AGV scheduling method considering charging facility resource limitation and flexible charging strategy

Abstract

The invention provides an AGV scheduling method considering charging facility resource limitation and a flexible charging strategy, and belongs to the technical field of industrial AGVs. According to the method, an industrial reality problem is converted into a mixed integer linear programming model, and an adaptive large neighborhood search algorithm is designed to solve the problem. The method comprises the following specific steps: firstly, generating an initial feasible solution through a three-stage heuristic method; and then introducing a destruction operator and a repair operator to iterate the initial solution through a local search mechanism, and finally further optimizing the solution by using neighborhood search to obtain an approximate optimal solution. The resource limitation of the charging facilities is considered, and the constraint conditions are constructed to realize flexible scheduling of AGV charging, so that an effective solution is provided for the AGV scheduling problem with limited charging facilities. The problem that quick response cannot be realized due to the lack of real-time performance and dynamic adjustment capability of the method in the prior art is solved; and the overall scheduling efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial AGVs, and particularly to an AGV scheduling method considering resource limitations of charging facilities and flexible charging strategies. Background Art

[0002] Industrial AGVs have the advantages of simple structure, high working efficiency, and strong controllability, and thus are widely applied. However, with the booming development of industrial AGVs, charging problems such as insufficient resources of AGV charging facilities, delays in transportation operations, and long charging waiting times of AGVs have become increasingly serious.

[0003] In the prior art, for the scheduling method of AGV charging, in order to simplify the problem, it is often assumed that the charging facilities can provide charging services for an infinite number of AGVs simultaneously. Although this assumption is helpful for solving the problem theoretically, it is far from the actual situation in practical applications; and a fixed charging strategy is adopted, that is, the battery replacement strategy or the AGV needs to be fully charged in each charging operation; the fact that AGV charging scheduling is prone to AGV idleness or delays in transportation operations is not considered; with the increase in the number of AGVs and the improvement of the complexity of transportation operations, the existing methods cannot find the optimal solution or approximate optimal solution within a reasonable time, especially in large-scale industrial applications. In addition, the scalability of the existing algorithms is also an important issue. With the expansion of the production scale and the change of scheduling requirements, the algorithms need to be able to adapt to the growing problem scale and more complex constraint conditions.

[0004] Therefore, an AGV scheduling method considering resource limitations of charging facilities and flexible charging strategies is needed. Summary of the Invention

[0005] In view of this, the present invention provides an AGV scheduling method considering resource limitations of charging facilities and flexible charging strategies, which solves the problems of inability to respond quickly and schedule flexibly in the existing AGV charging scheduling methods.

[0006] For this purpose, the present invention provides the following technical solutions:

[0007] An AGV scheduling method considering resource limitations of charging facilities and flexible charging strategies, comprising:

[0008] Taking the transportation operation assignment variable, the charging timing and duration variable, the task sequence of transportation operations on the same AGV, the AGV charging sequence on each charging facility, the AGV assignment variable at the charging facility, and the completion time of all transportation operations and charging operations as decision variables; taking the minimization of the maximum completion time of all transportation operations as the objective function; and taking the transportation operation scheduling constraint, the battery energy flow balance constraint, and the charging scheduling constraint as constraint conditions to construct a mixed integer linear programming model;

[0009] Solve the mixed-integer linear programming model through an adaptive large neighborhood search algorithm to obtain an AGV scheduling scheme;

[0010] The solution of the mixed-integer linear programming model through the adaptive large neighborhood search algorithm includes:

[0011] Generate an initial feasible solution through a three-stage heuristic method;

[0012] Expand the search scope of the initial feasible solution space by introducing a destruction operator and a repair operator;

[0013] Introduce a local search mechanism to update the optimal solution obtained through iterative search.

