Multi-work-type collaborative construction deployment optimization method and system

Abstract

The invention discloses a deployment optimization method and system for multi-type collaborative construction, and relates to the technical field of intelligent scheduling, and the method comprises the steps: collecting the pre-construction data of each type of work, and carrying out the hierarchical splitting of a construction task according to the type of work and a working stage, and obtaining construction sub-tasks; constructing a dependency relationship of construction sub-tasks by utilizing topological sorting and a critical path method, and defining a sub-task execution constraint and a work type scheduling constraint by using a constraint planning method; generating a preliminary scheduling scheme by using a heuristic algorithm, and optimizing sub-task execution sorting according to the sub-task execution constraint and the work type scheduling constraint; and dynamically adjusting the execution sorting of the subtasks according to the dependency relationship of the subtasks by using a parallel scheduling algorithm to obtain a scheduling scheme of collaborative construction. According to the method, resource conflicts and delay risks are effectively avoided, high parallelism and load balancing of multiple work types of tasks are achieved, and the construction efficiency and reliability can be remarkably improved in a complex construction environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent scheduling, in particular to a deployment optimization method and system for multi-work-type collaborative construction. BACKGROUND

[0002] In recent years, with the continuous advancement of the intelligentization and informatization process of the construction industry, multi-work-type collaborative construction deployment technology has gradually become a research hotspot. The traditional construction management mode is mainly based on static planning, which is difficult to cope with actual changes on site and difficult to achieve efficient resource collaboration and schedule optimization. In recent years, big data scheduling technology based on BIM (Building Information Modeling) and cloud platforms has developed rapidly, providing visual and traceable management tools for construction sites through the coupling of three-dimensional models, construction progress, and cost data. At the same time, the development of Internet of Things (IoT) technology enables real-time collection of equipment status and personnel location, providing accurate on-site data support for scheduling decisions. On the other hand, scheduling algorithms for mixed integer programming (MIP) and constraint programming (CP) have made important progress in scientific research, achieving a comprehensive balance of schedule, cost, and resource utilization through the establishment of multi-objective optimization models. However, existing research is mostly limited to resource allocation for a single work type or a small number of work types, and mostly uses offline optimization strategies, lacking in-depth exploration of multi-work-type parallel construction and real-time dynamic adjustment.

[0003] Although the above technologies have achieved certain results in improving construction visualization and static optimization, there are still many shortcomings in practical application. First, traditional BIM scheduling and cloud platform management are mainly based on preset parameters, making it difficult to quickly respond to on-site disturbances such as equipment failures, personnel absences, or sudden task changes, resulting in a disconnect between the scheduling plan and the actual situation on site. Second, existing multi-objective scheduling models mostly aim to minimize schedule or cost, ignoring the collaborative constraints and load balancing between different work types, which can easily lead to resource conflicts or idle phenomena. Third, mixed integer programming and constraint programming algorithms have low solving efficiency in large-scale multi-work-type scenarios, making it difficult to meet real-time or near-real-time scheduling needs. SUMMARY

[0004] In view of the above problems, the present application is proposed.

[0005] Therefore, the technical problem solved by the present application is that existing multi-work-type collaborative construction scheduling technology generally has weak dynamic response capability, many resource conflicts, and low scheduling efficiency.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides a deployment optimization method for multi-work-type collaborative construction, comprising,

[0008] Collect pre-construction data for each type of work, and perform hierarchical decomposition of construction tasks according to the type of work and work phase to obtain construction subtasks;

[0009] Use topological sorting and critical path method to build the dependency relationship of construction subtasks, and use constraint programming method to define subtask execution constraints and work type scheduling constraints;

[0010] Generate a preliminary scheduling plan using a heuristic algorithm and optimize the subtask execution order based on subtask execution constraints and work type scheduling constraints;

[0011] A parallel scheduling algorithm is used to dynamically adjust the execution order of subtasks according to their dependencies, and a scheduling solution for collaborative construction is obtained.

[0012] As a preferred solution of the method for optimizing the deployment of multi-type collaborative construction according to the present invention, the hierarchical splitting includes processing pre-construction information of each type of work to obtain construction task information and type of work information, and converting the information into work stage and type of work splitting rules; for each construction task, splitting it into multiple subtasks according to the work stage, each stage including the start time and end time of a complete task; when splitting the work stage, the subtasks are matched with each type of work, and the subtasks are split into multiple subtasks according to the resources required by the type of work;

[0013] Define the properties of each split subtask, including: task name, required resources, duration, time window, and dependencies.

[0014] As a preferred solution of the method for optimizing the deployment of multi-type collaborative construction according to the present invention, the method comprises: creating a node for each subtask, adding directed edges between subtask nodes in the graph according to the dependency relationship between each two subtasks, calculating the in-degree of the node and initializing the task queue; taking a task from the queue Join the execution queue; for each task Dependent subsequent tasks ,Will The entry degree is reduced by 1; if When the in-degree of becomes zero, it is added to the queue; repeat the process to obtain the topological sorting of all subtasks;

[0015] Use the critical path method to calculate the total duration of the project and identify critical tasks; find the critical path by calculating the time window of each subtask, and traverse to generate a critical path task table.

