An intelligent optimization method for rapid construction organization design of large-scale projects

Abstract

The present invention discloses an intelligent compilation and optimization method for rapid construction organization design of large-scale projects, which relates to the technical field of construction organization design compilation. The method solves the problem of lack of effective construction organization design compilation methods in current large-scale projects with few people. The method aims at minimizing costs for generalized large-scale projects that may have multiple types of processes and discontinuous execution times of activities in the same process, and has the following beneficial effects: an efficient encoding and decoding method decouples the two problems of solution mode selection and solution activity start time in multi-mode resource-constrained project scheduling problems, effectively improves the operation efficiency of the algorithm, adaptively selects compilation strategies, adaptively allocates appropriate compilation strategies to newly generated individuals, reduces the computational consumption of invalid individuals, and can effectively improve the solution quality, shorten the local search technology of the critical path, and reduce it without changing the critical path as much as possible, thereby effectively achieving the purpose of compressing the construction period.

Description

Technical Field

[0001] The invention relates to the technical field of construction organization compilation, and in particular to an intelligent compilation optimization method for rapid construction organization design of a large-scale project. Background Art

[0002] In today's society, with the continuous development and application of intelligent technology and automation technology, the engineering field is experiencing widespread penetration of automated equipment, which is manifested in the reduction of labor costs and the increase in the number of mechanized equipment. Large-scale engineering construction is gradually moving towards less-manned and unmanned. Taking the construction of bridges in railway engineering as an example, most of its processes require the use of automated equipment such as loaders, beam transporters, and bridge erection machines for construction. However, in such large-scale, less-manned projects, the steps required for engineering construction are extremely complicated, and the number of large-scale machinery is very limited. Therefore, how to carry out effective construction organization design and preparation in large-scale, less-manned projects, so as to reasonably allocate mechanical resources within a given construction period, maximize resource utilization, and thus reduce the cost of the project, has become an urgent problem to be solved.

[0003] The problem of construction organization design preparation can usually be solved by modeling it as a resource-constrained project scheduling problem. This problem aims to achieve certain optimization goals, such as shortest construction period and minimum cost, by reasonably arranging the start time of activities under the conditions of activity timing constraints and resource constraints. The specific description is as follows: In a project, there are activities. Due to technical and organizational requirements, there is a precedence relationship between some activities, that is, the activity cannot be started before any of its predecessor activities are completed. Activity 1 represents the beginning of the project, and activity represents the end of the project, both of which are virtual activities. It is known that the execution time of the activity is, the demand for the th resource is, and the occupation of the th resource by the activity at any time does not exceed its upper limit. The problem to be solved is to schedule the start time of each activity under the premise of satisfying the above constraints. The goal is to shorten the total construction period of the project or minimize the total cost of the project. In RCPSP, only one execution mode is considered for any activity, that is, the time and resources required to complete an activity are fixed. However, in practical problems, an activity may have multiple execution modes. For the same activity, when different execution modes are adopted, the required resources and completion time are also different. This type of RCPSP with multiple execution modes is called a multi-mode resource-constrained project scheduling problem. MRCPSP is more general than RCPSP, so it is more practical to model the construction organization design compilation problem of large-scale sparsely-manned projects into MRCPSP for solution.

[0004] The most representative method for solving MRCPSP is the branch-and-bound method, which is an exact algorithm. The branch-and-bound method is the most commonly used algorithm for solving integer programming problems. It uses the breadth-first search strategy to search and iterate, selects different branch variables and sub-problems for branching, and calculates a target lower bound for the solution set in each subset to determine the optimal solution. However, when the scale of the problem increases, the search space of the branch-and-bound method will grow exponentially, and the complexity of its decision tree will also increase rapidly. Therefore, it is difficult for the branch-and-bound method to solve large-scale problems with limited solution resources and solution time. In order to solve larger-scale problems, researchers have developed approximate algorithms such as ant colony algorithms and particle swarm algorithms. Although these approximate algorithms can find approximately optimal solutions for medium-scale problems with dozens of activities, they are prone to fall into local optimality in large-scale problems with hundreds or even thousands of activities and cannot find acceptable and effective solutions. Due to the lack of effective algorithms for solving large-scale MRCPSP, currently, large-scale projects with few people usually use the traditional critical path method to prepare construction organization plans based on manual experience, and direct on-site construction operations through manual scheduling. This method usually consumes a lot of manpower and time costs to complete the preparation of the plan. Summary of the invention

