Quantum-inspired method for scheduling of precast construction components

Abstract

The application provides a quantum heuristic-based building component prefabrication task scheduling method, and relates to the technical field of task cooperative scheduling. The method abstracts building prefabrication tasks into a quantum optimization model, encodes a task allocation scheme with a quantum bit probability amplitude; through Monte Carlo observation, a quantum superposition state is collapsed into a classical feasible solution, and a weighted function is used for multi-objective quantification evaluation. A controlled phase gate is introduced to realize directional crossover between chromosomes, a simulated annealing temperature decay strategy is fused to dynamically adjust the mutation strength, an adaptive quantum rotation gate is used to adjust the probability amplitude distribution in real time according to the fitness difference, and the population quality is continuously improved. The resource load and process dependence are monitored, and when conflicts or overloads occur, intelligent repair and elastic scheduling are performed; the iteration termination adopts three criteria of algebraic threshold, fitness improvement rate and population diversity, and after convergence, the final generation individuals are comprehensively weighed, and the global optimal scheduling scheme is output.

Description

Technical Field

[0001] The present invention relates to the technical field of task collaborative scheduling, and in particular to a method for scheduling building component prefabrication tasks based on quantum inspiration. Background Art

[0002] Against the backdrop of the accelerating industrialization of construction engineering, precast concrete components have become a mainstream construction method due to their advantages in standardization, controllability, and construction efficiency. However, as project scale expands, component types increase, and production lines become more parallelized, traditional scheduling methods are increasingly facing insurmountable bottlenecks. On the one hand, classic priority rules, heuristic scheduling, and integer programming models are prone to local optima when faced with the "high-variety, low-batch, high-speed" precast factories. They are unable to find a feasible solution that balances schedule, cost, and energy consumption within the vast combination space. On the other hand, frequent on-site disturbances (such as equipment failures, raw material delays, and urgent orders) often prevent the existing static scheduling schemes from adapting in real time, often leading to idle resources or conflicts, delivery delays, and even chain reactions of production stoppages.

[0003] In recent years, while some research has attempted to introduce intelligent optimization frameworks such as genetic algorithms, ant colony algorithms, or reinforcement learning, these approaches, limited by the classical "0 / 1" bit encoding scheme, have led to an exponential expansion of the solution space and a sharp increase in computational time with the scale of the task. Furthermore, these algorithms lack unified and robust mechanisms for multi-objective trade-offs, dynamic constraint repair, and convergence criterion design, resulting in insufficient stability in the actual workshop environment. More critically, task scheduling in prefabrication scenarios is characterized by highly coupled process dependencies, multi-dimensional resource constraints, and real-time requirements. While existing quantum-inspired scheduling methods for cloud computing or mobile edge computing have theoretically demonstrated the parallel exploration advantages brought by quantum superposition, entanglement, and measurement, they have yet to be deeply adapted to the discrete manufacturing characteristics of prefabrication plants. Significant gaps remain in how quantum encoding maps component processes, how quantum observations rapidly generate executable workshop instructions, and how quantum evolution can be integrated with real-time perturbation closed-loop feedback at the production site. Therefore, the industry urgently needs a new task collaborative scheduling method that can fully utilize the parallel advantages of quantum computing and deeply integrate the process constraints and dynamic resource characteristics of prefabrication factories, so as to break through the limitations of traditional and existing intelligent algorithms in global optimization capabilities, dynamic response speed and multi-objective balance, thereby significantly improving the production efficiency and delivery reliability of building component prefabrication factories in complex and uncertain environments.

[0004] In view of this, this application is filed. Summary of the Invention

[0005] The present invention provides a quantum-inspired method for scheduling building component prefabrication tasks, which can at least partially improve the above-mentioned problems.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] A quantum-inspired method for scheduling building component prefabrication tasks, comprising:

[0008] Obtain the current pending task scheduling plan, encode the task scheduling plan using quantum bit encoding, generate a single chromosome, and initialize multiple single chromosomes to generate an initial population;

[0009] Convert the chromosomes in the initial population into a classical scheduling scheme and calculate the fitness of the classical scheduling scheme;

[0010] A controlled phase gate is used to perform gene exchange processing on the quantum bits between multiple chromosomes. The temperature attenuation strategy is combined to dynamically adjust the mutation intensity. Based on the fitness of the classical scheduling scheme, the probability amplitude of the quantum bits of the chromosomes is dynamically adjusted through quantum gate operations.