[0016] As a preferred scheme of the deployment optimization method of the multi-variety collaborative construction, wherein: the constraint programming comprises obtaining task time variables and resource variables through subtask attributes, and obtaining constraint conditions by combining execution topologies and dependency relationships among subtasks; for subtasks with dependency relationships and , the subtask execution constraint specifically comprises:

[0017] an execution sequence constraint, if subtask must be executed after subtask is completed, then the end time of subtask must be less than the start time of subtask ;

[0018] a time window constraint, each subtask has a time window , which determines the earliest start time and the latest end time of the subtask, and the start time of the subtask must be within the time window thereof;

[0019] a resource limitation constraint, for the same resource, if two subtasks use the resource in the same time period and the total amount of the resource is greater than a certain amount, then the two subtasks cannot be executed simultaneously;

[0020] the variety scheduling constraint comprises the following conditions that need to be met among subtasks and of different varieties in the same work stage, the earliest start time of is less than the earliest start time of ; for subtasks and of different stages, the earliest start time of is greater than the latest end time of ; wherein, and represent different work levels obtained by division, and is prior to .

[0021] As a preferred scheme of the deployment optimization method of the multi-variety collaborative construction, wherein: the generating of the preliminary deployment scheme by using the heuristic algorithm comprises: setting each individual to represent a selectable local deployment scheme, the gene structure of the individual comprises subtask start time, time window and resource allocation, and the order of the gene sequence is the execution order of the subtasks;

[0022] initializing the population, randomly generating the gene structure, and following the subtask execution constraint and the variety scheduling constraint;

[0023] A fitness function is constructed, and the function target includes minimizing total duration, resource use efficiency; the violation degree of the penalty term is calculated comprehensively and added to the target value to obtain the fitness evaluation;

[0024] The selection is performed according to the fitness evaluation sequence by using a double selection elimination method; the preliminary selection is performed by using random selection, and the selection is divided into a preliminary winning group and a preliminary losing group; the individual with higher fitness is taken into the preliminary winning group, and the rest is taken into the preliminary losing group; the secondary selection is performed by using a high-order selection method, and the individuals with high order and low order in the same group are compared; in the preliminary winning group, the winning individual is reserved, and the loser is taken into the loser group; in the preliminary losing group, the winning individual is taken into the loser group, and the loser is eliminated; the winner is reserved by using the high-order selection method in the loser group, and the loser is eliminated; the individuals reserved in the preliminary winning group and the loser group jointly constitute a new generation population;

[0025] The offspring individual is generated by using a sequential crossover method; the start time sequence of the generated subtask needs to meet the time window constraint and the execution sequence constraint during the execution of the crossover; the solution after the crossover is verified by constraint programming, and the constraint conflict item is screened out; the mutation operation is performed; the start time in the subtask execution sequence is adjusted within the time window constraint, and the resource allocation of the task is adjusted within the resource limitation constraint to obtain the offspring individual.

[0026] As a preferred scheme of the deployment optimization method for the multi-work-specification collaborative construction, the optimization subtask execution ordering includes, when a new individual is generated in each generation, verifying whether the generated scheduling scheme meets all constraint conditions by constraint programming; if the constraint is violated, the individual is punished and adjusted again; when the maximum number of iterations is reached, the iteration is stopped, and the gene structure corresponding to the individual with the highest fitness is finally output; the execution ordering of the optimized subtask is obtained by analyzing the gene structure, and the preliminary scheduling scheme is obtained according to the subtask execution ordering.

[0027] As a preferred scheme of the deployment optimization method for the multi-work-specification collaborative construction, the parallel scheduling algorithm includes, reconstructing a dependency relationship graph according to the preliminary scheduling scheme, and analyzing parallel execution conditions; initializing a scheduling list to record the start time, end time and allocated resources of each subtask; selecting a subtask that has no predecessor task and whose predecessor task has been completed from the reconstructed topological ordering, and adding the subtask to a parallel execution subtask set.

[0028] The process of parallel task allocation specifically includes:

[0029] In the first round of scheduling, a subtask is selected from the parallel execution subtask set, resource allocation is performed on the subtask, and the start time and end time of the subtask are calculated;

[0030] Through multiple rounds of scheduling, the subtasks in the task graph are gradually selected, scheduling is performed according to the dependency relationship and priority, and no resource conflict is detected.

[0031] Verify whether the constraint planning is met, keep load balance for scheduling all parallel tasks, after each scheduling is completed, dynamically update the start time and end time of the to-be-scheduled subtasks according to the subtask execution, and obtain a final scheduling scheme of collaborative construction.

[0032] In a second aspect, the present application provides a multi-task collaborative construction deployment optimization system, comprising,

[0033] A task allocation module, a progress tracking module and a conflict management module.

[0034] The task allocation module is used for calculating the task list, execution time and resource requirement of each task type, and performing subtask allocation.

[0035] The progress tracking module is used for updating the task progress in real time, and ensuring the progress synchronization of each task type with other task types.

[0036] The conflict management module is used for providing an automatic optimization scheme when conflicts occur between task types, and supporting manual intervention adjustment.

[0037] In a third aspect, a computer device comprises a memory and a processor, the memory stores a computer program, and the processor executes the computer program to implement the steps of the multi-task collaborative construction deployment optimization method.

[0038] In a fourth aspect, a computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the multi-task collaborative construction deployment optimization method.