[0014] Furthermore, the transportation operation assignment variables, charging timing and duration variables, task sequence of transportation operations on the same AGV, AGV charging sequence on each charging facility, AGV assignment variables on the charging facility, and completion times of all transportation operations and charging operations include:

[0015] v jkr : If transportation operation j is executed after the r-th charging operation of AGV k , then v jkr is equal to 1, otherwise 0;

[0016] w kr : If AGV k executes the r-th charging operation, then w kr is equal to 1, otherwise 0;

[0017] δ kr : Charging duration of the r-th charging operation of AGV k ;

[0018] Δ kr : Remaining battery power of AGV k before executing the r-th charging operation;

[0019] y krs : If the r-th charging operation of AGV k is executed on charging facility s, then y krs is equal to 1, otherwise 0;

[0020] x krk′r′ : If the r-th charging operation of AGV k and the r'-th charging operation of AGV k′ are assigned to the same charging facility, and the r-th charging operation of AGV k is executed before the r'-th charging operation of AGV k′ , then x krk′r′ is equal to 1, otherwise 0;

[0021] C kr : The completion time of the r-th charging operation of the AGV k

[0022] Furthermore, the transportation operation scheduling constraints include:

[0023] Each transportation operation is assigned to the AGV k after the r-th charging operation:

[0024]

[0025] v jkr ≤ w kr

[0026] When the AGV k performs the r-th charging operation, the number of transportation operations j assigned to the AGV k is greater than or equal to 1:

[0027]

[0028] When the AGV k does not perform the (r - 1)-th charging operation, then the r-th charging operation is not performed:

[0029] w kr ≤ w k,r-1 .

[0030] Furthermore, the battery energy flow balance constraints include:

[0031] The remaining battery power of the AGV k before the r-th charging operation is equal to the remaining battery power of the AGV k before the (r - 1)-th charging operation plus the battery power of the (r - 1)-th charging operation minus the energy consumption of the transportation operations performed between the (r - 1)-th and r-th charging operations:

[0032]

[0033] Impose a battery capacity limit on each charging operation:

[0034] τδ kr + Δ kr ≤ b

[0035] The remaining battery power of the AGV k is non - negative:

[0036] Δ kr ≥ 0

[0037] When the AGV k ​If the r-th charging operation of the [AGV] is not executed, the charging duration of the r-th charging operation is 0:

[0038]

[0039] AGV k is initially in a charged state, and the initial battery level is equal to the battery capacity of the [AGV] k :

[0040] w k1 = 1

[0041] Δ k1 = b

[0042] AGV k 's r-th charging operation must be assigned to a charging facility s:

[0043]

[0044] where J: the set of transportation operations, represented by indices i, j; K: the set of [AGVs], represented by index k; R: the set of charging operations, represented by index r; S: the set of charging facilities, represented by index s.

[0045] Furthermore, the charging scheduling constraints include:

[0046] If the r-th charging operation of the [AGV] k and the r'-th charging operation of the [AGV] k′ are assigned to the same charging facility s, and r' is executed after r, then the completion time of r' is greater than or equal to the completion time of r plus the charging time of r':

[0047] C k′r′ ≥ C kr + δ k′r′ - M(3 - x krk′r′ - y kra - y k′r′s )

[0048] C kr ≥ C k′r′ + δ kr - M(2 + x krk′r′ - y krs - y k′r′s )

[0049] AGV k 's completion time of the r-th charging operation is greater than or equal to the sum of the completion time of the (r - 1)-th charging operation, the charging duration of the r-th charging operation, the charging preparation time, and the processing time of the transportation operations of the [AGV] after the (r - 1)-th charging operation: k :

[0050]

[0051] AGV k The maximum completion time is greater than or equal to the completion time of the r-th charging operation of the AGV k and the total processing time of the transportation operations after the r-th charging operation:

[0052] C max ≥ C kr + ∑ j∈J v jkr p j .

[0053] Furthermore, the generation of the initial feasible solution by the three-stage heuristic method includes:

[0054] The first stage: Using the branch and bound method, assign transportation operations to the AGVs;

[0055] The second stage: By the Arc-flow method, assign transportation operations and charging operations to each AGV;

[0056] The third stage: Schedule the charging operations on a preset number of charging facilities according to the first-come-first-charge scheduling rule.

