Coordinated scheduling method for intermediate transfer of prefabricated cabins with self-learning mechanism

Abstract

The present invention relates to a collaborative scheduling method for intermediate transfer of prefabricated cabins with a self-learning mechanism, and belongs to the technical field of large cruise ship manufacturing. The scheduling scheme described in the collaborative scheduling of the complex transfer link of the prefabricated cabins of cruise ships is usually unable to effectively implement the constraint judgment and infeasible solution processing process during the algorithm. The following steps are included: population initialization; fitness calculation; self-learning low-level scheduling rule update; genetic operation. The present invention realizes the control of the low-level scheduling rule components, and completes the solution of specific transfer tasks and equipment scheduling schemes according to the execution rule scheme mapping generated by the sequence of low-level scheduling rule components; the collaborative scheduling scheme for the intermediate transfer link of the prefabricated cabins of cruise ships can be solved without performing complex problem constraint judgment and invalid solution repair processes during the algorithm operation.

Description

Technical Field

[0001] The invention relates to a prefabricated cabin intermediate transfer coordinated scheduling method with a self-learning mechanism, and belongs to the technical field of large cruise ship manufacturing. Background Art

[0002] Prefabricated cabins are a typical type of intermediate product with a large number in the process of cruise ship construction. The intermediate transfer link involves the coordinated operation of multiple types of equipment, and the coordination of various types of equipment is a complex scheduling process. Reasonable formulation of scheduling plans for various types of equipment in the intermediate transfer link is a necessary condition for improving the utilization rate of enterprise production resources, reducing waste, and reducing production costs under the premise of ensuring that the original production resource configuration remains unchanged. However, since the intermediate transfer link of cruise ship prefabricated cabins is an operation process in which a series of transportation equipment influence and restrict each other, it has the characteristics of multiple types of equipment participation, multiple types of task execution, and multi-link process connection. It is difficult to clearly formulate an effective scheduling plan in the actual production process of the enterprise. Therefore, studying the collaborative scheduling method for the intermediate transfer link of cruise ship prefabricated cabins is of great significance to improving the production efficiency of enterprises and reducing production costs.

[0003] Prior art document 1 "Scheduling management and optimization analysis of intermediate products transfer in a shipyard for cruise ships", Liu et al., PLOS ONE, 17 (3), e0265047, which gives "A cruise ship intermediate product transportation scheduling optimization method based on genetic algorithm", which realizes the description of the scheduling plan of transportation tasks and equipment through double-layer chromosome encoding. Document 2 "Vehicle Routing and Scheduling Optimization of Ship Steel Distribution Center under Green Shipbuilding Mode", Li et al., Sustainability, 2019, 11, 4248, which gives "A shipbuilding intermediate product distribution and transportation scheduling method based on intelligent water drop algorithm", completes the description of the scheduling plan by inserting the central node of the solution sequence. However, for the coordinated scheduling of the transfer link of cruise ship prefabricated cabins with multi-type equipment participation, multi-type task execution, and multi-link process connection constraints, directly describing the scheduling plan usually cannot effectively realize the constraint judgment and infeasible solution processing process in the algorithm, so it is necessary to further study the scheduling method that can avoid the need for process constraint judgment and infeasible solution repair. The publication number is CN114841581A, and the name of the invention is GEP-VNS evolutionary dynamic job shop scheduling rules, and its feature selection method does not disclose the self-learning mechanism. The publication number is CN116795054A, and the name of the invention is a scheduling method for intermediate products in a discrete manufacturing mode, which is the scheduling solution of the solution result.