[0039] The multi-task collaborative construction deployment optimization method provided by the present application can finely generate construction subtasks by hierarchically splitting the pre-construction data of each task type, and can make the task execution order and construction period evaluation more accurate based on the topological sorting and critical path method to construct the dependency relationship; the constraint programming method is introduced to uniformly define the execution order, time window and resource limitation constraint, so as to effectively avoid resource conflicts and delay risks; the genetic algorithm and other heuristic strategies are used to quickly generate a preliminary scheduling scheme that meets various constraints, and the parallel scheduling algorithm is used to dynamically adjust the subtask execution order, so as to realize high parallelism and load balance of multi-task. The method has the advantages of shortening the construction period, improving the resource utilization rate, intelligently scheduling the scheme and having strong real-time response capability, and can significantly improve the construction efficiency and reliability in a complex construction environment. BRIEF DESCRIPTION OF DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor based on these drawings.

[0041] Figure 1 The overall flow chart of a multi-work-type collaborative construction deployment optimization method provided for the first embodiment of the present application. DETAILED DESCRIPTION

[0042] In order to make the above-mentioned purposes, features and advantages of the present application more apparent and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings in the specification. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should be within the protection scope of the present application.

[0043] Embodiment 1, refer to Figure 1 For an embodiment of the present application, a multi-work-type collaborative construction deployment optimization method is provided, which comprises:

[0044] S1: Collect pre-construction data of each work type, and perform hierarchical splitting on construction tasks according to work types and work stages to obtain construction sub-tasks.

[0045] Further, the pre-construction data is collected to obtain an original construction task set and corresponding pre-construction information, which comprises:

[0046] A task belonging work type set (such as civil engineering, installation, electrical, etc.). A task corresponding work stage set (such as foundation, main body, finishing, etc.). A predicted construction period, required resources, available time window, etc.

[0047] Define splitting rules, for each original task, perform hierarchical splitting according to work stages and work types , generate a construction sub-task for each pair , and obtain a sub-task set .

[0048] The hierarchical splitting includes processing each type of pre-construction information to obtain construction task information and type information, and converting into work stage and type splitting rules; for each construction task, splitting into multiple sub-tasks according to the work stage, and each stage containing the start time and end time of a complete task; when splitting the work stage, considering that the split sub-tasks match each type, and splitting into multiple sub-tasks according to the required resources of the type.

[0049] For each construction task, it is first divided according to its work stage. Each stage usually has a start time and end time. For example, the "civil construction" task in a building project can be divided into foundation construction, reinforcement installation, concrete pouring, etc.

[0050] The tasks of each type depend on the corresponding work stage. For example, "electrician" may need to perform wiring, equipment installation, etc., while "civil engineer" may need to perform piling, foundation pouring, etc. Define the criteria for splitting tasks for each type. For example, if a type has multiple operation steps (such as concrete construction including concrete pouring, compaction, etc.), these operations are split into sub-tasks.

[0051] Split tasks according to stages: each construction task is decomposed according to its work stage. For example, the "foundation construction" task of a building project can be split into: foundation excavation, foundation cushion construction, foundation pouring, etc.; the tasks split out of each stage are sub-tasks.

[0052] Assign types to each sub-task: according to the needs of each task stage, select the appropriate type. For example, "civil engineer" is assigned to the foundation excavation task; "reinforcement worker" is assigned to the reinforcement installation task; "electrician" is assigned to the electrical wiring task. Here, it is necessary to determine which types are needed for each task according to the needs of the construction site.

[0053] Calculate the duration and resource requirements: according to the split sub-tasks, calculate the duration of each sub-task (usually estimated according to historical data or experience), and clearly define the resources required for the task. For example, the "foundation pouring" task may require a concrete mixer, concrete, reinforcement, etc. The availability of resources (such as the number of equipment, the skills of personnel, etc.) will affect the scheduling of the task.

[0054] Define the attributes of each split sub-task, including: task name, required resources, duration, time window, dependency relationship.

[0055] Each level of sub-task has its own time window (TimeWindow) and resource requirements.

[0056] The dependency relationship of sub-tasks (such as "foundation pouring" depends on "foundation cushion construction") is reflected in the tree structure.

[0057] The final output is a set of all subtasks, each with explicit attributes including task name, required resources, duration, time window, dependencies, etc.

[0058] Subtask dependency graph: Generate a directed acyclic graph (DAG) representing the dependencies between subtasks. Each node is a subtask, and edges represent dependencies.

[0059] It should be noted that the algorithm splits construction tasks hierarchically, refines them by trade and work stage, splits complex construction tasks into multiple subtasks, and assigns resources, duration, time window, etc. to each subtask. Through this hierarchical task splitting method, the accuracy and flexibility of task scheduling can be effectively improved, providing a foundation for subsequent task sequencing, resource allocation and scheduling optimization.

[0060] S2: Use topological sorting and critical path method to build the dependency relationship of construction subtasks, and use constraint programming method to define subtask execution constraints and trade scheduling constraints.

[0061] Further, topological sorting is used to sort the directed acyclic graph (DAG) so that each subtask is executed after its dependent tasks are completed. The dependency relationship between construction tasks can be modeled as a directed graph.

[0062] The dependency relationship of the subtasks includes that the former (preceding task) must be completed before the latter (subsequent task) starts or ends to ensure that the subsequent task starts under the correct premise.

[0063] Input a set of construction subtasks Each subtask has a dependency relationship with other subtasks (for example, task must be completed before can start).