[0005] In response to the above problems, the present invention provides an intelligent optimization method for the preparation of rapid construction organization design for large-scale projects. It aims at the generalized large-scale project problem that may have multiple types of processes and the execution time of activities in the same process is discontinuous, which may generate idle costs. It aims to minimize costs and solve the problem of the lack of effective construction organization design preparation methods in current large-scale projects with few people.

[0006] The technical solution of the present invention is:

[0007] An intelligent optimization method for rapid construction organization design of large-scale projects, specifically comprising the following steps:

[0008] S1. Read in basic project information: read in process logic relationship, activity indicator system, environmental logistics constraints and reduction relationship table, and total resource constraints;

[0009] S2, random initialization: Randomly initialize the population. For any individual, randomly initialize its previous gene position to any random integer that satisfies the shift constraint, and initialize the J+1th gene position to 1;

[0010] S3, Adaptive staffing strategy selection: Adaptively assign appropriate staffing strategies to individuals according to different shift arrangements and use them as the start time;

[0011] S4, fitness value calculation: For individuals whose construction period does not meet the constraints, calculate the work cost under the current shift arrangement, calculate the time that exceeds the construction period constraint and multiply it by the penalty coefficient ε as the penalty value, and take the sum of the work cost and the penalty value as the fitness value of the individual; for individuals whose construction period meets the constraints, the cost target calculated according to the optimal cost objective function is used as the fitness value;

[0012] S5. Optimize invalid individuals: For individuals that do not meet the construction period constraints, shorten the critical path through local search;

[0013] S6, individual selection: Each time, two individuals are randomly selected from the population with equal probability, the fitness of the two individuals is compared, and the individual with the lower fitness value is selected to enter the mating pool;

[0014] S7, single-point crossover: The individuals in the mating pool are randomly paired, and the two paired individuals have a probability C r Choose whether to crossover. For the two parent individuals to be crossed, randomly select any position on the chromosome as the crossover position, exchange the gene information after the crossover position in the chromosome of the parent individual, generate two new offspring individuals and add them to the next generation population. For the individuals that do not crossover, directly add them to the next generation population;

[0015] S8, random mutation: Each individual in the population has C m The probability of mutation is , and the first J gene positions of the individual to be mutated are also C m The probability of selecting a mutation point is , and the mutation point will randomly select a new shift number within the optional shift number of the gene position;

[0016] S9, output end project schedule: repeat S3-S8 until the maximum number of iterations is reached, then select the best individual from the last generation population for decoding, compile the project according to the specified compilation strategy and shift arrangement, and output it as a construction schedule;

[0017] The compilation strategies in S3 include sequential compilation strategies, overall compilation strategies and serial compilation strategies;

[0018] The sequential scheduling strategy is to calculate the earliest start time of the activity according to the immediate predecessor constraint and the shift arrangement, and use it as the start time. This strategy does not consider minimizing idle costs and resource constraints;

[0019] The overall compilation strategy is to regard the activities in each process as a whole, arrange all activities in each process to be executed continuously, so that the idle cost is 0, and arrange the earliest start time of all activities in the process that meet the immediate predecessor constraint as the start time for the first activity in the process;

[0020] The serial scheduling strategy is to calculate the earliest start time and the latest start time of all activities according to the predecessor constraints and the shift schedule, and consider the resource constraints to arrange the appropriate start time of the activities within the range of the earliest start time to the latest start time;

[0021] The adaptive compilation strategy selection specifically includes the following steps:

[0022] S3.1. Calculate the shortest construction period using the critical path method based on the shift arrangement;