[0004] Therefore, it is urgent to propose a collaborative scheduling method for intermediate transfer of prefabricated cabins with a self-learning mechanism to solve the above technical problems. Summary of the invention

[0005] The present invention aims to overcome the defects of the prior art and provide a method for coordinated scheduling of intermediate transfer of prefabricated cabins with a self-learning mechanism. The method is based on a hyper-heuristic algorithm framework and uses genetic operations as high-level strategies to achieve control of low-level scheduling rule components, and completes the solution of specific transfer tasks and equipment scheduling solutions according to the execution rule solution mapping generated by the sequence of low-level scheduling rule components. The present invention can achieve the solution of the coordinated scheduling solution for the intermediate transfer link of prefabricated cabins of cruise ships without performing complex problem constraint judgment and invalid solution repair processes during the algorithm operation process. A brief overview of the present invention is given below to provide a basic understanding of certain aspects of the present invention. It should be understood that this overview is not an exhaustive overview of the present invention. It is not intended to determine the key or important parts of the present invention, nor is it intended to limit the scope of the present invention.

[0006] The technical solution of the present invention:

[0007] The coordinated scheduling method for intermediate transfer of prefabricated cabins with a self-learning mechanism includes the following steps:

[0008] Step 1: Using the population size of the advanced genetic strategy in the hyper-heuristic algorithm, according to the number of work tasks and the number of transfer stages, the initial scheduling plan for the coordinated operation of the intermediate transfer of the prefabricated cabin is generated through a random strategy;

[0009] Step 2: Fitness calculation, used to evaluate each initial scheduling solution;

[0010] Step 3: Self-learning low-level scheduling rule updates to continuously optimize the combination of scheduling rule components;

[0011] Step 4: Genetic operation is used to complete the collaborative scheduling solution.

[0012] Preferred: In step 1, the initial population pop generated by random strategy init :

[0013] pop init =rand(ps,(n·m)) (1)

[0014] Where ps is the population size of the advanced genetic strategy in the hyper-heuristic algorithm, n is the number of collaborative tasks, and m is the number of transfer stages. rand(ps, (n·m)) represents the random generation of a random number matrix with ps rows and (n·m) columns that satisfies the (0,1) uniform distribution. Each row vector of the random number matrix represents a chromosome of the genetic algorithm, and the chromosome corresponds to a scheduling rule solution to the job problem.

[0015] Step 2: Fitness calculation, including the following steps:

[0016] Step 2.1: Mapping probability of each low-level scheduling rule in the initial state;

[0017] Step 2.2: Map the probability value range of each low-level scheduling rule in order;

[0018] Step 2.3: According to the probability value of the gene position on the chromosome, determine the low-level scheduling rule sr corresponding to each gene position, complete the mapping of the low-level scheduling rule components corresponding to all positions on a chromosome, and obtain the complete scheduling rule to obtain the maximum fuzzy completion time AI consistency index with plan;

[0019] Step 2.4: Combine the two optimization objectives into one by weighting, construct the optimization objective function; calculate the individual fitness value;

[0020] Step 3: Evaluate the effectiveness and applicability of these low-level scheduling rules by calculating the number of occurrences of the low-level scheduling rules on the chromosome, the number of occurrences immediately after the low-level scheduling rules, and the set of low-level scheduling rules on all chromosomes;

[0021] Step 4: Use roulette method to select population.

[0022] Preferably: In step 2.1, the set of low-level scheduling rules is represented as L, and the number of low-level scheduling rules is represented as n llh , then the mapping probability p of each low-level scheduling rule in the initial state llh :

[0023]

[0024] Preferably: In step 2.2, the mapping probability value range of each low-level scheduling rule in order is

[0025]

[0026] Preferred: In step 2.3, according to the mapping probability of the low-level scheduling rule, the low-level scheduling rule sr corresponding to each gene position is determined, and the low-level scheduling rule component mapping corresponding to all positions on a chromosome is completed to obtain the complete scheduling rule, and then the allocation of the intermediate transfer task of the cruise ship prefabricated cabin and the scheduling of the equipment are completed according to the complete scheduling rule, and then the maximum fuzzy completion time is obtained. AI consistency index with plan;

[0027]

[0028]

[0029] Among them, f1 and f2 represent the optimization objectives of the coordinated scheduling problem of intermediate transfer of prefabricated cabins on cruise ships. represents the fuzzy completion time of each stage of the prefabricated cabin transfer task, max represents the maximum fuzzy completion time of each transfer stage, AI i It indicates the degree of consistency between the completion time of each stage of the prefabricated cabin transfer task and the plan. represents the planned fuzzy completion time of the prefabricated cabin transfer task, area represents the area enclosed by the fuzzy time and the coordinate axis, n represents the number of transfer stages required to perform the transfer task, i represents the specific task of the transfer stage, i=1,2,…,n.