[0064] Build a directed acyclic graph (DAG) and create a node for each subtask. According to the dependency relationship between subtasks (for example, task depends on task ), add directed edges between tasks in the graph. Ensure that the graph has no loops, as a loop will cause circular dependencies and cannot be executed.

[0065] Calculate the in-degree of each node: the in-degree refers to how many tasks depend on the current task. For example, the in-degree of task is 1 (it depends on ). For each subtask , calculate its in-degree, which represents how many other tasks depend on it. Initialize the task queue: put all the sub-tasks with in-degree zero (no other tasks depend on them) into a queue. These tasks are the ones that can be executed first.

[0066] The dependency relationship of the construction sub-tasks includes creating a node for each sub-task, adding directed edges between the sub-task nodes in the graph according to the dependency relationship between each two sub-tasks, calculating the in-degree of the nodes and initializing the task queue; taking a task from the queue and adding it to the execution queue; for each subsequent task that depends on the task , reduce the in-degree of the task by 1; if the in-degree of the task becomes zero, add it to the queue; repeat the execution to get the topological sorting of all sub-tasks.

[0067] Calculate the total duration of the project and identify the critical tasks using the critical path method; find the critical path by calculating the time window of each sub-task, and traverse to generate the critical path task table.

[0068] The execution order of the construction sub-tasks (obtained through topological sorting), and the work steps of each sub-task:

[0069] For each task, initialize the earliest start time (ES) to 0 and the latest finish time (LF) to infinity. Calculate the earliest start time (ES) by traversing the topologically sorted tasks and calculating the earliest start time (ES) of each task. For each task , its earliest start time is the latest finish time of all its predecessor tasks (i.e. the earliest finish time of the predecessor tasks) plus the duration of the task.

[0070] Calculate the latest finish time (LF): start from the last task and calculate the latest finish time (LF). For each task , its latest finish time is the earliest start time of all its successor tasks minus the duration of the task.

[0071] Calculate the critical path, which is the path on which the earliest start time (ES) and the latest finish time (LF) are equal.

[0072] Output the earliest start time (ES), the latest finish time (LF), and the critical path task list.

[0073] The constraint planning includes obtaining task time variables and resource variables through sub-task attributes, and obtaining constraint conditions by combining the execution topology and dependency relationship between sub-tasks; for sub-tasks and , the sub-task execution constraints specifically include:

[0074] execution order constraint, if a subtask must be executed after the completion of a subtask , then the end time of the subtask must be less than the start time of the subtask .

[0075] time window constraint, each subtask has a time window , which determines the earliest start time and the latest end time of the subtask, and the start time of the subtask must be within its time window.

[0076] resource limitation constraint, for the same resource, if two subtasks use the resource in the same time period, and the total usage is greater than the total amount of the resource, then the two subtasks cannot be executed at the same time.

[0077] The start time, end time and duration of each subtask are key variables. Set the following variables: : the start time of subtask . : the end time of subtask . : the duration of subtask (fixed, known).

[0078] For each resource (such as equipment, personnel), define its availability: : represents whether the resource is available at time point . : represents whether the resource is occupied by subtask .

[0079] Define the constraint condition, execution order constraint: the execution order between subtasks is determined according to the dependency relationship. For subtasks and with dependency relationship, the following constraints need to be met: if subtask must be executed after the completion of subtask (i.e. is the predecessor task of ), then there needs to be:

[0080]

[0081] Here, the end time of subtask must be less than the start time of subtask .

[0082] Time window constraint: each subtask has a time window which specifies the earliest start time and the latest end time of the task. The constraint is as follows:

[0083]

[0084] The start time of subtask start_i must be within its time window.

[0085] Resource limitation constraint: resource limitation constraint ensures that no resource conflict occurs during the execution of different subtasks. Assuming that resource r needs to be occupied when subtask i is executed, the constraint is as follows:

[0086] Resource occupation constraint:

[0087]

[0088] Resource availability constraint:

[0089]

[0090] wherein, is the maximum capacity of resource r, indicating the maximum available number of the resource at each time t, avoiding over-allocation of resources.

[0091] Non-overlapping constraint: for the same resource r, different subtasks i and j need to meet the following condition if they use the resource in the same time period:

[0092]

[0093] That is, subtasks i and j must not overlap in time, otherwise they cannot share the same resource.

[0094] The said work type scheduling constraint includes that the earliest start time of different work type subtasks and needs to meet, the earliest start time of ; for different stage subtasks and , the earliest start time of is greater than the latest end time of ; wherein, and indicate different work levels obtained by division, and precedes .

[0095] In the actual construction process, multiple types of work may be alternately performed at different stages, and therefore, each task needs to be hierarchically divided according to its type, and the scheduling of tasks between different types needs to be ensured not to conflict. The specific constraint conditions are as follows: type hierarchical division: it is assumed that the task set has been divided into multiple levels according to the type, and each level contains one or more subtasks (such as the types of civil engineering, equipment installation, decoration, etc.).

