized optimization models, treating scheduling as a static mathematical programming problem and attempting to calculate the globally optimal solution through a central processing unit. This architecture proves rigid and fragile when faced with the inherent high dynamism and uncertainty of construction sites; any unplanned disturbance can cause the entire solution to fail, requiring time-consuming and laborious recalculation. Existing algorithms often fall into the trap of piling up technologies, blindly introducing complex deep learning models while ignoring the high cost of data annotation and the scarcity of effective samples in the construction field. This results in poor generalization ability of the models in practical applications, rendering them merely laboratory results. Most scheduling systems treat equipment as isolated resources, optimizing only their individual travel and load, completely ignoring the objectively existing collaborative relationships and clustering effects between equipment, such as the coordinated operation of cranes and concrete pump trucks. This fragmented optimization often leads to local optima while resulting in low global efficiency. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an intelligent scheduling method for building equipment based on AI algorithms. This method solves the problems of existing building equipment scheduling methods that rely too heavily on centralized static optimization models, which cannot effectively cope with dynamic disturbances at the construction site. Furthermore, it addresses the issues of poor generalization ability and low global efficiency caused by neglecting data scarcity and equipment coordination relationships.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent scheduling method for building equipment based on AI algorithms, comprising the following steps:

[0006] S1. Construct a three-tiered progressive decision-making framework for building equipment scheduling, including macro, meso, and micro levels;

[0007] S2. At the macro level, based on historical and real-time data, an improved time series forecasting algorithm is used to predict trends and make strategic plans for equipment demand at construction sites;

[0008] S3. At the meso-level, an improved combinatorial auction algorithm is used to optimize the matching of devices and tasks, while also considering the synergistic effect between devices;

[0009] S4. At the micro level, a distributed negotiation algorithm is used to enable devices to make autonomous decisions and adjust in real time;

[0010] S5. Monitor the operating status of the scheduling system through a resilient scheduling algorithm and quickly restore system performance when disturbances occur;

[0011] S6. Integrate the decision results from each layer and output the final equipment scheduling plan.

[0012] Preferably, the improved time series prediction algorithm described in step S2 is: the algorithm is named the Trend-Seasonal Prediction Adaptive Algorithm; this algorithm adds a feature recognition module specific to building scenes based on the traditional ARIMA model; the prediction model is shown below: .

[0013] Preferably, the improved combinatorial auction algorithm described in step S3 is specifically: the algorithm is named "Combined Auction Allocation Algorithm Considering Synergy Effect"; [equipment / device] For task sets The formula for calculating the bid value is:

[0014] .

[0015] Preferably, the improved combinatorial auction algorithm employs a greedy optimization strategy for task allocation: let the task set be... The equipment set is Initialization allocation result Unassigned task set Repeat the following steps until... Or unable to continue allocation:

[0016] a) For each unassigned task subset Calculate the bid value of all devices or groups of devices for this subset of the task;

[0017] b) Select the allocation combination with the highest cost performance. The formula for calculating cost-effectiveness is as follows: ;

[0018] c) Assignment combination Add to allocation results ;

[0019] d) From Remove the assigned task subset Remove from available devices Output the final allocation result. .

[0020] Preferably, the distributed negotiation algorithm mentioned in step S4 is specifically: the algorithm name is a distributed negotiation algorithm based on bee behavior; the device is defined. For the task The pheromone intensity is The pheromone update rule is as follows: .

[0021] Preferably, the distributed negotiation algorithm includes the following negotiation phases:

[0022] Reconnaissance Phase: Randomly select a proportion from all idle devices. The equipment is used as reconnaissance equipment, and these devices actively explore the surrounding available missions;

[0023] Evaluation Phase: The reconnaissance equipment assesses the fitness of the detected missions. If the fitness is higher than a threshold... Then they return to the base to recruit;

[0024] Recruitment Phase: High-fitness reconnaissance equipment recruits other equipment through a pheromone mechanism, with a recruitment quantity... It is determined by the following formula: ;

[0025] Execution phase: The recruited devices go to perform the mission and update the pheromones on the path after the mission is completed.

[0026] Preferably, the resilient scheduling algorithm described in step S5 is specifically: the algorithm is named a resilient scheduling algorithm based on ant colony behavior; and the resilience index of the scheduling system is defined. for: .