[0030] Preferably: in step 2.4, the two optimization objectives are combined into one by weighting to construct an optimization objective function F;

[0031] F=ω1*f1+ω2*1 / f2 (6)

[0032] Among them, ω1 and ω2 represent the weights corresponding to the two optimization objectives respectively, and ω1+ω2=1 is used to represent the tendency of the scheduling process for the two optimization objectives;

[0033] In step 2.5, the individual fitness value can be calculated as:

[0034] fit=1 / F (7)

[0035] Preferably: In step 3, define ρ h,q is the number of occurrences of the low-level scheduling rule component h at position q on the chromosome, η h',h is the number of times the low-level scheduling rule component h appears immediately after h', Ω q is the set of low-level scheduling rules for all chromosomes in the population at position q, r is the specific low-level scheduling rule, r∈Ω q , the subscripts his and cur are the historical population and the current population in the algorithm iteration process, respectively, and the algorithm cumulative knowledge probability can be obtained and algorithms to explore knowledge probability

[0036]

[0037]

[0038] The current iteration number of the algorithm is represented as Iter, and the maximum iteration number is represented as maxIter. According to equations (8) and (9), the probability of updating the self-learning low-level scheduling rule μ can be calculated: h,q :

[0039]

[0040] Preferably: in step 4, after the relevant parameters for genetic operation are obtained according to formulas (1)-(7), a roulette wheel method is used to select the population, and half of the chromosomes in the population are obtained to enter the offspring population. The roulette wheel selection probability P(s) of individual s in the population can be calculated as:

[0041]

[0042] Among them, fit s is the fitness value corresponding to individual s in the population, fit i is the fitness value corresponding to individual i in the population; the other half of the offspring population updates the scheduling rules of each gene position of the parent chromosome according to equations (8)-(10), and the two parts together constitute a new offspring population; then the genetic operation is completed through uniform crossover and single-point mutation strategies, and then the coordinated scheduling solution for the intermediate transfer link of the cruise ship prefabricated cabin is solved through continuous iteration.

[0043] Preferably: the low-level scheduling rules adopt first-come-first-served, shortest time priority, longest time priority, shortest remaining time priority, longest remaining time priority, least total time proportion priority, and longest total time proportion priority, a total of seven low-level scheduling rule components constitute a low-level scheduling rule component set, and the method of the present invention is used to solve the coordinated scheduling scheme for the intermediate transfer link of the cruise ship prefabricated cabin.

[0044] The present invention has the following beneficial effects:

[0045] 1. The method of the present invention introduces a self-learning mechanism, which guides the convergence direction by learning the historical knowledge and exploration knowledge generated in the solution process, thereby effectively reducing the search space of the scheduling rules.

[0046] 2. The solution obtained by the method of the present invention is a dispatch rule solution, which is more universal and more suitable for dynamic environmental changes.

[0047] 3. The present invention is based on a hyper-heuristic algorithm framework, using genetic operations as high-level strategies to achieve control over low-level scheduling rule components, and completes the solution of specific transfer tasks and equipment scheduling solutions based on the execution rule solution mapping generated by the low-level scheduling rule component sequence; the present invention can achieve the solution of the coordinated scheduling solution for the intermediate transfer link of the cruise ship prefabricated cabin without performing complex problem constraint judgment and invalid solution repair process during the algorithm operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 The flowchart of the collaborative scheduling method for intermediate transfer of prefabricated cabins with a self-learning mechanism.