[0057] Furthermore, the expansion of the search scope of the initial feasible solution space by introducing the destruction operator and the repair operator includes:

[0058] Use the roulette wheel selection mechanism to select the destruction operator or the repair operator;

[0059] The destruction operator destroys on the basis of the current solution to explore a new solution space;

[0060] After the destruction operation, the repair operator uses the repair operator to reconstruct the solution and restore the feasibility of the solution.

[0061] Furthermore, the destruction operator includes:

[0062] Random transportation operation removal, critical transportation operation removal, worst transportation operation removal, random charging operation removal, critical charging operation removal, worst charging operation removal;

[0063] The repair operator includes: random transportation operation insertion, greedy transportation operation insertion, regret value transportation operation insertion, random charging operation insertion, greedy charging operation insertion, and regret value charging operation insertion.

[0064] Furthermore, the local search mechanism includes a splitting operator and a moving operator;

[0065] The splitting operator identifies the AGV with the longest charging completion timek and the r-th charging operation of the AGV k has a waiting time; determine the charging facility s where the r-th charging operation of the AGV k is located, identify that the previous charging operation of the r-th charging operation of the AGV on the charging facility s is the r'-th charging operation of the AGV k , divide the r'-th charging operation of the AGV k′ into two charging operations r' k′ and r' 1 , and allocate the transportation operations assigned after the r'-th charging operation to the divided r'-th 2 charging operation and after the r'-th 1 charging operation; 2

[0066] The movement operator first identifies the AGV with the longest charging completion time k , and the r-th charging operation of the AGV k has a waiting time; determine the charging facility s where the r-th charging operation of the AGV k is located, identify that the previous charging operation of the r-th charging operation of the AGV on the charging facility s is the r'-th charging operation of the AGV k ; move the transportation operations assigned after the r-th k′ charging operation of the AGV k′ to after the r ′ +n-th charging operation of the AGV k′ . ′

[0067] Furthermore, the input parameters of the mixed-integer linear programming model include:

[0068] p j : the processing time of transportation operation j, e j : the energy consumption of transportation operation j, b: the battery capacity of each AGV, τ: the charging rate of each AGV, and λ: the setup time of each charging operation.

[0069] The present invention constructs a mixed-integer linear programming model and an adaptive large neighborhood search algorithm to consider the AGV scheduling problem with charging facility resource constraints and flexible charging strategies, so as to obtain a set of AGV scheduling schemes; for the adaptive large neighborhood search algorithm, an initial feasible solution is generated by a three-stage heuristic method; the initial solution is improved by introducing a destruction operator and a repair operator to effectively explore the solution space, and a local search mechanism is introduced to improve the optimal solution obtained through iterative search, thereby providing an effective solution for the AGV scheduling problem. Furthermore, it solves the lack of real-time performance and dynamic adjustment ability of the existing technical methods and cannot respond quickly; improves the overall scheduling efficiency. Brief Description of the Drawings

[0070] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0071] Figure 1 It is a flowchart of the method in the embodiment of the present invention. Detailed Embodiments

[0072] In order to enable those skilled in the art of the present technology to better understand the solutions of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0073] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such used data can be interchanged under appropriate circumstances so that the embodiments of the present invention described here can be implemented in an order other than those illustrated or described here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0074] The present invention provides an AGV scheduling method considering charging facility resource constraints and flexible charging strategies. By transforming industrial reality problems into a mixed integer linear programming model and designing an adaptive large neighborhood search (adaptive large neighborhood search algorithm) to solve this problem: first, generate an initial feasible solution through a three-stage heuristic method; then introduce a destruction operator and a repair operator to iterate the initial solution through a local search mechanism, and finally further optimize the solution using neighborhood search to obtain an approximate optimal solution. The present invention considers the resource constraints of charging facilities and constructs constraint conditions to achieve flexible scheduling of AGV charging, thereby providing an effective solution for the AGV scheduling problem with limited charging facilities. Furthermore, it solves the problem that the existing technical methods lack real-time performance and dynamic adjustment capabilities and cannot respond quickly; and improves the overall scheduling efficiency.