[0027] Preferably, the resilience recovery mechanism includes the following steps:

[0028] Identify disturbance types: equipment failure, task change, environmental change, or resource shortage;

[0029] Select a recovery strategy based on the type of disturbance:

[0030] For equipment failures, an equipment replacement strategy should be adopted;

[0031] For task changes, a task redistribution strategy is adopted, using an improved Hungarian algorithm for fast redistribution;

[0032] In response to environmental changes, a planned adjustment strategy is adopted to recalculate equipment operating times and routes;

[0033] After the recovery strategy is executed, the system resilience metrics are recalculated until... .

[0034] Preferably, the decision fusion in step S6 adopts a weighted fusion method:

[0035] Output strategic planning solutions at the macro level The weight is Mid-level output task allocation scheme The weight is Real-time adjustment scheme for microscopic layer output The weight is Final scheduling scheme for: During the scheduling process, In the later stages of dispatching or in emergency situations, .

[0036] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the AI-based intelligent scheduling method for building equipment.

[0037] This invention provides an intelligent scheduling method for building equipment based on AI algorithms. It has the following beneficial effects:

[0038] This invention provides an intelligent scheduling method for building equipment based on AI algorithms. This technology constructs a hierarchical and distributed intelligent scheduling ecosystem. This technology designs a three-layer progressive decision-making framework of macro, meso, and micro levels. Each layer adopts a targeted and improved lightweight algorithm to decompose complex global problems into manageable local problems. While ensuring the quality of decision-making, it significantly reduces the real-time computing burden. Through an improved combinatorial auction algorithm, the synergistic effect between equipment is quantified in task allocation, making the scheduling results more in line with the physical laws of team collaboration in actual operations, thereby improving the hidden efficiency improvement potential.

[0039] This invention provides an intelligent scheduling method for building equipment based on AI algorithms. Through a distributed negotiation algorithm, this invention enables individual equipment to make autonomous decisions and negotiate with its neighbors based on simple rules, allowing the overall system intelligence to move beyond local interactions among the equipment group. This decentralized design improves the system's response speed to sudden disturbances. The scheduling algorithm, like an immune system, can continuously monitor the system's state. Once performance degradation is detected, it can automatically trigger multiple preset recovery strategies to ensure that the scheduling scheme can smoothly adapt to rather than collapse under various internal and external changes. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the overall process of the method of the present invention;

[0041] Figure 2 This is a schematic diagram of the three-layer decision-making framework and data interaction process of the present invention;

[0042] Figure 3 This is a schematic diagram of the system deployment and hardware architecture of the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] like Figure 1-3 As shown, this embodiment of the invention provides an intelligent scheduling method for building equipment based on AI algorithms, including the following steps:

[0045] S1. Construct a three-tiered progressive decision-making framework for building equipment scheduling, including macro, meso, and micro levels;

[0046] S2. At the macro level, based on historical and real-time data, an improved time series forecasting algorithm is used to predict trends and make strategic plans for equipment demand at construction sites;

[0047] S3. At the meso-level, an improved combinatorial auction algorithm is used to optimize the matching of devices and tasks, while also considering the synergistic effect between devices;

[0048] S4. At the micro level, a distributed negotiation algorithm is used to enable devices to make autonomous decisions and adjust in real time;

[0049] S5. Monitor the operating status of the scheduling system through a resilient scheduling algorithm and quickly restore system performance when disturbances occur;

[0050] S6. Integrate the decision results from each layer and output the final equipment scheduling plan.

[0051] The improved time series prediction algorithm in step S2 is as follows:

[0052] The algorithm is called Trend-Seasonal Forecast Adaptive Algorithm (TSP-AD).

[0053] This algorithm adds a feature recognition module specific to architectural scenes to the traditional ARIMA model;

[0054] The prediction model is shown below:

[0055]

[0056] in: For time Forecasted equipment demand; The trend component is calculated using the following formula:

[0057]

[0058] in This is the actual value from the previous period. This is the trend value from the previous period. α is the smoothing coefficient (0 < α < 1). This represents the project schedule change rate. This is the schedule impact factor; The seasonal component is calculated using the following formula:

[0059]

[0060] in The length of the seasonal cycle (7 days or 30 days in architectural scenarios). γ is the seasonal smoothing coefficient (0 < γ < 1); The formula for calculating the impact component of a causal event is as follows:

[0061]

[0062] in For the first An exogenous event in time The intensity of the impact, For the corresponding weights, The number of exogenous events to be considered; This is the random error term.

[0063] The improved combinatorial auction algorithm in step S3 is as follows:

[0064] The algorithm is named Combinatorial Auction Allocation Algorithm Considering Synergy (CAA-IS).