[0049] Figure 2 Composition diagram for low-level scheduling rule components.

[0050] Figure 3 Box plot of the results of 30 solutions of the test case for the collaborative scheduling method for intermediate transfer of prefabricated cabins with a self-learning mechanism; the A-axis represents the algorithm, and the B-axis represents the objective function value, dimensionless units.

[0051] Figure 4 A coordinated scheduling scheme for intermediate transfer of prefabricated cabins of cruise ships is a coordinated scheduling method for intermediate transfer of prefabricated cabins with a self-learning mechanism; the A-axis represents time in min, and the B-axis represents the equipment number. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention is described below by the specific embodiments shown in the accompanying drawings. However, it should be understood that these descriptions are only exemplary and are not intended to limit the scope of the present invention. In addition, in the following description, the description of well-known structures and technologies is omitted to avoid unnecessary confusion of the concept of the present invention.

[0053] Specific implementation method 1: Combination Figure 1-4 This embodiment is described. The prefabricated cabin intermediate transfer coordinated scheduling method with a self-learning mechanism in this embodiment adopts a generation hyper-heuristic algorithm framework, uses a low-level heuristic rule (low-level scheduling rule) sequence as the chromosome encoding of the high-level genetic algorithm framework to construct a complete scheduling solution, and self-learns the rule association knowledge of each genetic iteration process, adopts a probability update formula until the relevant low-level scheduling rules are generated, and includes the following steps:

[0054] Step 1: Using the population number of advanced genetic strategies in the hyper-heuristic algorithm (the number of scheduling rule schemes for all prefabricated cabin intermediate transfer collaborative operations), the population has diversity, and then according to the number of operating tasks required to be performed in each transfer stage and the number of transfer stages required to complete the transfer task, the initial scheduling scheme (population) for the prefabricated cabin intermediate transfer collaborative operation is generated through a random strategy to improve the optimization ability; that is, the number of operating tasks refers to the number of tasks required to be performed in each transfer stage, and the transfer stage refers to the total number of stages required to complete the transfer task;

[0055] In step 1, the initial population pop generated by random strategy init (A collection of initial scheduling rule schemes):

[0056] pop init =rand(ps,(n·m)) (1)

[0057] Where ps is the population size of the high-level genetic strategy in the hyper-heuristic algorithm, n is the number of collaborative tasks for intermediate transfer of cruise ship prefabricated cabins, and m is the number of transfer stages; rand(ps, (n·m)) represents the random generation of a random number matrix with ps rows and (n·m) columns that satisfies the (0,1) uniform distribution. Each row vector of the random number matrix represents a chromosome of the genetic algorithm, and the chromosome corresponds to a scheduling rule solution to the collaborative operation problem of intermediate transfer of cruise ship prefabricated cabins (one of the scheduling rule solutions for the collaborative operation of intermediate transfer of all prefabricated cabins); the random number on each chromosome represents the mapping probability of the corresponding low-level heuristic rule;

[0058] Step 2: Fitness calculation, used to evaluate the pros and cons of each initial scheduling solution, such as Figure 2 The low-level scheduling rules adopt first-come-first-served, shortest time priority, longest time priority, shortest remaining time priority, longest remaining time priority, least total time ratio priority, and longest total time ratio priority, a total of seven low-level scheduling rule components constitute a low-level scheduling rule component set, and the method of the present invention is used to solve the coordinated scheduling scheme of the intermediate transfer link of the cruise ship prefabricated cabin;

[0059] Step 2.1: The mapping probability of each low-level scheduling rule (low-level heuristic rule) in the initial state (the initial state refers to the starting state of the algorithm iteration number);

[0060] In step 2.1, the set of low-level scheduling rules is represented as L, and the number of low-level scheduling rules is represented as n llh , then the mapping probability p of each low-level scheduling rule in the initial state llh :