[0011] Monitor the resource occupancy and task dependency of the offspring chromosomes obtained after gene exchange in real time, and optimize the corresponding scheme of the offspring chromosomes based on the monitoring results;

[0012] The optimized solution is judged and screened according to the preset conditions and preset comprehensive indicators to obtain the current optimal solution.

[0013] In summary, this method first uniformly encodes tasks, processes, and resource constraints into quantum bit probability amplitudes, allowing a single chromosome to implicitly represent a vast number of feasible solutions. This superposition is then collapsed into a classical scheduling solution through quantum observation, and its fitness is rapidly evaluated using weighted functions of duration, cost, and energy consumption. Subsequently, controlled phase gates are used to achieve directed crossover between chromosomes. A simulated annealing temperature decay strategy is combined to dynamically adjust the mutation intensity. Quantum gates are adaptively rotated based on fitness differences to continuously optimize the probability amplitude distribution, balancing global exploration with local refinement. During system operation, resource load and task dependencies are monitored at high frequency. When timing conflicts or resource overruns are detected, intelligent remediation is triggered based on task criticality: non-critical tasks are right-shifted, while critical tasks utilize reserved elastic resources. The iterative process is terminated using a triple criterion: an algebraic threshold, fitness improvement rate, and population diversity. Finally, the final generation of elite solutions is weighted and ranked using the analytic hierarchy process to output the optimal scheduling solution in terms of duration, cost, and energy consumption. It aims to solve the pain points of traditional scheduling methods in discrete manufacturing scenarios, such as being prone to falling into local optimality, poor dynamic adaptability, and difficulty in balancing multiple objectives. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 4 is a flow chart of a method for scheduling building component prefabrication tasks based on quantum inspiration provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0015] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0016] refer to Figure 1 As shown, the first embodiment of the present invention discloses a quantum-inspired method for scheduling building component prefabrication tasks, which can be executed by a quantum-inspired building component prefabrication task scheduling device (hereinafter referred to as the scheduling device), and in particular, by one or more processors within the scheduling device to implement the following method:

[0017] S1, obtain the current task scheduling plan to be processed, encode the task scheduling plan using quantum bit encoding, generate a single chromosome, and initialize multiple single chromosomes to generate an initial population;

[0018] Specifically, step S1 further includes: obtaining a task scheduling plan to be processed, encoding the optional steps of each task in the task scheduling plan using a quantum bit encoding method, generating multiple quantum bits corresponding to the tasks, and obtaining a single chromosome, wherein the single chromosome is composed of multiple quantum bits;

[0019] The quantum bit is expressed as a probability amplitude, and the formula for the quantum bit of the jth process of the i-th task is: , 、 are all ground states, 、 are all complex probability amplitudes, which satisfy the normalization condition , is the probability of not selecting the process, that is, it collapses to The probability of is the probability of selecting this process, which collapses to probability;

[0020] Combining the greedy algorithm with the probability amplitude, multiple single chromosomes are randomly initialized to generate the initial population.

[0021] In this embodiment, the ERP or MES interface first pulls the entire list of tasks to be produced that day in real time. Each record contains the component number, the set of optional processes, the type and quantity of required resources, the earliest start time, and the latest end time. Then, step S1 is entered: The entire production blueprint is compressed into a chromosome for "microfilm", which significantly improves the expression ability and search efficiency of the solution space. Specifically, for any task i, its j-th candidate process is mapped to a quantum state, where Directly represents the probability that the process is actually selected, for example | =0.7 means that the probability of Task 2 selecting Process 3 is 70%. This encoding method makes a chromosome composed of the quantum bits of all tasks sequentially spliced ​​together; for example, in the scenario of 5 tasks × 3 processes, a single chromosome can implicitly carry 2 15 It saves a lot of storage space compared with traditional 0-1 encoding and far exceeds the linear expression capability of traditional binary encoding.