[0065] equipment For task sets The formula for calculating the bid value is:

[0066]

[0067] in: For equipment Performing tasks alone The fundamental value:

[0068]

[0069] in: For equipment Current location to task Distance between locations; For equipment Execute the task Estimated energy consumption; For equipment Maintenance status indicators; The degree of matching between equipment operator skills and task requirements; These are the weighting coefficients, and ; The normalization function maps each index to the [0,1] interval; For equipment group Synergistic effect function:

[0070]

[0071] in: This refers to the number of devices in the equipment group. For equipment With equipment The historical collaborative efficiency coefficient; For task set The intensity of demand for equipment collaboration; This represents the synergistic effect weighting coefficient.

[0072] The improved combinatorial auction algorithm employs a greedy optimization strategy for task allocation:

[0073] Let the task set be The equipment set is ;

[0074] Initialization allocation results Unassigned task set ;

[0075] Repeat the following steps until Or unable to continue allocation:

[0076] a) For each unassigned task subset Calculate the bid value of all devices or groups of devices for this subset of the task;

[0077] b) Select the allocation combination with the highest cost performance. The formula for calculating cost-effectiveness is as follows:

[0078]

[0079] in: For equipment The cost per unit time used; For task subset The average duration of the task in the middle; This is the project duration impact factor;

[0080] c) Assignment combination Add to allocation results ;

[0081] d) From Remove the assigned task subset Remove from available devices Output the final allocation result. .

[0082] The distributed negotiation algorithm in step S4 is as follows:

[0083] The algorithm is named the Distributed Negotiation Algorithm Based on Bee Behavior (DN-BA).

[0084] Define device For the task The pheromone intensity is ;

[0085] The pheromone update rules are as follows:

[0086]

[0087] in: ρ is the pheromone evaporation coefficient, 0 < ρ < 1; The pheromone increment is calculated as follows:

[0088]

[0089] in: The pheromone constant; For equipment Execute the task The fitness is calculated using the following formula:

[0090]

[0091] in The distance from the device to the task. For skill matching, This is the adjustment coefficient; For equipment Execute the task The actual time consumed; the probability of the device selecting the task is:

[0092]

[0093] in, To adjust the parameters, the influence of pheromones and distance were controlled separately.

[0094] The distributed negotiation algorithm includes the following negotiation phases:

[0095] Reconnaissance Phase: Randomly select a proportion from all idle devices. The equipment is used as reconnaissance equipment, and these devices actively explore the surrounding available missions;

[0096] Evaluation Phase: The reconnaissance equipment assesses the fitness of the detected missions. If the fitness is higher than a threshold... Then they return to the base to recruit;

[0097] Recruitment Phase: High-fitness reconnaissance equipment recruits other equipment through a pheromone mechanism, with a recruitment quantity... It is determined by the following formula:

[0098]

[0099] in Total number of available devices;

[0100] Execution phase: The recruited devices go to perform the mission and update the pheromones on the path after the mission is completed.

[0101] The resilient scheduling algorithm in step S5 is as follows:

[0102] The algorithm is named the Resilient Scheduling Algorithm Based on Ant Colony Behavior (RS-AB).

[0103] Define the resilience index of the scheduling system for:

[0104]

[0105] in: The number of time periods monitored; For the first Recovery time for each disturbance event; The time decay constant; For the first The actual amount of work completed in a given time period; For the first The planned workload for a given time period; when the system resilience index Below the threshold At that time, the resilience recovery mechanism is triggered.

[0106] The resilience recovery mechanism includes the following steps:

[0107] Identify disturbance types: equipment failure, task change, environmental change, or resource shortage;

[0108] Select a recovery strategy based on the type of disturbance:

[0109] For equipment failures, a replacement strategy should be adopted to replace the equipment. The selection criteria are:

[0110]

[0111] in For the set of available devices, The distance from the replacement equipment to the faulty equipment. For the speed of equipment movement, and These are the capability values ​​for replacing equipment and faulty equipment, respectively.

[0112] For task changes, a task redistribution strategy is adopted, using an improved Hungarian algorithm for fast redistribution;

[0113] In response to environmental changes, a planned adjustment strategy is adopted to recalculate equipment operating times and routes;

[0114] After the recovery strategy is executed, the system resilience metrics are recalculated until... .