[0061]

[0062] Step 2.2: Map the probability value range of each low-level scheduling rule in order;

[0063] In step 2.2, the mapping probability range of each low-level scheduling rule in order is

[0064]

[0065] Step 2.3: A chromosome represents a scheduling rule solution to the problem, and the genes on the chromosome represent low-level heuristic scheduling rules. The low-level heuristic scheduling rules can be further decoded to obtain a specific scheduling plan; according to the probability value of the gene position on the chromosome, the low-level scheduling rule sr corresponding to each gene position is determined, and the low-level scheduling rule component mapping corresponding to all positions (genes) on a chromosome is completed to obtain a complete scheduling rule, and then the allocation of intermediate transfer tasks of the cruise ship prefabricated cabin and the scheduling of equipment are completed according to the low-level scheduling rules, and the maximum fuzzy completion time is obtained. AI consistency index with plan;

[0066] In step 2.3, according to the mapping probability of the low-level scheduling rule, the low-level scheduling rule sr corresponding to each gene position is determined, and the complete scheduling rule is obtained after the mapping of the low-level scheduling rule components corresponding to all positions on a chromosome is completed. Then, according to the complete scheduling rule, the allocation of intermediate transfer tasks of the cruise ship prefabricated cabin and the scheduling of equipment are completed, and the maximum fuzzy completion time is obtained. AI consistency index with plan;

[0067]

[0068]

[0069] Among them, f1 and f2 represent the optimization objectives of the coordinated scheduling problem of intermediate transfer of prefabricated cabins on cruise ships. represents the fuzzy completion time of each stage of the prefabricated cabin transfer task, max represents the maximum fuzzy completion time of each transfer stage, AI i It indicates the degree of consistency between the completion time of each stage of the prefabricated cabin transfer task and the plan. represents the planned fuzzy completion time of the prefabricated cabin transfer task, area represents the area enclosed by the fuzzy time and the coordinate axis, n represents the number of transfer stages required to be performed in the transfer task, i represents the specific operation task of the transfer process, i=1,2,…,n, n is a positive integer;

[0070] Step 2.4: Combine the two optimization objectives into one by weighting and construct the optimization objective function;

[0071] In step 2.4, the two optimization objectives are combined into one by weighting to construct the optimization objective function F;

[0072] F=ω1*f1+ω2*1 / f2 (6)

[0073] Among them, ω1 and ω2 represent the weights corresponding to the two optimization objectives respectively, and ω1+ω2=1 is used to represent the tendency of the scheduling process for the two optimization objectives;

[0074] Step 2.5: Calculate the individual fitness value; the individual represents a chromosome, that is, a scheduling rule solution to the problem;

[0075] In step 2.5, the individual fitness value can be calculated as:

[0076] fit=1 / F (7)

[0077] Step 3: Self-learning low-level scheduling rule updates to continuously optimize the combination of scheduling rule components;

[0078] Step 3: Evaluate the effectiveness and applicability of these low-level scheduling rules by calculating the number of occurrences of the low-level scheduling rules on the chromosome, the number of occurrences immediately after the low-level scheduling rules, and the set of low-level scheduling rules on all chromosomes; through the iterative process of the algorithm, continuously optimize the combination of scheduling rules, improve the algorithm's search ability and efficiency, while maintaining a certain exploration ability to avoid falling into the local optimal solution;

[0079] In step 3, define ρ h,q is the number of occurrences of the low-level scheduling rule component h at position q on the chromosome, η h',h is the number of times the low-level scheduling rule component h appears immediately after h', Ω q is the set of low-level scheduling rules for all chromosomes in the population at position q, r is the specific low-level scheduling rule, r∈Ω q , the subscripts his and cur are the historical population and the current population in the algorithm iteration process, respectively, and the algorithm cumulative knowledge probability can be obtained and algorithms to explore knowledge probability

[0080]

[0081]