[0023] S3.2. For individuals whose shortest duration does not meet the duration constraint, use the sequential compilation strategy to set the position of the chromosome's gene to 1, and do not execute the following steps; otherwise, use the reverse compilation strategy to set the position of the chromosome's gene to 2, and continue to execute the following steps;

[0024] S3.3. Use the reverse compilation strategy to compile the schedule according to the shift arrangement. If the compiled schedule does not meet the construction period constraint, the J+1th gene bit of the chromosome is reset to 1, and the following steps are no longer executed; otherwise, the compiled schedule is checked for resource constraints;

[0025] S3.4. If the inverse compilation strategy does not meet the resource constraints, the serial compilation strategy is used to reset the chromosome's gene position to 3; otherwise, no operation is performed;

[0026] The optimal objective function of cost in S4 is:

[0027]

[0028] Among them, n represents the number of process types, p i It is expressed as the number of activities included in the i-th process, c i It is expressed as the single-shift work cost of the i-th process, s i,j is the number of shifts arranged for the jth activity of the ith process, ρ is the ratio of idle cost to labor cost, d i,j represents the working time required to complete the jth activity of the ith process, start i,j It is represented by the start time and finish time of the jth activity of the ith process. i,j It is represented as the end time of the jth activity of the ith process, r i,j,k represents the quantity of the kth resource used for the jth activity of the ith process;

[0029] The local search for shortening the critical path in S5 specifically includes the following steps:

[0030] S5.1. Calculate the critical path using the critical path method based on the shift schedule;

[0031] S5.2. Traverse all activities on the critical path except virtual activities. For the ath activity on the critical path, its duration under the current shift arrangement is d1. Calculate the free float time of all the immediate predecessor activities on the non-critical path of the a+1th activity on the critical path, and record the minimum value as Δd min If the number of shifts for the ath activity does not reach the maximum limit, calculate the duration after adding one shift and add d2. If d1-d2>Δd min , then adding a shift to the ath activity will not make other immediate predecessor activities of the a+1th activity become activities on the critical path, and the ath activity will be added to the candidate activity set;

[0032] S5.3. For all activities in the candidate activity set, calculate the additional cost after adding one shift, select the activity with the smallest additional cost, and add one shift for it.

[0033] As a preferred solution of the present invention: the random initialization in S2 includes a coding method, and the coding information in the coding method only includes the shift arrangement and the compilation strategy. The specific process is as follows:

[0034] S2.1. First, number the different J activities according to the order of the logical relationship of the process, and number the different compilation strategies;

[0035] S2.2. The chromosome is represented by a one-dimensional array of length J, Δ = (θ1, θ2, ... θ J ,θ J+1 ), where the first J genes of the chromosome represent the number of shifts used for the work, and the J+1th gene position represents the compilation strategy to be used for the entire project;

[0036] S2.3. The project is organized according to the organization strategy specified by the J+1th gene position of the chromosome and the work shift arrangement contained in the first J gene positions.

[0037] The beneficial effects of the present invention are:

[0038] 1. The present invention proposes an efficient encoding and decoding method for the multi-mode resource-constrained project scheduling problem. By decoupling the two problems of solving mode selection and solving activity start time in the multi-mode resource-constrained task scheduling problem, the search space is effectively narrowed, the operation efficiency of the algorithm is greatly improved, and appropriate compilation strategies can be added according to actual needs to achieve further expansion of the algorithm;

[0039] 2. The present invention proposes an adaptive selection compilation strategy, and the algorithm will adaptively assign different compilation strategies to individuals. First, invalid solutions are quickly identified through the calculation of the critical path and the simplest compilation strategy is assigned to them, which reduces the computational consumption on the evaluation of invalid individuals, thereby speeding up the operation efficiency of the algorithm. Then, feasible compilation strategies are assigned to feasible candidate solutions to effectively improve the quality of the solution.