[0115] The decision fusion in step S6 adopts a weighted fusion method:

[0116] Output strategic planning solutions at the macro level The weight is ;

[0117] Mid-level output task allocation scheme The weight is ;

[0118] Real-time adjustment scheme for micro-level output The weight is ;

[0119] Final scheduling scheme for:

[0120]

[0121] Where the weight coefficients satisfy Furthermore, the weight values ​​are dynamically adjusted based on the scheduling phase:

[0122] In the initial stage of scheduling, ;

[0123] During the scheduling process, ;

[0124] In the later stages of dispatching or in emergency situations .

[0125] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements an intelligent scheduling method for building equipment based on an AI algorithm.

[0126] This invention first integrates project planning data, historical equipment usage records, and real-time external data sources, including meteorological information and material supply chain status. At this stage, an improved time series forecasting algorithm begins operation. Instead of simply fitting historical curves, it proactively identifies the unique cyclical patterns of construction activities, such as the recurring patterns of concrete pouring and formwork erection, while simultaneously analyzing the potential impact of various external events on equipment demand. The algorithm generates a detailed equipment demand forecast curve for a future period by dynamically weighting and fusing trend components, seasonal components, and causal event components. The advantage of this process is that it transforms traditional passive demand estimation into intelligent forecasting that proactively senses the project's pulse and the external environment. This allows equipment scheduling to be based on forward-looking planning, laying a solid strategic foundation for subsequent precise scheduling and effectively avoiding project delays or equipment idleness caused by insufficient resource prediction.

[0127] Entering the medium-term task intelligent allocation stage, the system starts the improved combinatorial auction algorithm based on the macro prediction and the real-time task list. First, for each task package to be allocated, the algorithm calculates the comprehensive value bids of all available devices or device combinations. This value calculation is highly innovative. It not only evaluates the status, location distance, and energy consumption cost of the devices themselves but also, uniquely, adds a synergistic effect value that quantifies the historical cooperation efficiency of the device combination and the intensity of the current task's requirement for multi-device cooperation. Subsequently, the algorithm simulates the auction process. Through a greedy selection strategy with multiple rounds of iteration, it always gives priority to allocating the device and task combination with the highest current performance-price ratio and updates the available resource pool in real time. The outstanding advantage of this mechanism is that it subverts the isolated allocation mode of "seeing only the trees but not the forest" in traditional scheduling. By modeling the cooperation relationship between devices and endowing it with economic value, the scheduling result can naturally give rise to an efficient cooperation team. Especially in complex scenarios that require multi-machine type linkage, such as earthwork excavation and large component installation, it can significantly improve the overall operation coherence and resource integration efficiency.

[0128] In the micro autonomous execution and adjustment stage, the distributed negotiation algorithm starts to dominate the real-time behavior of the devices. In this stage, it no longer relies on the centralized instructions of the central system. Each device is endowed with the autonomous decision-making ability based on simple rules. The device senses the surrounding environment and the "digital pheromones" released by other devices through local sensors and communication modules. The intensity of this pheromone represents the task attractiveness and completion quality. Like individuals in a bee colony, some devices actively scout for new tasks, and some follow the high-intensity pheromones to the task points that have been verified effective. The whole process includes multiple bionic links such as autonomous scouting, evaluation, recruitment, and execution, forming a continuously operating distributed negotiation network. The core advantage of this decentralized architecture lies in its unparalleled robustness and agility. When unplanned tasks, sudden device failures, or local traffic jams occur on-site, the pheromone distribution will change rapidly, and the surrounding devices can adjust their strategies almost in real time to achieve dynamic rebalancing of tasks. The whole system demonstrates strong self-organization and self-adaptability, ensuring the effective implementation of scheduling instructions in a complex dynamic environment.

[0129] Ultimately, the system ensures overall stability and optimal operation through a resilience monitoring and multi-layered decision-making fusion phase. The resilience scheduling algorithm, acting as the system's immune system, continuously runs, defining and monitoring resilience indicators that reflect system health. If any disturbance causes these indicators to decline, it automatically triggers diagnostic and recovery mechanisms, either quickly matching alternative equipment or initiating a task reallocation process to ensure rapid system performance recovery. Simultaneously, a dynamic weight fusion module is always running. Based on project phase, system load, and stability status, it intelligently adjusts the fusion weights of the outputs from the macro-strategic planning, meso-level task allocation, and micro-level autonomous decision-making layers. For example, it respects the macro-plan during stable periods and relies more on micro-level autonomous negotiation during turbulent periods.