[0075] The specific method steps include:

[0076] S1. First, characterize the problem as a mixed-integer linear programming model.

[0077] S11. Define binary variables and continuous variables to represent the scheduling decisions of AGVs s .

[0078] In this embodiment, variables are defined to represent whether a certain transportation operation is executed by a specific AGV and whether the charging operation is performed at a specific charging facility, including:

[0079] 1. Define sets:

[0080] J: The set of transportation operations, represented by indices i, j;

[0081] K: The set of AGVs (Automated Guided Vehicles), represented by index k;

[0082] R: The set of charging operations, represented by index r;

[0083] S: The set of charging facilities, represented by index s;

[0084] 2. Define variables:

[0085] v jkr : If transportation operation j is scheduled to be executed after the r-th charging operation of AGV k , then v jkr equals 1, otherwise 0;

[0086] w kr : If AGV k executes the r-th charging operation, then w kr equals 1, otherwise 0;

[0087] δ kr : The charging duration of the r-th charging operation of AGV k ;

[0088] Δ kr : The remaining battery level of AGV k before executing the r-th charging operation;

[0089] y krs : If the r-th charging operation of AGV k is performed at charging facility s, then y krs equals 1, otherwise 0;

[0090] x krk′r′ : If the r-th charging operation of AGV k and the r-th charging operation of AGV k′ r′ The r-th charging operation is assigned to the same charging facility, and the AGV k performs its r-th charging operation before the r'-th charging operation of the AGV k′ , then x krk′r′ equals 1, otherwise 0;

[0091] C kr : The completion time of the r-th charging operation of the AGV k .

[0092] 3. Define parameters:

[0093] p j : The processing time of transportation operation j;

[0094] e j : The energy consumption of transportation operation j;

[0095] b: The battery capacity of each AGV;

[0096] τ: The charging rate of each AGV;

[0097] M: A sufficiently large number for modeling constraints;

[0098] λ: The setup time for each charging operation.

[0099] S12. Establish the objective function. The objective function is to minimize the maximum completion time of all transportation operations, that is, to minimize the makespan. This function calculates the maximum time for all AGVs S to complete their transportation operations and charging operations. Specifically, minimize the maximum completion time of all transportation operations:

[0100] min z = C max

[0101] S13. Construct the constraint conditions; To ensure that the scheduling of the AGV s meets the limitations of the actual operation, the constraint conditions include: Each transportation operation can only be completed by one AGV; The AGV s must be charged before the battery runs out; A charging facility can only provide charging service to one AGV at the same time; The number of charging facilities is limited; The AGV s needs to share these charging facilities; The power consumption and charging strategy of the AGV s need to conform to the actual battery capacity and charging rate. The specific constraint conditions are as follows:

[0102] 1) Transportation operation scheduling constraint:

[0103] Each transportation operation is assigned after the r-th charging operation of the AGV k :

[0104]

[0105] v jkr ≤ w kr

[0106] When the AGV k performs the r - th charging operation, the number of transportation jobs j assigned to the AGV k is greater than or equal to 1:

[0107]

[0108] When the AGV k does not perform the (r - 1)-th charging operation, then the r - th charging operation is not performed:

[0109] w kr ≤ w k,r-1 .

[0110] 2) Battery energy flow balance constraint:

[0111] The remaining battery power of the AGV k before the r - th charging operation is equal to the remaining battery power of the AGV k before the (r - 1)-th charging operation plus the battery power of the (r - 1)-th charging operation minus the energy consumption of the transportation jobs performed between the (r - 1)-th charging operation and the r - th charging operation:

[0112]

[0113] Impose a battery capacity limit on each charging operation:

[0114] τδ kr + Δ kr ≤ b

[0115] The remaining battery power of the AGV k is non - negative:

[0116] Δ kr ≥ 0

[0117] When the r - th charging operation of the AGV k is not performed, then the charging duration of the r - th charging operation is 0:

[0118]

[0119] The AGV k is initially in a charged state, and the initial battery power is equal to the battery capacity of the AGV k :