[0082] The current iteration number of the algorithm is represented as Iter, and the maximum iteration number is represented as maxIter. According to equations (8) and (9), the probability of updating the self-learning low-level scheduling rule μ can be calculated: h,q :

[0083]

[0084] Step 4: Genetic operation is used to complete the collaborative scheduling solution;

[0085] The relevant parameters for genetic operations are calculated by equations (1)-(7); then, the roulette method is used to select the population, that is, the probability of each individual being selected and inherited to the next generation population is calculated according to the fitness calculation function value of the individual, and whether each individual is inherited to the next generation population is determined by simulating the roulette operation; half of the offspring population is obtained by the roulette method, and the other half of the offspring population is updated according to equations (8)-(10) to update the scheduling rules of each gene position of the parent chromosome; by simulating the process of natural selection and genetic transmission, continuous iterative optimization is carried out, and finally the optimal solution of the coordinated scheduling scheme for the intermediate transfer link of the cruise ship prefabricated cabin is found.

[0086] In step 4, after obtaining the relevant parameters for genetic operations according to equations (1)-(7), the roulette wheel method is used to select the population, and half of the chromosomes in the population enter the offspring population. The roulette wheel selection probability P(s) of individual s in the population can be calculated as:

[0087]

[0088] Among them, fit s is the fitness value corresponding to individual s in the population, fit i is the fitness value corresponding to individual i in the population; the other half of the offspring population updates the scheduling rules of each gene position of the parent chromosome according to equations (8)-(10), and the two parts together constitute a new offspring population; then the genetic operation is completed through uniform crossover and single-point mutation strategies, and then the coordinated scheduling solution of the intermediate transfer link of the cruise ship prefabricated cabin is solved through continuous iteration;

[0089] Determine whether the number of algorithm iterations Iter = maxIter is true. If so, calculate the scheduling plan and output the result. Otherwise, return to step 2, the fitness calculation process, and continue iterative calculation until Iter = maxIter is true.

[0090] Embodiment 1:

[0091] This embodiment adopts seven low-level heuristic rule components, namely, first-come-first-served, shortest time priority, longest time priority, shortest remaining time priority, longest remaining time priority, least total time proportion priority, and longest total time proportion priority, to form a low-level heuristic rule component set, and adopts the method of the present invention to solve the coordinated scheduling scheme for the intermediate transfer link of the cruise ship prefabricated cabin. Figure 1 Detailed description of the implementation process:

[0092] Step 1: Randomly generate the initial population by combining the parameters n, m of the coordinated scheduling of intermediate transfer of cruise prefabricated cabins and the algorithm parameters ps;

[0093] Step 2: Fitness calculation, including the following steps:

[0094] Step 2.1: Mapping probability of each low-level scheduling rule in the initial state;

[0095] Step 2.2: Map the probability value range of each low-level scheduling rule in order;

[0096] Step 2.3: According to the probability value of the gene position on the chromosome, determine the low-level scheduling rule sr corresponding to each gene position, complete the mapping of the low-level scheduling rule components corresponding to all positions on a chromosome, and then obtain the complete scheduling rule. Then, according to the low-level scheduling rule, complete the allocation of intermediate transfer tasks of the cruise ship prefabricated cabin and the scheduling of equipment, and then obtain the maximum fuzzy completion time AI consistency index with plan;

[0097] Step 2.4: Combine the two optimization objectives into one by weighting, construct the optimization objective function; calculate the individual fitness value;

[0098] Step 3: Calculate the self-learning update probability information on each gene position of the chromosome according to formula (8)-(10), select half of the individuals in the population to use the self-learning update mechanism for update calculation; at the same time, take ω1=ω2=0.5, combine formula (2)-(7) to complete the population fitness calculation, the initial population adapts to the problem, and selects half of the chromosomes to enter the offspring population according to the roulette probability according to formula (11);

[0099] Step 4: Perform crossover and mutation operations on the offspring population;

[0100] Further, determine whether the number of algorithm iterations Iter = maxIter is established. If so, perform scheduling calculations and output the results. Otherwise, return to the fitness calculation process and continue iterative calculations until Iter = maxIter is established.