[0022] Among them, the quantum bit can be understood as a DNA in a chromosome. The overall task scheduling solution is composed of multiple quantum bits, which can be understood as a chromosome composed of several DNAs. Compared with traditional binary coding, this encoding method has stronger linear expression capabilities, reduces the overall coding volume, reduces search time, and improves search efficiency. The probability amplitude is the complex coefficient in the quantum state. , is the weight that describes the quantum bit when it is in superposition state. For example, the quantum state middle, and They represent the probability of getting 0 or 1 when measured. It is a form of quantum bit.

[0023] Furthermore, when initializing the chromosomes, 30% of the chromosomes are quickly generated by the greedy algorithm: first, they are sorted locally by “shortest processing time first”, and then each process is assigned a |β| close to 1. 2 , so that the initial population contains a high-quality "skeleton"; the remaining 70% of chromosomes are assigned amplitudes in a uniform random manner to ensure genetic diversity; that is, for all 、 ,set up , In a typical 20-task instance, this hybrid initialization strategy significantly improved the optimal fitness of the first generation compared to a purely random approach, while maintaining sufficient search breadth, laying the foundation for a population with both a "high starting point" and "multi-path" for subsequent quantum evolution.

[0024] S2, converts the chromosomes in the initial population into the classical scheduling scheme and calculates the fitness of the classical scheduling scheme;

[0025] Specifically, step S2 further includes: converting the chromosomes in the initial population into a classical scheduling scheme based on the quantum state collapse mechanism, and maintaining probability conservation through dynamic normalization processing, wherein, for the i-th task, calculating the cumulative probability distribution of its j-th process , generate a [0,1] uniformly distributed random number r, select The minimum process is the final process, is the complex probability amplitude;

[0026] The multi-objective weighted function is used to quantify the quality of the classic scheduling scheme and obtain the fitness of the current classic scheduling scheme. The formula is: .

[0027] In this embodiment, quantum measurement is performed on each chromosome in the population task by task. The cumulative probability distribution of the m candidate processes of task i is first calculated. Then, a hardware random number generator extracts r in the interval [0,1]. If , then the jth process is the final process. This process naturally introduces the randomness of quantum measurement, allowing the same chromosome to generate multiple classical solutions. For example, in a 20-task scenario, an average of 2 valid variants are generated per observation. To prevent the probability amplitude from drifting after multiple observations, the system performs dynamic normalization immediately after each measurement, that is, recalculating after each observation. , which not only maintains mathematical rigor but also ensures the stability of subsequent iterations.

[0028] Once a classic solution is generated, the system immediately applies a multi-objective weighted function to "score" it. This function takes normalized duration, cost, and energy consumption as input, with weights set to 0.5, 0.3, and 0.2, respectively. It also adds a constraint penalty: Each time a resource limit or timing conflict occurs, the total fitness is deducted by a certain value. This allows the "on-time delivery, budget overrun, and energy consumption overrun" to be simultaneously factored into a scalar between 0 and 1. This design, validated on a real-world precast plant dataset, can complete a fitness evaluation within 200 milliseconds, two orders of magnitude faster than traditional discrete event simulation. Furthermore, the real-time feedback provided by the penalty significantly improves search efficiency.

[0029] S3 uses a controlled phase gate to perform gene exchange processing on the quantum bits between multiple chromosomes, combines the temperature attenuation strategy to dynamically adjust the mutation intensity, and dynamically adjusts the probability amplitude of the quantum bits of the chromosomes through quantum gate operations according to the fitness of the classical scheduling scheme;