[0040] 3. The present invention proposes a local search technology for shortening the critical path. For individuals that do not meet the construction period constraints, the critical path is shortened through local search, thereby effectively compressing the construction period. Different from the previous method of simply searching for the work with the lowest rush cost on the critical path, this method aims to shorten the critical path without changing the critical path as much as possible. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is the algorithm flow chart of the present invention;

[0042] Figure 2 It is a Gantt comparison chart between the intelligent optimization algorithm of the present invention and the traditional algorithm;

[0043] Figure 3 It is the daily fund usage graph and the maximum usage graph of each resource of the intelligent optimization algorithm of the present invention;

[0044] Figure 4 It is the daily fund usage graph and the maximum usage graph of each resource of the traditional algorithm;

[0045] Figure 5 It is the algorithm planning progress diagram of the present invention. DETAILED DESCRIPTION

[0046] The embodiments of the present invention will be further described below in conjunction with the accompanying drawings.

[0047] Example 1:.

[0048] like Figure 1 As shown, the bridge construction project in railway construction is modeled as an example: there are n types of processes in this project, and the i-th type of process includes p i activities, with questions covering Activities in the same process with the same set of personnel are called a single shift. Due to the constraints of the construction environment, the maximum number of available shifts for the i-th process is limited to s i,max , different execution modes are equivalent to using different numbers of shifts, that is, the activities in the i-th process have s i,max In this execution mode, discontinuous execution of activities in the same process will result in vacant shifts. If a process shift is vacant, the vacancy cost is calculated as ρ times the labor cost. The constraints that any feasible construction schedule needs to meet are as follows:

[0049] 1. Order constraint: immediate precedence constraint;

[0050] 2. Resource constraints: The usage of each type of work and machinery at the same time cannot exceed the maximum supply resources;

[0051] 3. Duration constraint: The project has an earliest completion time T min and the latest completion time T max , the project duration must be within [T min ,T max ] to be completed;

[0052] 4. Work shift constraint: When multiple work shifts work together to complete a job, there will be a certain loss in construction rate. And due to the constraints of the construction environment, the number of work shifts assigned to each job cannot be greater than the maximum available number of work shifts for this type of work.

[0053] Specifically, an intelligent optimization method for rapid construction organization design of a large-scale project includes the following steps when used:

[0054] 1. Read in basic project information: read in process logic relationship, activity indicator system, environmental logistics constraints and reduction relationship table, and total resource constraints;

[0055] 2. Random initialization: The present invention proposes an efficient encoding and decoding method, which is as follows: the encoding information only includes the shift arrangement and the compilation strategy. First, the different J activities are numbered according to the order of the process logic relationship, and then the different compilation strategies are numbered. Therefore, the chromosome is represented by a one-dimensional array with a length of J+1. Δ=(θ1,θ2,...θ J ,θ J+1 ), where the first J gene bits in the chromosome represent the number of shifts used for the work, and the J+1th gene bit represents the organization strategy to be used for the entire project. During decoding, the project is organized according to the organization strategy specified by the J+1th gene bit of the chromosome and the shift arrangement contained in the first J gene bits. The population is randomly initialized. For any individual, its first J gene bits are randomly initialized to any random integer that satisfies the shift constraint, and the J+1th gene bit is initialized to 1;

[0056] 3. Adaptive staffing strategy selection: The algorithm adaptively assigns appropriate staffing strategies to individuals according to different shift arrangements to speed up the algorithm's operating efficiency and improve the algorithm's running time. The algorithm currently provides three staffing strategies, as follows:

[0057] (1) Sequential compilation strategy: Calculate the earliest start time of the activity according to the predecessor constraint and the shift schedule, and use it as the start time. This strategy does not consider minimizing idle costs and resource constraints, so the compilation speed is the fastest, but the quality of the compiled schedule is the worst;

[0058] (2) Overall compilation strategy: treat the activities in each process as a whole, arrange all activities in each process to be executed continuously, so that the idle cost is 0, and arrange the earliest start time for all activities in the process to meet the immediate predecessor constraint as the start time for the first activity in the process. It has the lowest cost under the same shift arrangement. When the number of activities in all processes is 1, it is equivalent to the sequential compilation strategy, and its compilation speed is slightly slower than the sequential compilation strategy;

[0059] (3) Serial scheduling strategy: Calculate the earliest start time and the latest start time of all activities according to the predecessor constraints and the shift schedule, and consider resource constraints to schedule the appropriate start time of the activities within the range of the earliest start time to the latest start time. Under the same shift schedule, it has the shortest completion time and meets the resource constraints;

[0060] Except for the overall compilation strategy, the other two strategies are existing compilation strategies.