[0101] To illustrate the beneficial effects of the method of the present invention, a comparative test is conducted between the embodiment of the method of the present invention and a commonly used scheduling algorithm; Figure 3 The figure shows the box plot obtained by calculating 30 times the method of the present invention in the test example. It can be seen that the median value of the objective function value obtained by the coordinated scheduling method of the intermediate transfer link of the cruise prefabricated cabin with a self-learning mechanism of the present invention under the current optimization weight setting is 2.7535, which is much smaller than the commonly used scheduling method (often greater than 5), indicating that the method of the present invention has good stability and optimization accuracy under the test example;

[0102] Attached Figure 4The figure shows the coordinated scheduling scheme for the intermediate transfer link of the cruise ship prefabricated cabin calculated by the method of the present invention. The maximum fuzzy completion time is [247, 287, 328] (unit: min), the plan consistency index is 0.488, and the optimization target value is 2.336, which is much smaller than the commonly used scheduling method (often greater than 4.5). This shows that the method has good stability and optimization accuracy in the calculation of actual scheduling schemes.

[0103] In summary, the collaborative scheduling method for the intermediate transfer link of cruise ship prefabricated cabins with a self-learning mechanism in the present invention has good stability and optimization accuracy in the calculation of test examples and actual scheduling plans, and avoids the constraint processing process of complex business processes. It is suitable for the transfer scheduling link of cruise ship prefabricated cabins, and is of great significance for improving the utilization rate of production resources of cruise ship manufacturing enterprises, improving transportation efficiency, and thus reducing enterprise cost consumption.

[0104] It should be noted that in the above embodiments, as long as the technical solutions are not contradictory, they can be arranged and combined, and those skilled in the art can exhaust all possibilities based on the mathematical knowledge of arrangement and combination. Therefore, the present invention will no longer describe the technical solutions after arrangement and combination one by one, but it should be understood that the technical solutions after arrangement and combination have been disclosed by the present invention.