[0030] Specifically, step S3 further includes: using a controlled phase gate to perform gene exchange processing on the quantum bits between multiple chromosomes, and establishing a quantum entangled state dynamic adjustment crossover strategy, wherein, for any two parent chromosomes and Perform controlled phase gate operation to generate daughter chromosomes , is the cross phase angle, It is the base of the natural logarithm in mathematics, approximately equal to 2.71828. is a matrix representation of a controlled phase gate (CX gate) that links parental genes through quantum entanglement, To represent the tensor product, if , but , 、 are the fitness of the two parent chromosomes respectively, if ,but , the child generation enhances the inheritance of the parent generation If the gene ,but , the child generation enhances the inheritance of the parent generation genes;

[0031] Introducing adaptive crossover probability , when premature convergence conditions are detected, the cross strength is automatically increased, where the premature convergence conditions are , is the population diversity index, It is a hyperbolic tangent function that smoothly maps diversity to the probability adjustment interval to avoid parameter mutations.

[0032] By integrating simulated annealing and quantum mutation mechanisms, the temperature attenuation strategy is used to dynamically control the mutation intensity. ;

[0033] When in high temperature stage hour( is the number of iterations), perform large-scale mutation operations, including process exchange, resource reallocation, and mutation probability As the temperature decreases, it decays exponentially from the initial value, and its mathematical expression is: , Represents the fitness improvement, which is used to quantify the quality difference between the current solution and the historical optimal solution. , is the fitness value of the historical optimal solution in the population up to the current generation, is the fitness value of the individual solution currently being evaluated, if , indicating that the current solution is inferior to the historical optimal solution. , it is close to convergence.

[0034] For gene sites that have fallen into local stagnation during the evolution process, a directed mutation strategy is introduced to enforce neighborhood search for gene sites that have not improved for 10 generations, and to perform neighborhood operations on stagnant sites.

[0035] Adopting the adaptive rotation angle strategy, quantum gate operation is performed according to the fitness of the classical scheduling scheme to dynamically adjust the distribution of probability amplitude, wherein the fitness difference between the current solution and the historical optimal solution is , calculate the rotation angle, the formula is: , is the fitness of the current solution, is the fitness of the historical optimal solution, is the learning rate;

[0036] Perform a unitary transformation on each qubit, the formula is: ,in, 、 are the probability amplitudes of the converted process j of task i, is the matrix transpose operator symbol, is the adaptive rotation angle.

[0037] In this example, a controlled phase gate (CX) acts as a "molecular scissors" to perform gene exchange on parent chromosomes. Any two parent chromosomes undergo CX gates to produce entangled offspring. The crossover phase angle is determined by the fitness difference between the parents, enabling the directional transfer of high-quality genes. The crossover intensity is adjusted in real time by adaptive probability. Once the population diversity index falls below 0.3, the system immediately increases the crossover rate, rapidly opening new search channels and effectively suppressing premature convergence. Quantum entanglement can be understood as the strong correlation between two or more quantum systems.

[0038] Furthermore, the adaptive crossover profile is based on population diversity to avoid premature convergence and local optima. A decrease in the diversity index D enhances genetic recombination. For example, if most chromosomes converge to the same process sequence, the adaptive mechanism will increase the crossover frequency to introduce new combinations.

[0039] Immediately following crossover, the algorithm initiates temperature-decayed quantum mutation. The annealing temperature remains high until the 30th generation, at which point the system boldly executes drastic mutations such as process swaps and resource reallocations. The probability of mutations decays exponentially as the temperature decreases, maintaining exploration enthusiasm while gradually tightening convergence precision. For loci that have not improved for 10 consecutive generations, the algorithm applies a directed neighborhood search: the probability amplitude of that locus is perturbed within a ±5% range and renormalized. Simply put, the purpose of implementing a forced neighborhood search is to precisely intervene at loci that have reached local stagnation during evolution, breaking the deadlock through structured perturbations. Next, neighborhood operations, such as process swaps and resource reallocation, are applied to these stagnant loci (based on the neighborhood).

[0040] Finally, an adaptive revolving gate fine-tunes the probability amplitude. The system calculates the fitness difference between the current solution and the historical optimal solution, generates a rotation angle according to a formula, and updates the chromosome bit by bit using a unitary matrix. This rotation strategy improves average fitness over 50 generations in a scenario with 20 tasks and 3 targets, while maintaining a population diversity index above 0.4. This ensures that the final scheduling solution achieves the triple advantages of shortest duration, lowest cost, and lowest energy consumption.