[0096] For example, if task belongs to the type level, and task belongs to the type level, the coordination between types needs to be ensured during scheduling:

[0097]

[0098] This indicates that some types (such as civil engineering) must be completed before the subsequent type (such as equipment installation) can begin. Type priority: for different levels of types, there may be priority constraints. It is assumed that the task of type must be completed before the task of type , and the constraint is:

[0099]

[0100] This ensures that the tasks of some types are executed in order and do not conflict with the execution across types. Dependency constraint between types: in multi-type construction, the tasks of some types must wait for the completion of other types. The dependency relationship between types can be defined to ensure coordination between types as needed. For example, civil engineering work must be completed before equipment installation, and then:

[0101]

[0102] where and represent the start time of the equipment installation task and the end time of the civil engineering foundation task, respectively.

[0103] It should be noted that by introducing the constraints of type hierarchical division, the tasks of different types can be executed in the appropriate order while avoiding resource conflicts between tasks. The above constraint conditions include the execution order constraint of subtasks, the time window constraint, the resource limitation constraint, and the scheduling constraint for the type hierarchy structure, which can effectively ensure that each task is performed in a coordinated and efficient manner while meeting the dependency relationship and resource limitations.

[0104] S3: generate a preliminary scheduling scheme using a heuristic algorithm, and optimize the execution order of subtasks according to the subtask execution constraints and type scheduling constraints.

[0105] Further, the generating the preliminary scheduling scheme by using the heuristic algorithm comprises: each individual represents a selectable local scheduling scheme, a gene structure of the individual comprises a subtask start time, a time window and a resource allocation, and an order of the gene sequence is an execution order of the subtasks.

[0106] Initializing a population, and randomly generating a gene structure in compliance with a subtask execution constraint and a worker scheduling constraint;

[0107] Constructing a fitness function, a function target comprises minimizing a total duration and a resource use efficiency, a violation degree of a penalty term is comprehensively calculated and added to a target value to obtain a fitness evaluation.

[0108] The method for double selection and elimination is selected according to a fitness evaluation sequence; a preliminary selection uses random selection, and is divided into a preliminary winning group and a preliminary losing group, individuals with higher fitness are taken into the preliminary winning group, and the rest are taken into the preliminary losing group; a secondary selection uses a sequential selection method, individuals with high and low sequences in the same group are compared, individuals selected twice are reserved, and individuals discarded twice are eliminated; in the preliminary winning group, winning individuals are reserved, and losers are taken into a loser group; in the preliminary losing group, winning individuals are taken into the loser group, and losers are eliminated; in the loser group, a sequential selection method is used, winning individuals are reserved, and losers are eliminated; the individuals reserved in the preliminary winning group and the loser group jointly constitute a new generation population.

[0109] A sequential crossover method is used to generate offspring individuals, a subtask start time sequence generated when the crossover is performed needs to meet time window constraints and execution sequence constraints, and the solution after the crossover is verified through constraint programming to filter out constraint conflict items; a mutation operation is performed, the start time in the subtask execution sequence is adjusted within the time window constraint, and the resource allocation of the task is adjusted within the resource limitation constraint to obtain offspring individuals. The initial round of random grouping and the second round of sequential confrontation ensure that individuals at different fitness levels have the opportunity to "revive", avoiding the loss of gene diversity caused by directly eliminating all low-mid fitness solutions.

[0110] The optimizing the subtask execution sequence comprises: when a new individual is generated in each generation, whether the generated scheduling scheme meets all constraint conditions is verified through constraint programming; if the constraint is violated, the individual will be punished and adjusted again; when a maximum iteration number is reached, the iteration is stopped, and finally the gene structure corresponding to the individual with the highest fitness is output; the optimized subtask execution sequence is obtained by analyzing the gene structure, and the preliminary scheduling scheme is obtained according to the subtask execution sequence.

[0111] It should be noted that the proposed method combines a modified genetic algorithm with constraint programming. When generating preliminary scheduling solutions, the genetic algorithm performs a global search, while constraint programming verifies that each individual meets task dependencies, time windows, resource constraints, and work-level constraints. This combination fully leverages the genetic algorithm's global search capabilities while ensuring that the scheduling solution meets actual construction requirements and various constraints, thus providing an effective optimization solution for complex, multi-work construction scheduling.

[0112] S4: Use a parallel scheduling algorithm to dynamically adjust the subtask execution order according to the subtask dependencies to obtain a scheduling plan for collaborative construction.

[0113] Furthermore, the parallel scheduling algorithm includes reconstructing the dependency graph based on the preliminary scheduling plan and analyzing the parallel execution status; initializing a scheduling list, recording the start time, end time and allocated resources of each subtask, and selecting subtasks that currently have no predecessor tasks or have completed predecessor tasks from the reconstructed topological sorting, and adding them to the parallel execution subtask set.

[0114] Read the reconstruction subtask set, each subtask Including construction period , resource requirements , time window and precursor collection .

[0115] Read resource set , each resource Available quantity at each moment . Build dependency graph and topological sort, build directed acyclic graph (DAG), and nodes are tasks ,side If and only if .

[0116] Perform topological sorting on the DAG to obtain a linear sequence , ensuring that all predecessors go first.

[0117] Critical Path Method (CPM) pre-calculation, in topological order calculate:

[0118]

[0119] Inverse topology calculation:

[0120]

[0121] in for The direct successor set of . Initialize the ready queue and put all the predecessor sets The task is added to the ready queue . Record the current global time pointer .

[0122] Parallel task allocation is performed, specifically including:

[0123] First round of scheduling, select subtasks from the set of parallel execution subtasks, allocate resources to the subtasks, and calculate their start time and end time.