[0120] w k1 = 1

[0121] Δk1 = b

[0122] AGV k The r-th charging operation of the AGV must be assigned to a charging facility s:

[0123]

[0124] 3) Charging transportation operation scheduling constraints:

[0125] If the r-th charging operation of the AGV k and the r-th charging operation of the AGV k′ are assigned to the same charging facility s, and the r ′ is executed after the r ′ , then the completion time of the r ′ is greater than or equal to the completion time of the r plus the charging time of the r ′ :

[0126] C k′r′ ≥ C kr + δ k′r′ - M(3 - x krk′r′ - y krs - y k′r′s )

[0127] C kr ≥ C k′r′ + δ kr - M(2 + x krk′r′ - y krs - y k′r′s )

[0128] AGV k The completion time of the r-th charging operation of the AGV is greater than or equal to the sum of the completion time of the (r - 1)-th charging operation, the charging duration of the r-th charging operation, the charging preparation time, and the processing time of the transportation operation of the AGV after the (r - 1)-th charging operation: k :

[0129]

[0130] AGV k The maximum completion time is greater than or equal to the completion time of the r-th charging operation of the AGV and the total processing time of the transportation operation after the r-th charging operation: k :

[0131]

[0132] Define the domain of the decision variables, including:

[0133] v jk,r ∈ {0, 1}

[0134] y krs ∈ {0, 1}

[0135] w kr ∈ {0, 1}

[0136] x krk′r′ ∈ {0, 1}

[0137] Structural performance analysis and lower bounds, including:

[0138] Minimize the makespan of all transportation operations:

[0139] min z = C max

[0140] Ensure that for each AGV k , its completion time C max is at least equal to the sum of the processing times of all jobs assigned to it on this AGV:

[0141]

[0142] Ensure that each job j is assigned to exactly one AGV:

[0143]

[0144] Define the domain of decision variables:

[0145] v jk ∈ {0, 1}

[0146] Calculate the lower bound of the AGV scheduling problem:

[0147]

[0148] S14. Integrated model; In this implementation, preferably, the objective function and all constraint conditions are integrated into a MILP model.

[0149] S2. Model solution, using the adaptive large neighborhood search method, generate an initial solution, design destruction and repair operators to optimize the initial solution, and introduce a local search mechanism to further optimize the solutions obtained during iteration. The solution obtained provides the transportation job assignment and charging schedule for the AGV s , including:

[0150] S21. Generate an initial solution, using a three - stage heuristic method to generate an initial feasible solution. In the first stage, assign transportation jobs to the AGV s, similar to a parallel machine scheduling problem, solved using the branch and bound method. Second stage: Using the Arc-flow method, further allocate transportation operations to charging operations. Third stage: Through the "first come, first served" scheduling rule, for all AGVs s Arrange charging operations on a limited number of charging facilities and form a feasible charging schedule.

[0151] In the first stage, allocate transportation operations to AGVs s , including:

[0152] Minimize the total completion time in the AGV scheduling problem:

[0153] min z = C max

[0154] This constraint is used to calculate the lower bound of the maximum completion time of each AGV:

[0155]

[0156] where Γ k represents the number of charging operations of the AGV k ;

[0157] Ensure that the number of charging times of the AGV k can at least meet its minimum charging requirement for completing all operations, and the key part that the AGV k will not stop due to insufficient power during the execution of transportation operations:

[0158]

[0159] Ensure that the number of charging operations of the AGV k does not exceed the number of operations assigned to it:

[0160]

[0161] Ensure that each transportation operation j is exactly assigned to one AGV k :

[0162]

[0163] Define the value range of the variable v jk :

[0164] v jk ∈ {0, 1}

[0165] Define that the variable Γ k can only take non-negative integers:

[0166]

[0167] Second stage: Using the Arc-flow method, further allocate transportation operations to charging operations and schedule the operations on each AGV, including:

[0168] Minimize the maximum completion time of a specific AGV k :

[0169]