[0061] The specific steps for selecting an adaptive compilation strategy are as follows:

[0062] (1) Calculate the shortest construction period using the critical path method based on the shift arrangement;

[0063] (2) For individuals whose shortest duration does not meet the duration constraint, use the sequential compilation strategy to set the J+1th gene position on the chromosome to 1, and do not execute the following steps; otherwise, use the reverse compilation strategy to set the J+1th gene position on the chromosome to 2, and continue to execute the following steps;

[0064] (3) According to the shift arrangement, the schedule is compiled using the reverse compilation strategy. If the compiled schedule does not meet the construction period constraint, the J+1th gene bit of the chromosome is reset to 1, and the following steps are no longer executed; otherwise, the resource constraint check is performed on the compiled schedule;

[0065] (4) If the inverse compilation strategy does not meet the resource constraints, the serial compilation strategy is used to reset the J+1th gene position of the chromosome to 3; otherwise, no operation is performed;

[0066] 4. Calculation of fitness value: For individuals whose construction period does not meet the constraints, calculate the work cost under the current shift arrangement, calculate the time that exceeds the construction period constraint and multiply it by a very large penalty coefficient ε as the penalty value, and take the sum of the work cost and the penalty value as the fitness value of the individual; for individuals whose construction period meets the constraints, calculate according to the formula The calculated cost target is used as the fitness value;

[0067] 5. Local search to shorten the critical path: For individuals that do not meet the construction period constraints, the critical path is shortened through local search, which includes the following steps:

[0068] (1) Calculate the critical path using the critical path method based on the shift schedule;

[0069] (2) Traverse all activities on the critical path except virtual activities. For the ath activity on the critical path, its duration under the current shift schedule is d1. Calculate the free float time of all the immediate predecessor activities on the non-critical path of the a+1th activity on the critical path, and record the minimum value as Δd min If the ath activity can be increased by one shift, then calculate the duration of one shift and add d2. If d1-d2>Δd min , then adding a shift to the ath activity will not make other immediate predecessor activities of the a+1th activity become activities on the critical path, and the ath activity will be added to the candidate activity set;

[0070] (3) For all activities in the candidate activity set, calculate the additional cost after adding one shift, select the activity with the smallest additional cost, and add one shift to it;

[0071] 6. Binary tournament individual selection: Each time, two individuals are randomly selected from the population with equal probability, the fitness of the two individuals is compared, and the individual with the lower fitness value is selected to enter the mating pool;

[0072] 7. Single-point crossover: Randomly pair the individuals in the mating pool, and the two paired individuals have a probability C r Choose whether to crossover. For the two parent individuals to be crossed, randomly select any position on the chromosome as the crossover position, exchange the gene information after the crossover position in the chromosome of the parent individual, generate two new offspring individuals and add them to the next generation population. For the individuals that do not crossover, directly add them to the next generation population;

[0073] 8. Random mutation: Each individual in the population has a mutation probability C m Perform the mutation operation, and use C in the first J gene positions of the individual to be mutated m The probability of selecting a mutation point is , and the mutation point will randomly select a new shift number within the optional shift number of the gene position;

[0074] 9. Population iteration: Repeat steps 3-8 until the maximum number of iterations is reached;

[0075] 10. Output the construction schedule: Select the best individual from the last generation of the population for decoding, compile the project according to the specified compilation strategy and shift arrangement, and output it as a construction schedule.