[0124] Multi-round scheduling, gradually select subtasks in the task graph, schedule according to dependency relationship and priority, and detect no resource conflict.

[0125] All subtasks without predecessor dependency (i.e. ) and the current time pointer fall within their time window are added to the ready queue . At this time, the global time pointer is set to zero, and all resources are ready according to their initial available amount .

[0126] Ready task identification in each round of scheduling, first check each subtask in the queue :

[0127] Determine whether the current time has reached its earliest start time . Confirm that the current time is still within its allowed execution time window . Confirm that the current available amount of resources is sufficient to meet the needs of the task . The task that meets the above three conditions can be selected as a candidate set that can be executed in parallel in this round.

[0128] Task parallel allocation is performed simultaneously for all candidate tasks:

[0129] Set the actual start time of the candidate task to the larger value of the current time and . Deduct the required resources of each task from the global resource pool to ensure that the resources are not occupied by other tasks within the interval . Calculate and record their completion time .

[0130] Ready queue update, after each round of allocation, check the just completed task and all its successor tasks: if all predecessors of a successor task have been completed, insert it into so that it enters the next round of candidate check. At the same time, remove the tasks that have been executed, and keep Only tasks that have not been scheduled or have not completed all predecessors are included in the global time.

[0131] The global time is updated to the earliest completion time among the scheduled tasks in the current round. This not only advances the progress, but also ensures that subsequent rounds can discover new executable tasks in time. If there is resource contention or exceeds the parallelism limit at the same time point, the candidate tasks in the ready queue are reordered according to the following priority function:

[0132]

[0133] where a larger indicates a more urgent or less dependent task, which should be allocated resources first. Repeat the steps until the ready queue is empty - at this point, all subtasks have been scheduled to start and allocated completion times, and the scheduling loop ends.

[0134] Verify whether the constraint plan is met, keep load balancing for all parallel tasks, and dynamically update the start time and end time of the to-be-scheduled subtasks according to the execution of the subtasks after each scheduling is completed, to obtain the final scheduling scheme of collaborative construction.

[0135] It should be noted that the main steps of the parallel scheduling algorithm are to calculate the dependency relationship of tasks through topological sorting and critical path method, identify tasks that can be executed in parallel, then dynamically adjust the execution order of tasks based on resource and time window constraints, and perform parallel task scheduling. Through load balancing, resource conflict detection and scheduling optimization, it is ensured that all tasks can be completed on time, reducing the construction period and cost, and improving the construction efficiency. The core advantage of this algorithm is that it can dynamically adjust the task execution order according to the actual construction situation, and fully utilize the available resources to maximize parallelism, thereby shortening the overall construction period.

[0136] Embodiment 2 provides a multi-workpiece collaborative construction scheduling optimization method, which is an embodiment of the present application. In order to verify the beneficial effects of the present application, economic benefit calculation and simulation experiments are carried out for scientific demonstration.

[0137] First, this embodiment selects a municipal comprehensive pipe gallery construction project in a certain city as the verification object, and the total task of the project is 50, covering three types of work, namely civil engineering, pipeline installation and electrical work. First, through historical project data and on-site survey data, pre-construction information of each type of work is collected, including task duration estimation (unit: days), number of personnel required (person / day), main equipment model and available period, material supply cycle and construction site environment constraints, etc., to form a pre-construction database. According to the "hierarchical splitting rule by work type and work phase", 50 construction tasks are split into 150 sub-tasks, and each sub-task defines five attributes: task name (such as "pipeline foundation excavation_phase 1"), required resources (personnel, excavator, concrete pump truck), duration (days), time window (earliest start, latest end) and dependency relationship (indicated by the number of previous tasks).

[0138] After the data preparation is completed, the self-developed script is used to import the sub-task information into the constraint programming solver (Google OR-Tools CP-SAT), and the topological sorting algorithm (based on directed acyclic graph) is used to automatically generate a preliminary execution sequence. Subsequently, the critical path method (CPM) is used to calculate the earliest start time (ES), the latest end time (LF) and the total duration of the sequence, and to identify the critical task set. The constraint programming part takes the execution order constraint, time window constraint, resource limitation constraint and work type scheduling constraint as the core to build a mathematical model: the execution order constraint ensures that the sub-task can start after the parent task ends; the time window constraint limits the start time of the sub-task within the preset interval; the resource limitation constraint ensures that the total usage of each resource within the same time period does not exceed the standard; and the work type scheduling constraint specifies the sequence relationship of different work types in the same phase.

[0139] Based on the constraint model, a genetic algorithm is used as a heuristic to generate a preliminary scheduling scheme. The population size is set to 50 individuals, and the chromosome genes include sub-task start time, allocated resource identifier and phase flag; the fitness function consists of two parts: "total duration minimization" and "resource utilization rate maximization", and a penalty term is added to individuals that violate the constraints. Evolution iteration is performed for 100 generations, and the crossover operation uses sequential crossover, and the mutation operation fine-tunes the gene value within the time window range. After crossover and mutation, the constraint programming module is called to verify and filter out conflicting individuals. Finally, several local optimal scheduling schemes are obtained, and the one with the highest fitness is recorded as the preliminary scheme.