[0041] S4, monitors the resource occupancy and task dependency of the offspring chromosomes obtained after the gene exchange process in real time, and optimizes the corresponding scheme of the offspring chromosomes based on the monitoring results;

[0042] Specifically, step S4 further includes: constructing a multi-dimensional dynamic monitoring system based on the resource load matrix Track the second-level utilization of class m resources in real time to obtain the resource occupancy rate. When the resource occupancy rate exceeds the preset threshold, an early warning is triggered and the task is started to shift right. The preset threshold is , is the utilization rate of the mth resource at time t (usually in percentage form);

[0043] When a task with a timing conflict is detected, Rearrange the processes of tasks with time sequence conflicts in descending order;

[0044] When a task with a resource conflict is detected, a dynamic priority mechanism is used to perform real-time scheduling optimization for the task with the resource conflict. Among them, automatic classification is performed based on the task criticality index CI, an intelligent right-shift strategy is implemented for non-critical tasks, and a preset-sized elastic resource pool is reserved for critical tasks.

[0045] In this embodiment, after gene exchange is complete and the offspring chromosomes have achieved their initial morphology, they are immediately projected into a digital twin space synchronized with the actual production line, entering the "real-time correction" phase in step S4. Based on the resource load matrix, the multi-dimensional dynamic monitoring system tracks the second-by-second utilization of m types of resources in real time at a sampling frequency of 10 Hz. Once resource utilization exceeds a threshold, a millisecond-level warning is issued and a "task right shift" micro-scheduling is automatically triggered.

[0046] For timing conflicts between processes (using topological sorting to identify cyclic dependencies, for example, when randomly generating the initial population or in subsequent iterations, two processes may occupy the same machine at the same time), press The order of conflicting tasks is reordered in descending order, with high-probability processes being prioritized. This not only preserves the preference for high-quality genes in the quantum genetics stage, but also ensures a logical loop. For resource conflicts (real-time calculation of resource load rates for each period, such as when resource allocation exceeds the available resource threshold at the same time), a task criticality index (CI) is introduced to grade all tasks in real time: Non-critical tasks are marked as "driftable" and intelligent right shift is directly implemented; The critical tasks are designated as "protected areas", for which a 15% elastic resource pool is reserved. In subsequent periods, the idle capacity of adjacent periods can be dynamically borrowed to ensure that the critical path is not lengthened.

[0047] Simply put, step S4 can be understood as a test of the offspring produced in step S3. During the hybridization process, in order to inherit high-quality genes, the aforementioned resource overrun and timing conflicts may occur. Real-time monitoring of resource usage and task dependencies, along with intelligent repair mechanisms (such as timing conflict rescheduling and dynamic priority adjustment), ensures the feasibility of scheduling solutions in dynamic environments.

[0048] S5, judge and screen the optimized solution according to the preset conditions and preset comprehensive indicators to obtain the current optimal solution.

[0049] Specifically, step S5 further includes: judging the optimized solution according to preset conditions, terminating the optimization and saving the current optimal solution when the solution meets any two of the preset conditions, wherein the preset conditions include three items: the first item is monitoring the number of iterations, judging whether the number of iterations exceeds the algebraic threshold, the second item is detecting the fitness improvement rate, judging whether the fitness improvement rate of the population is less than the change threshold, and the third item is detecting the diversity of the population, judging whether the population diversity index is greater than the minimum diversity threshold;

[0050] The optimized solution is judged according to the preset conditions. When the solution meets any two of the preset conditions, the optimization is terminated and the current optimal solution is saved. The preset conditions include three items: the first item is to monitor the number of iterations to determine whether the number of iterations exceeds the algebraic threshold; the second item is to detect the fitness improvement rate to determine whether the fitness improvement rate of the population is less than the change threshold; the third item is to detect the diversity of the population to determine whether the population diversity index is greater than the minimum diversity threshold.