[0105] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A collaborative scheduling method for intermediate transfer of prefabricated cabins with a self-learning mechanism, characterized by: The following steps are involved: Step 1: Using the population size of the advanced genetic strategy in the hyper-heuristic algorithm, according to the number of work tasks and the number of transfer stages, the initial scheduling plan for the coordinated operation of the intermediate transfer of the prefabricated cabin is generated through a random strategy; In step 1, the initial population generated by random strategy : (1) in, is the population size of the advanced genetic strategy in the hyper-heuristic algorithm, is the number of collaborative tasks, is the number of transfer stages; Indicates that random generation satisfies Evenly distributed OK, A random number matrix with columns, each row vector of which represents a chromosome of the genetic algorithm, and the chromosome corresponds to a scheduling rule solution of the job problem; Step 2: Fitness calculation, used to evaluate each initial scheduling scheme; Step 2: Fitness calculation, including the following steps: Step 2.1: Mapping probability of each low-level scheduling rule in the initial state; In step 2.1, the low-level scheduling rule set is represented as , the number of low-level scheduling rules is expressed as , then the mapping probability of each low-level scheduling rule in the initial state is : (2) Step 2.2: Map the probability value range of each low-level scheduling rule in order; In step 2.2, the mapping probability value range of each low-level scheduling rule in order is : (3) Step 2.3: Determine the low-level scheduling rules corresponding to each gene position based on the probability value of the gene position on the chromosome After completing the mapping of low-level scheduling rule components corresponding to all positions on a chromosome, the complete scheduling rule is obtained, and the maximum fuzzy completion time is obtained. Consistency with plan index ; In step 2.3, the low-level scheduling rules corresponding to each gene position are determined according to the mapping probability of the low-level scheduling rules. After completing the mapping of the low-level scheduling rule components corresponding to all positions on a chromosome, the complete scheduling rule is obtained, and then the allocation of intermediate transfer tasks of the cruise ship prefabricated cabin and the scheduling of equipment are completed according to the complete scheduling rule, and then the maximum fuzzy completion time is obtained. Consistency with plan index ; (4) = (5) in, represents the optimization objective of the coordinated scheduling problem of intermediate transfer of prefabricated cabins on cruise ships, Indicates the fuzzy completion time of each transfer stage of the prefabricated cabin transfer task, represents the maximum fuzzy completion time of each transfer stage operation, It indicates the degree of consistency between the completion time of each stage of the prefabricated cabin transfer task and the plan. represents the planned fuzzy completion time of the prefabricated cabin transfer task, represents the area enclosed by the fuzzy time and the coordinate axis, Indicates the specific number of operations in the transfer phase that need to be performed in the transfer task. Indicates the specific tasks in the transfer phase. ; Step 2.4: Combine the two optimization objectives into one by weighting and construct the optimization objective function; Step 2.5: Calculate individual fitness value; Step 3: Self-learning low-level scheduling rule updates to continuously optimize the combination of scheduling rule components; Step 3: Evaluate the effectiveness and applicability of these low-level scheduling rules by calculating the number of occurrences of the low-level scheduling rules on the chromosome, the number of occurrences immediately after the low-level scheduling rules, and the set of low-level scheduling rules on all chromosomes; In step three, define For low- and medium-level scheduling rule components Location on chromosome The number of occurrences, For low-level scheduling rules components exist The number of times it appears immediately after, For all chromosomes in the population at position The set of low-level scheduling rules that appear on is a specific low-level scheduling rule, , subscript and are the historical population and the current population in the algorithm iteration process, respectively, and the algorithm cumulative knowledge probability is obtained and algorithms to explore knowledge probability : (8) (9) The current iteration number of the algorithm is expressed as , the maximum number of iterations is expressed as , according to formula (8) and formula (9), the probability of updating the self-learning low-level scheduling rules is calculated : (10) Step 4: Genetic operation is used to complete the collaborative scheduling solution; Step 4: Use roulette wheel method to select population; In step 4, after obtaining the relevant parameters for genetic operations according to equations (1)-(7), a roulette wheel method is used to select the population, and half of the chromosomes in the population are entered into the offspring population. The probability of choosing a roulette wheel Calculated as: (11) in, For individuals in a population The corresponding fitness value; the other half of the offspring population updates the scheduling rules of each gene position of the parent chromosome according to equations (8)-(10), and the two parts together constitute a new offspring population; then the genetic operation is completed through uniform crossover and single-point mutation strategies, and then the coordinated scheduling solution for the intermediate transfer link of the cruise ship prefabricated cabin is solved through continuous iteration.

2. The coordinated scheduling method for intermediate transfer of prefabricated cabins with a self-learning mechanism according to claim 1 is characterized in that: In step 2.4, the two optimization objectives are combined into one by weighting to construct the optimization objective function ; (6) in, and Respectively represent the weights corresponding to the two optimization objectives, , which is used to indicate the tendency of the scheduling process towards the two optimization objectives.

3. The coordinated scheduling method for intermediate transfer of prefabricated cabins with a self-learning mechanism according to claim 2 is characterized in that: In step 2.5, the individual fitness value is calculated as: (7)。 4. The coordinated scheduling method for intermediate transfer of prefabricated cabins with a self-learning mechanism according to claim 1 is characterized in that: The low-level scheduling rules adopt seven low-level scheduling rule components, including first-come-first-served, shortest time priority, longest time priority, shortest remaining time priority, longest remaining time priority, least total time proportion priority, and longest total time proportion priority, to form a low-level scheduling rule component set. This method is used to solve the collaborative scheduling scheme for the intermediate transfer link of prefabricated cabins of cruise ships.

Images