[0140] For the preliminary scheme, further use parallel scheduling algorithm for dynamic optimization. Based on the preliminary scheme, the subtask dependency graph is reconstructed, and in a multi-thread environment, the executable subtask set is selected according to the topological sorting, and parallel task allocation is performed. In each round of scheduling, the resources of each subtask in the current executable set are allocated and the start and end times are calculated, and then the actual scheduling result is used to dynamically update the executable set of the next round. The process is repeated until all subtasks are completed. In order to ensure data reliability, seven schemes including Baseline1 (only CPM), Baseline2 (constraint programming + CPM), Baseline3 (heuristic scheduling), and Proposal1~Proposal4 (different parameter combinations of the method) are executed 10 times, and the average value is taken as the final experimental data.

[0141] Compared with Baseline1 (only using critical path method), Baseline2 (integrating constraint programming) shortens the total project duration by about 10 days and improves the resource utilization by 8 percentage points, verifying the effect of constraint programming on resource bottleneck relief; Baseline3 (only heuristic scheduling) further improves the duration and resource utilization, indicating that genetic algorithm has significant effect in task ordering and resource allocation optimization. In contrast, Proposal1~Proposal4 have the most obvious performance improvement after introducing parallel scheduling algorithm: taking Proposal4 as an example, the total project duration is only 89 days, which is 31 days less than Baseline1 (25.8% reduction), and the resource utilization reaches 88%, which is improved by 26 percentage points; the number of overtime subtasks decreases from 20 to 4, and the average task delay decreases from 2.5 days to 0.6 days, significantly reducing the delay caused by process waiting and resource conflict. The parallelism index increases from 3.2 of Baseline1 to 7.8 of Proposal4, indicating that the parallel scheduling algorithm can maximize the parallel potential of subtasks, shorten the critical path length (60 days) and improve the overall scheduling efficiency (90%).

[0142] Embodiment 3 provides a multi-trade collaborative construction deployment optimization system, which is an embodiment of the present application, comprising a task allocation module, a progress tracking module, and a conflict management module.

[0143] The task allocation module is configured to calculate the task list, execution time, and resource requirement of each trade, and perform subtask allocation.

[0144] The progress tracking module is configured to update the task progress in real time and ensure that the progress of each trade is synchronized with that of other trades.

[0145] The conflict management module is configured to provide an automatic optimization scheme when conflicts occur between trades and support manual intervention adjustment.

[0146] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the parts of the present application that essentially contribute to the prior art or the parts of the technical solutions of the present application can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0147] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a list of executable instructions for implementing logic functions, which can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus or device, such as a computer-based system, a system including a processor or other system that can fetch instructions from an instruction execution system, apparatus or device and execute the instructions, or in conjunction with these instruction execution systems, apparatus or devices. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transport programs for use by or in connection with an instruction execution system, apparatus or device, or in conjunction with these instruction execution systems, apparatus or devices.

[0148] More specific examples (non-exhaustive list) of the computer-readable medium include the following: an electrical connection having one or more wires (electrical devices), a portable computer diskette (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CD ROM). In addition, the computer readable medium can even be paper or other suitable medium on which the program can be printed, because the program can be obtained electronically, for example, by optical scanning of the paper or other medium, followed by editing, interpreting or otherwise processing, if necessary, in other suitable ways, to be electronically obtained and then stored in the computer memory.

[0149] It should be understood that portions of the present application can be implemented in hardware, software, firmware, or combinations thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, implementation can be with any or a combination of the following technologies, which are all well known in the art: a discrete logic circuit having logic gates for implementing logic functions upon an application of data signals, an application specific integrated circuit having appropriate combinational logic gates, a programmable gate array (PGA), a field programmable gate array (FPGA), etc. It should be understood that the above-described embodiments are merely given as examples of the technical solution of the present application and are not to be interpreted in a limiting manner, and that for persons skilled in the art, modifications or equivalent replacements can be made to the technical solution of the present application without departing from the spirit and scope of the present application, and all such modifications or equivalent replacements shall be encompassed in the scope of the claims of the present application.

Claims

1. A method for optimizing the deployment of multi-type collaborative construction, characterized in that: include: Collect pre-construction data for each type of work, and perform hierarchical decomposition of construction tasks according to the type of work and work phase to obtain construction subtasks; Use topological sorting and critical path method to build the dependency relationship of construction subtasks, and use constraint programming method to define subtask execution constraints and work type scheduling constraints; Generate a preliminary scheduling plan using a heuristic algorithm and optimize the subtask execution order based on subtask execution constraints and work type scheduling constraints; A parallel scheduling algorithm is used to dynamically adjust the execution order of subtasks according to their dependencies, and a scheduling solution for collaborative construction is obtained.

2. The method for optimizing the deployment of multi-type collaborative construction according to claim 1, characterized in that: The hierarchical splitting includes processing pre-construction information of each type of work to obtain construction task information and type of work information, and converting the information into work stage and type of work splitting rules; for each construction task, splitting it into multiple subtasks according to the work stage, each stage including the start time and end time of a complete task; when splitting the work stage, considering that the split subtasks match each type of work, splitting it into multiple subtasks according to the resources required by the type of work; Define the properties of each split subtask, including: task name, required resources, duration, time window, and dependencies.