[0051] The multiple chromosomes screened are re-sorted and screened according to the preset comprehensive indicators to obtain the current optimal solution, which is as follows:

[0052] According to the AHP method, the weights of construction period, cost and energy consumption are scored, and a judgment matrix is ​​constructed, and the maximum eigenvector is used to represent the weight of each factor;

[0053] Calculate the preset solution with the highest fitness of the last generation population and obtain the current optimal solution.

[0054] In this embodiment, when quantum evolution comes to an end, a composite criterion (such as algebraic threshold, fitness improvement rate, and population diversity) is used to control the convergence of the algorithm to ensure that a high-quality solution is output within a reasonable time. Specifically, the optimized solution is judged according to the preset conditions. As long as two indicators send signals at the same time, the generation stops immediately and the current optimization is locked in the "safe". Specifically, the preset conditions are three conditions for the algorithm to stop. One is the number of iterations (for example, the maximum number of iterations G max = 200 generations), one is the fitness change rate of several populations in several consecutive generations (for example, after the current solution has been iterated for K = 10 generations, the change rate of the population fitness is less than the threshold = 0.001), and the other is the diversity of the population (for example, the minimum diversity D min =10).

[0055] However, the final optimal solution has not yet been achieved. The AHP method is first used to determine the weights for the three factors of duration, energy consumption, and cost. Experts may adjust the weights based on practical considerations. The top-ranked individuals from the final generation are then selected to determine the optimal solution based on their weights. By comprehensively considering multiple key performance indicators (KPIs) such as duration, cost, and energy consumption, the quality of the scheduling solution is quantified using weighting functions and penalty terms to achieve a multi-objective balance.

[0056] Simply put, this method combines quantum computing principles (such as qubit superposition, quantum measurement, and quantum gate operations) with classical heuristic algorithms (such as genetic algorithms and simulated annealing) to solve the problem of coordinated task scheduling in building component prefabrication plants. By incorporating dynamic adjustment strategies such as quantum crossover, annealing mutation, and quantum rotating gates, it adaptively optimizes search directions based on population diversity and fitness differences, avoiding local optima.

[0057] In summary, the quantum-inspired method for scheduling building component prefabrication tasks effectively overcomes these limitations by integrating quantum computing principles with classical heuristic algorithms, significantly improving the optimization effect and practicality of the scheduling scheme. Specifically, first, to address the problem that traditional algorithms are prone to falling into local optimality, quantum-inspired algorithms utilize the superposition state characteristics of quantum bits to explore multiple potential solutions simultaneously. For example, in a quantum genetic algorithm, each quantum chromosome represents multiple possible task sequences through probability amplitude encoding, thereby maintaining a richer diversity of solution spaces during the search process. Combined with the dynamic adjustment mechanism of the quantum revolving gate, the algorithm can adaptively update the search direction based on the difference between the current solution and the global optimal solution, avoiding premature convergence to a suboptimal solution. This method demonstrates stronger global search capabilities in complex scheduling scenarios.

[0058] Secondly, the proposed constraint repair mechanism and dynamic resource allocation strategy effectively address the lack of adaptability of traditional methods in dynamic environments. The solution generated through quantum observation detects timing conflicts and resource overruns in real time and prioritizes the most optimal adjustment solution based on probability amplitude. For example, when a task cannot be started on time due to resource constraints, the algorithm quickly evaluates alternative solutions, delaying or reallocating tasks to meet the constraints. This dynamic responsiveness makes it particularly suitable for scenarios such as building prefabrication sites, where task priorities and resource availability frequently change.

[0059] Furthermore, quantum-inspired algorithms significantly improve computational efficiency through solution space compression and adaptive optimization strategies. Quantum bit encoding converts the solutions explicitly stored in traditional genetic algorithms into probability amplitudes, reducing memory usage and computational overhead. Furthermore, an adaptive mutation mechanism combined with simulated annealing balances exploration and exploitation, avoiding ineffective iterations and accelerating convergence. Multiple termination conditions (such as algebraic thresholds and a solution improvement stagnation check) further ensure that the algorithm produces high-quality solutions within a reasonable timeframe.