[0076] like Figure 2-Figure 5 As shown, in order to better evaluate the algorithm performance of the intelligent optimization algorithm proposed in the present invention, a comparative analysis experiment is conducted between the traditional algorithm structure and the algorithm results of the present invention:

[0077] The algorithm sets the population size to 20, the number of evolutionary iterations to 500, the number of activities to 217, the minimum duration to 650 days, and the maximum duration to 750 days. By analyzing the project costs, vacancy costs, and algorithm running time of different methods, the optimal solution obtained by the intelligent optimization algorithm proposed in the present invention is significantly less than the comparison algorithm in terms of completion costs, as shown in the following table:

[0078]

[0079]

[0080] As shown in the above table, compared with the traditional method, the method proposed by the present invention reduces the completion cost by about 400,000 and has no idle cost; in terms of algorithm operation efficiency, the running time of the intelligent optimization algorithm proposed by the present invention is also less than 1 second; in terms of construction period, the construction period obtained by the algorithm proposed by the present invention within the specified construction period range is shorter than that of the traditional algorithm.

[0081] Further, such as Figure 2 As shown, it can be clearly observed that the intelligent optimization algorithm proposed by the present invention can arrange the construction period more reasonably while satisfying the resource constraints. The intelligent optimization algorithm proposed by the present invention can obtain multiple effective solutions, and the cost of multiple groups of effective solutions is lower than that of the comparison algorithm while satisfying the construction period. The result given by this algorithm is the optimal solution among the effective solutions. It can be seen that the process is not so compactly executed within the construction period, which can reduce the number of shifts and reduce costs.

[0082] Further, such as Figure 3 and Figure 4 As shown by Figure 3 From the daily fund usage chart, we can see that in order to reduce idle costs as much as possible, funds are mainly spent in the subsequent period of the project. There will not be too much fund expenditure in the early stage, and the fund expenditure grows relatively steadily. From the maximum usage of each resource, we can observe that it does not exceed the resource limit, which is consistent with the Figure 4 By comparison, it can be found that the trends of the daily fund usage graph and the maximum resource usage graph are basically consistent, but the daily fund graph is more stable than the traditional algorithm, and the average daily expenditure and maximum resource usage are lower than the traditional algorithm.

[0083] Further, such as Figure 5 As shown, the intelligent optimization algorithm proposed in the present invention can quickly find a plan within a reasonable construction period, and then gradually find a better solution in iteration through local search. We tested the impact of different population sizes on the algorithm. Under different population sizes, the algorithm can find relatively good solutions. As the population size increases, the algorithm's local search capability will also increase accordingly.

[0084] The above-mentioned embodiments only express the specific implementation of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. An intelligent optimization method for rapid construction organization design of large-scale projects, characterized in that: The specific steps include: S1. Read in basic project information: read in process logic relationship, activity indicator system, environmental logistics constraints and reduction relationship table, and total resource constraints; S2, random initialization: Randomly initialize the population. For any individual, randomly initialize its previous gene position to any random integer that satisfies the shift constraint, and initialize the J+1th gene position to 1; S3, Adaptive staffing strategy selection: Adaptively assign appropriate staffing strategies to individuals according to different shift arrangements and use them as the start time; S4, fitness value calculation: For individuals whose construction period does not meet the constraints, calculate the work cost under the current shift arrangement, calculate the time that exceeds the construction period constraint and multiply it by the penalty coefficient ε as the penalty value, and take the sum of the work cost and the penalty value as the fitness value of the individual; for individuals whose construction period meets the constraints, the cost target calculated according to the optimal cost objective function is used as the fitness value; S5. Optimize invalid individuals: For individuals that do not meet the construction period constraints, shorten the critical path through local search; S6, individual selection: Each time, two individuals are randomly selected from the population with equal probability, the fitness of the two individuals is compared, and the individual with the lower fitness value is selected to enter the mating pool; S7, single-point crossover: The individuals in the mating pool are randomly paired, and the two paired individuals have a probability C r Choose whether to crossover. For the two parent individuals to be crossed, randomly select any position on the chromosome as the crossover position, exchange the gene information after the crossover position in the chromosome of the parent individual, generate two new offspring individuals and add them to the next generation population. For the individuals that do not crossover, directly add them to the next generation population; S8, random mutation: Each individual in the population has C m The probability of mutation is , and the first J gene positions of the individual to be mutated are also C m The probability of selecting a mutation point is , and the mutation point will randomly select a new shift number within the optional shift number of the gene position; S9, output end project schedule: repeat S3-S8 until the maximum number of iterations is reached, then select the best individual from the last generation population for decoding, compile the project according to the specified compilation strategy and shift arrangement, and output it as a construction schedule; The compilation strategies in S3 include sequential compilation strategies, overall compilation strategies and serial compilation strategies; The sequential scheduling strategy is to calculate the earliest start time of the activity according to the immediate predecessor constraint and the shift arrangement, and use it as the start time. This strategy does not consider minimizing idle costs and resource constraints; The overall compilation strategy is to regard the activities in each process as a whole, arrange all activities in each process to be executed continuously, so that the idle cost is 0, and arrange the earliest start time of all activities in the process that meet the immediate predecessor constraint as the start time for the first activity in the process; The serial scheduling strategy is to calculate the earliest start time and the latest start time of all activities according to the predecessor constraints and the shift schedule, and consider the resource constraints to arrange the appropriate start time of the activities within the range of the earliest start time to the latest start time; The adaptive compilation strategy selection specifically includes the following steps: S3.