3. The method for optimizing the deployment of multi-type collaborative construction according to claim 2, characterized in that: The construction of the dependency relationship of the construction subtasks includes creating a node for each subtask, adding directed edges between the subtask nodes in the graph according to the dependency relationship between each two subtasks, calculating the in-degree of the node and initializing the task queue; taking a task from the queue Join the execution queue; for each task Dependent subsequent tasks ,Will The entry degree is reduced by 1; if When the in-degree becomes zero, it is added to the queue; Repeat the execution to obtain the topological sorting of all subtasks; and Indicates the index of the subtask; Use the critical path method to calculate the total duration of the project and identify critical tasks; find the critical path by calculating the time window of each subtask, and traverse to generate a critical path task table.

4. The method for optimizing the deployment of multi-type collaborative construction according to claim 3, characterized in that: The constraint planning includes obtaining task time variables and resource variables through subtask attributes, and obtaining constraint conditions by combining the execution topology and dependency relationships between subtasks; for subtasks with dependencies, and ,Subtask execution constraints specifically include: Execution order constraints, if the subtask Must be in subtask After the completion of the execution, the subtask The end time of the task must be less than the subtask The start time of Time window constraint, each subtask There is a time window , determine the earliest start time of the subtask and the latest end time ,The start time of the subtask must be within its time window; Resource limitation constraints: for the same resource, if two subtasks use resources in the same time period, if the resource usage exceeds the total resource amount, the two subtasks cannot be executed at the same time; The work type scheduling constraints include subtasks of different work types in the same work phase and Time needs to be met, The earliest start time is less than The earliest start time; for subtasks at different stages and , The earliest start time is greater than The latest end time; and Represents the different levels of work obtained by division, and Before .

5. The method for optimizing the deployment of multi-type collaborative construction according to claim 4, characterized in that: The method of generating a preliminary scheduling plan using a heuristic algorithm includes assuming that each individual represents an optional local scheduling plan, the gene structure of the individual includes the subtask start time, time window and resource allocation, and the order of the gene sequence is the execution order of the subtasks; Initialize the population and randomly generate the genetic structure, following the subtask execution constraints and job scheduling constraints; Construct a fitness function whose objectives include minimizing the total construction period and maximizing resource utilization efficiency. Calculate the degree of violation of the penalty term and add it to the target value to obtain a fitness evaluation. The double selection elimination method is used to select based on the fitness evaluation sequence; the initial selection uses random selection to divide the group into the initial winning group and the initial losing group. Individuals with higher fitness enter the initial winning group, and the rest enter the initial losing group; the secondary selection uses the ranking selection method to select high-ranking and low-ranking individuals in the same group for comparison. In the initial winning group, the winning individuals are retained, and the losers enter the loser group; in the initial losing group, the winning individuals enter the loser group, and the losers are eliminated; in the loser group, the ranking selection method is used to retain the winning individuals and eliminate the losers; the individuals retained in the initial winning group and the loser group together constitute the new generation population; The offspring individuals are generated using the sequential crossover method. The subtask start time sequence generated during crossover must satisfy the time window constraints and execution order constraints. The solution after crossover is verified through constraint programming to filter out constraint conflict items. The mutation operation is performed to adjust the start time in the subtask execution sequence within the time window constraints and the resource allocation of the task within the resource limit constraints to obtain the offspring individuals.

6. The method for optimizing the deployment of multiple types of work in collaborative construction according to claim 5, characterized in that: The optimization subtask execution order includes, when generating new individuals in each generation, verifying whether the generated scheduling plan meets all constraints through constraint programming; if the constraints are violated, the individual will be penalized and readjusted; When the maximum number of iterations is reached, the iteration is stopped and the genetic structure corresponding to the individual with the highest fitness is finally output; the genetic structure is analyzed to obtain the optimized execution order of the subtasks, and a preliminary scheduling plan is obtained based on the execution order of the subtasks.

7. The method for optimizing the deployment of multiple types of work in collaborative construction according to claim 6, characterized in that: The parallel scheduling algorithm includes reconstructing a dependency graph based on a preliminary scheduling plan and analyzing the parallel execution status; initializing a scheduling list, recording the start time, end time and allocated resources of each subtask, and selecting subtasks that currently have no predecessor tasks or whose predecessor tasks have been completed from the reconstructed topological sorting, and adding them to the parallel execution subtask set; The process of parallel task allocation specifically includes: In the first round of scheduling, subtasks are selected from the set of parallel execution subtasks, resources are allocated to the subtasks, and their start and end times are calculated; Through multiple rounds of scheduling, subtasks in the task graph are gradually selected and scheduled according to dependencies and priorities, and no resource conflicts are detected; Check whether the constraint planning is satisfied and schedule all parallel tasks to maintain load balance. After each scheduling is completed, the start and end times of the subtasks to be scheduled are dynamically updated according to the execution status of the subtasks to obtain the final scheduling plan for collaborative construction.

8. A system using the deployment optimization method for multi-type collaborative construction according to any one of claims 1 to 7, characterized in that: Including task allocation module, progress tracking module, and conflict management module; The task allocation module is used to calculate the task list, execution time, resource requirements of each type of work, and perform subtask allocation; The progress tracking module is used to update the task progress in real time to ensure that the progress of each type of work is synchronized with the progress of other types of work; The conflict management module is used to provide automatic optimization solutions when conflicts arise between work types and supports manual intervention and adjustment.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for optimizing the deployment of multi-trade collaborative construction according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for optimizing the deployment of multi-trade collaborative construction according to any one of claims 1 to 7 are implemented.

Images