[0060] Finally, through multiple iterations of quantum genetic algorithm optimization, high-performing candidates are selected from the final generation population, and a multi-objective evaluation system is constructed to comprehensively evaluate these solutions. Specifically, the system comprehensively considers key performance indicators such as project duration, resource costs, and energy consumption, and uses a multi-criteria decision-making method to quantitatively analyze and compare various solutions, ultimately selecting the scheduling solution that achieves the best balance between multiple optimization objectives.

[0061] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A quantum-inspired method for scheduling building component prefabrication tasks, characterized in that: include: Obtain the current pending task scheduling plan, encode the task scheduling plan using quantum bit encoding, generate a single chromosome, and initialize multiple single chromosomes to generate an initial population; Convert the chromosomes in the initial population into a classical scheduling scheme and calculate the fitness of the classical scheduling scheme; A controlled phase gate is used to perform gene exchange processing on the quantum bits between multiple chromosomes. The temperature attenuation strategy is combined to dynamically adjust the mutation intensity. Based on the fitness of the classical scheduling scheme, the probability amplitude of the quantum bits of the chromosomes is dynamically adjusted through quantum gate operations. Monitor the resource occupancy and task dependency of the offspring chromosomes obtained after gene exchange in real time, and optimize the corresponding scheme of the offspring chromosomes based on the monitoring results; Judge and screen the optimized solutions based on the preset conditions and comprehensive indicators to obtain the current optimal solution; A controlled phase gate is used to perform gene exchange processing on quantum bits between multiple chromosomes, specifically: A controlled phase gate is used to perform gene exchange processing on quantum bits between multiple chromosomes, and a quantum entangled state dynamic adjustment crossover strategy is established, wherein for any two parent chromosomes and Perform controlled phase gate operation to generate daughter chromosomes , is the cross phase angle, is the base of natural logarithms in mathematics, is the tensor product, 、 are the fitness of the two parent chromosomes respectively; Introducing adaptive crossover probability , when premature convergence conditions are detected, the cross strength is automatically increased, where the premature convergence conditions are , is the population diversity index, is the hyperbolic tangent function.

2. The method for scheduling building component prefabrication tasks based on quantum inspiration according to claim 1 is characterized in that: Obtain the current pending task scheduling plan, encode the task scheduling plan using quantum bit encoding, generate a single chromosome, and initialize multiple single chromosomes to generate an initial population. Specifically: Obtaining a currently pending task scheduling plan, encoding the optional steps of each task in the task scheduling plan using a quantum bit encoding method, generating a plurality of quantum bits corresponding to the tasks, and obtaining a single chromosome, wherein the single chromosome is composed of a plurality of quantum bits; The quantum bit is expressed as a probability amplitude, and the formula for the quantum bit of the jth process of the i-th task is: , 、 are all ground states, 、 are all complex probability amplitudes, which satisfy the normalization condition , is the probability of not selecting the process, is the probability of selecting this process; Combining the greedy algorithm with the probability amplitude, multiple single chromosomes are randomly initialized to generate the initial population.

3. The method for scheduling building component prefabrication tasks based on quantum inspiration according to claim 1 is characterized in that: Convert the chromosomes in the initial population into a classical scheduling scheme and calculate the fitness of the classical scheduling scheme, specifically: Based on the quantum state collapse mechanism, the chromosomes in the initial population are converted into a classical scheduling scheme, and the probability conservation is maintained through dynamic normalization processing. For the i-th task, the cumulative probability distribution of its j-th process is calculated. , generate a [0,1] uniformly distributed random number r, select The minimum process is the final process, is the complex probability amplitude; The multi-objective weighted function is used to quantify the quality of the classic scheduling scheme and obtain the fitness of the current classic scheduling scheme. The formula is: .