1. Calculate the shortest construction period using the critical path method based on the shift arrangement; S3.

2. For individuals whose shortest duration does not meet the duration constraint, use the sequential compilation strategy to set the position of the chromosome's gene to 1, and do not execute the following steps; otherwise, use the reverse compilation strategy to set the position of the chromosome's gene to 2, and continue to execute the following steps; S3.

3. Use the reverse compilation strategy to compile the schedule according to the shift arrangement. If the compiled schedule does not meet the construction period constraint, reset the J+1th gene bit of the chromosome to 1 and do not execute the following steps; otherwise, perform a resource constraint check on the compiled schedule; S3.

4. If the inverse compilation strategy does not meet the resource constraints, the serial compilation strategy is used to reset the chromosome's gene position to 3; otherwise, no operation is performed; The optimal objective function of cost in S4 is: Among them, n represents the number of process types, p i It is expressed as the number of activities included in the i-th process, c i It is expressed as the single-shift work cost of the i-th process, s i,j is the number of shifts arranged for the jth activity of the ith process, ρ is the ratio of idle cost to labor cost, d i,j represents the working time required to complete the jth activity of the ith process, start i,j It is represented by the start time and finish time of the jth activity of the ith process. i,j It is represented as the end time of the jth activity of the ith process, r i,j,k represents the quantity of the kth resource used for the jth activity of the ith process; The local search for shortening the critical path in S5 specifically includes the following steps: S5.

1. Calculate the critical path using the critical path method based on the shift schedule; S5.

2. Traverse all activities on the critical path except virtual activities. For the ath activity on the critical path, its duration under the current shift arrangement is d1. Calculate the free float time of all the immediate predecessor activities on the non-critical path of the a+1th activity on the critical path, and record the minimum value as Δd min If the number of shifts for the ath activity does not reach the maximum limit, calculate the duration after adding one shift and add d2. If d1-d2>Δd min , then adding a shift to the ath activity will not make other immediate predecessor activities of the a+1th activity become activities on the critical path, and the ath activity will be added to the candidate activity set; S5.

3. For all activities in the candidate activity set, calculate the additional cost after adding one shift, select the activity with the smallest additional cost, and add one shift for it.

2. According to claim 1, the intelligent compilation and optimization method for rapid construction organization design of large-scale projects is characterized by: The random initialization in S2 includes a coding method, and the coding information in the coding method only includes the shift arrangement and the compilation strategy. The specific process is as follows: S2.

1. First, number the different J activities according to the order of the logical relationship of the process, and number the different compilation strategies; S2.

2. The chromosome is represented by a one-dimensional array of length J, Δ = (θ1, θ2, ... θ J ,θ J+1 ), where the first J gene positions of the chromosome represent the number of shifts used for the work, and the J+1th gene position represents the organization strategy to be used for the entire project; S2.

3. The project is organized according to the organization strategy specified by the J+1th gene position of the chromosome and the work shift arrangement contained in the first J gene positions.

Images