[0130] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent scheduling of building equipment based on AI algorithms, characterized in that, Includes the following steps: S1. Construct a three-tiered progressive decision-making framework for building equipment scheduling, including macro, meso, and micro levels; S2. At the macro level, based on historical and real-time data, an improved time series forecasting algorithm is used to predict trends and make strategic plans for equipment demand at construction sites; S3. At the meso-level, an improved combinatorial auction algorithm is used to optimize the matching of devices and tasks, while also considering the synergistic effect between devices; S4. At the micro level, a distributed negotiation algorithm is used to enable devices to make autonomous decisions and adjust in real time; S5. Monitor the operating status of the scheduling system through a resilient scheduling algorithm and quickly restore system performance when disturbances occur; S6. Integrate the decision results from each layer and output the final equipment scheduling plan.

2. The intelligent scheduling method for building equipment based on AI algorithm according to claim 1, characterized in that: The improved time series prediction algorithm described in step S2 is specifically called the Trend-Seasonal Prediction Adaptive Algorithm. This algorithm adds a feature recognition module specific to building scenes to the traditional ARIMA model. The prediction model is shown below: .

3. The intelligent scheduling method for building equipment based on AI algorithm according to claim 1, characterized in that: The improved combinatorial auction algorithm mentioned in step S3 is specifically: the algorithm is named the combinatorial auction allocation algorithm considering synergy effects; device For task sets The formula for calculating the bid value is: 。 4. The intelligent scheduling method for building equipment based on AI algorithm according to claim 3, characterized in that: The improved combinatorial auction algorithm described above employs a greedy optimization strategy for task allocation: Let the task set be... The equipment set is Initialization allocation result Unassigned task set ; Repeat the following steps until Or unable to continue allocation: a) For each unassigned task subset Calculate the bid value of all devices or groups of devices for this subset of the task; b) Select the allocation combination with the highest cost performance. The formula for calculating cost-effectiveness is as follows: ; c) Assignment combination Add to allocation results ; d) From Remove the assigned task subset Remove from available devices Output the final allocation result. .

5. The intelligent scheduling method for building equipment based on AI algorithm according to claim 1, characterized in that: The distributed negotiation algorithm mentioned in step S4 is specifically: the algorithm name is a distributed negotiation algorithm based on bee behavior; the device is defined. For the task The pheromone intensity is The pheromone update rule is as follows: .

6. The intelligent scheduling method for building equipment based on AI algorithm according to claim 5, characterized in that: The distributed negotiation algorithm includes the following negotiation phases: Reconnaissance Phase: Randomly select a proportion from all idle devices. The equipment is used as reconnaissance equipment, and these devices actively explore the surrounding available missions; Evaluation Phase: The reconnaissance equipment assesses the fitness of the detected missions. If the fitness is higher than a threshold... Then they return to the base to recruit; Recruitment Phase: High-fitness reconnaissance equipment recruits other equipment through a pheromone mechanism, with a recruitment quantity... It is determined by the following formula: ; Execution phase: The recruited devices go to perform the mission and update the pheromones on the path after the mission is completed.

7. The intelligent scheduling method for building equipment based on AI algorithm according to claim 1, characterized in that: The resilient scheduling algorithm described in step S5 is specifically: the algorithm is named the Resilient Scheduling Algorithm Based on Ant Colony Behavior; the resilience index of the scheduling system is defined. for: .

8. The intelligent scheduling method for building equipment based on AI algorithm according to claim 7, characterized in that: The resilience recovery mechanism includes the following steps: Identify disturbance types: equipment failure, task change, environmental change, or resource shortage; Select a recovery strategy based on the type of disturbance: For equipment failures, an equipment replacement strategy should be adopted; For task changes, a task redistribution strategy is adopted, using an improved Hungarian algorithm for fast redistribution; In response to environmental changes, a planned adjustment strategy is adopted to recalculate equipment operating times and routes; After the recovery strategy is executed, the system resilience metrics are recalculated until... .

9. The intelligent scheduling method for building equipment based on AI algorithm according to claim 1, characterized in that: The decision fusion described in step S6 adopts a weighted fusion method: Output strategic planning solutions at the macro level The weight is Mid-level output task allocation scheme The weight is Real-time adjustment scheme for microscopic layer output The weight is Final scheduling scheme for: During the scheduling process, In the later stages of dispatching or in emergency situations, .

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the intelligent scheduling method for building equipment based on AI algorithms as described in any one of claims 1-9.

Images