4. The method for scheduling building component prefabrication tasks based on quantum inspiration according to claim 1 is characterized in that: Combined with the temperature attenuation strategy, the mutation intensity is dynamically adjusted, specifically: By integrating simulated annealing and quantum mutation mechanisms, the temperature attenuation strategy is used to dynamically control the mutation intensity. ; When in high temperature stage When performing large-scale mutation operations, including process exchange and resource reallocation, the mutation probability As the temperature decreases, it decays exponentially from the initial value, and its mathematical expression is: , is the fitness improvement amount, is the number of iterations; For gene sites that have fallen into local stagnation during the evolution process, a directed mutation strategy is introduced to enforce neighborhood search for gene sites that have not been improved for 10 generations, and to perform neighborhood operations on stagnant sites.

5. The method for scheduling building component prefabrication tasks based on quantum inspiration according to claim 4 is characterized in that: According to the fitness of the classical scheduling scheme, the probability amplitude of the chromosome's quantum bits is dynamically adjusted through quantum gate operations, specifically: Adopting the adaptive rotation angle strategy, quantum gate operation is performed according to the fitness of the classical scheduling scheme to dynamically adjust the distribution of probability amplitude, wherein the fitness difference between the current solution and the historical optimal solution is , calculate the rotation angle, the formula is: , is the fitness of the current solution, is the fitness of the historical optimal solution, is the learning rate; Perform a unitary transformation on each qubit, the formula is: ,in, 、 are the probability amplitudes of the converted process j of task i, is the matrix transpose operator symbol, is the adaptive rotation angle.

6. The method for scheduling building component prefabrication tasks based on quantum inspiration according to claim 5 is characterized in that: The resource occupancy rate and task dependency of the offspring chromosomes obtained after gene exchange are monitored in real time, and optimization is performed based on the monitoring results, specifically: Build a multi-dimensional dynamic monitoring system based on the resource load matrix Track the second-level utilization of the mth type of resources in real time to obtain the resource occupancy rate. When the resource occupancy rate exceeds the preset threshold, an early warning is triggered and the task is started to shift right. The preset threshold is , is the utilization rate of the mth type of resource at time t; When a task with a timing conflict is detected, Rearrange the processes of tasks with time sequence conflicts in descending order; When a task with a resource conflict is detected, a dynamic priority mechanism is used to perform real-time scheduling optimization for the task with the resource conflict. Among them, automatic classification is performed based on the task criticality index CI, an intelligent right-shift strategy is implemented for non-critical tasks, and a preset-sized elastic resource pool is reserved for critical tasks.

7. The method for scheduling building component prefabrication tasks based on quantum inspiration according to claim 6 is characterized in that: Mission Criticality Index , The tasks are non-critical tasks. The task is a critical task, and the preset size is 15%.

8. The method for scheduling building component prefabrication tasks based on quantum inspiration according to claim 1 is characterized in that: The optimized solutions are judged and screened according to the preset conditions and comprehensive indicators to obtain the current optimal solution, which is as follows: The optimized solution is judged according to the preset conditions. When the solution meets any two of the preset conditions, the optimization is terminated and the current optimal solution is saved. The preset conditions include three items: the first item is to monitor the number of iterations to determine whether the number of iterations exceeds the algebraic threshold; the second item is to detect the fitness improvement rate to determine whether the fitness improvement rate of the population is less than the change threshold; the third item is to detect the diversity of the population to determine whether the population diversity index is greater than the minimum diversity threshold; Based on the quantum genetic mechanism, multiple chromosomes with high fitness rankings are selected from the last generation of the population of the current optimal solution, and the selected chromosomes are re-sorted and screened according to preset comprehensive indicators to obtain the current optimal solution. The preset comprehensive indicators include construction period, cost, and energy consumption.

9. The method for scheduling building component prefabrication tasks based on quantum inspiration according to claim 8, characterized in that: The multiple chromosomes screened are re-sorted and screened according to the preset comprehensive indicators to obtain the current optimal solution, which is as follows: According to the AHP method, the weights of construction period, cost and energy consumption are scored, and a judgment matrix is ​​constructed, and the maximum eigenvector is used to represent the weight of each factor; Calculate the preset solution with the highest fitness of the last generation population and obtain the current optimal solution.

Images