[0170] Calculate the lower bound of the maximum completion time of each AGV k to ensure that the maximum completion time of each AGV is at least equal to the sum of the processing time, charging time, and setup time of all the operations it undertakes:

[0171]

[0172] Ensure that each job j is assigned to exactly one charging operation:

[0173]

[0174] h jr : If transportation job j is executed after the r-th charging operation on AGV k , then h jr = 1, otherwise 0;

[0175] The constraint ensures that if charging operation r occurs, then the previous charging operation r - 1 must also occur:

[0176] l r ≤l r-1

[0177] l r : If the r-th charging operation is executed, then l r = 1, otherwise 0;

[0178] If charging operation r occurs, then at least one job is assigned after this charging operation:

[0179] Ensure that if job j is assigned after charging operation r, then charging operation r must occur:

[0180] Ensure that before any charging operation r, the remaining battery power of AGV k is sufficient to complete all the jobs assigned after this charging operation:

[0181]

[0182] Define the value range of variable h jr :

[0183] hjr ∈ {0, 1}

[0184] Define the variable l r The value range of:

[0185] l r ∈ {0, 1}

[0186] The third stage: Arrange charging operations on a limited number of charging facilities through the "first come, first served" scheduling rule and form a feasible charging schedule, including:

[0187] Minimize the maximum completion time for all AGVs to complete all jobs:

[0188] min z = C max

[0189] AGV k The remaining power Δ before the r-th charging operation of the kr Calculation of:

[0190]

[0191] AGV k The charging amount of the r-th charging operation cannot exceed the battery capacity b of the AGV:

[0192]

[0193] For AGV k For the r-th charging operation, a charging facility must be selected from all charging facilities s for charging:

[0194]

[0195] AGV k The completion time of the r-th charging operation of the k,r-1 is at least the completion time C of its previous charging operation kr plus the duration δ of the current charging operation max

[0196]

[0197] The maximum completion time C of all transportation operations max is at least the completion time C of the r-th charging operation of AGV k kr plus the sum of the processing times of all transportation operations performed after this charging operation:

[0198]

[0199] S22. Select design destruction and repair operators. The destruction operator is used to perform destruction on the basis of the current solution to explore a new solution space. Randomly remove a certain number of transportation operations, remove the transportation operations that contribute the most to the total completion time, remove the transportation operations that result in the highest total cost, randomly remove charging operations, remove the most time-consuming charging operations, and remove the charging operations with the highest cost. The repair operator is used to reconstruct the solution using the repair operator after the destruction operation to restore the feasibility of the solution. Specifically, it includes randomly inserting the removed transportation operations, greedily inserting transportation operations according to the specific attributes of the transportation operations, considering the flexibility of inserting transportation operations, using regret values to guide the insertion of transportation operations, randomly inserting the removed charging operations, and greedily inserting charging operations according to the specific attributes of the charging operations.

[0200] S23. Through the local search mechanism, use the splitting operator and the moving operator to identify charging operations with waiting time and reduce the waiting time of AGVs.

[0201] The splitting operator identifies the AGV with the longest charging completion time k , and the r-th charging operation of the AGV k has a waiting time; determine the charging facility s where the r-th charging operation of the AGV k is located, identify the previous charging operation of the r-th charging operation of the AGV k on the charging facility s as the r-th charging operation of the AGV k′ , split the r-th charging operation of the AGV ′ into two charging operations r k′ and r ′ , assign the transportation operations after the r-th charging operation to the split r-th charging operation and after the r-th charging operation; 1 ′ and r 2 ′ , and after the r-th charging operation ′ ; 1 ′ charge operation and after the r-th charging operation; 2 ′ charge operation;

[0202] The moving operator first identifies the AGV with the longest charging completion time k , and the r-th charging operation of the AGV k has a waiting time; determine the charging facility s where the r-th charging operation of the AGV k is located, identify the previous charging operation of the r-th charging operation of the AGV k on the charging facility s as the r-th charging operation of the AGV k′ ; assign the transportation operations assigned to the r-th charging operation of the AGV ′ to the r-th charging operation of the AGV k′ ; ′Transport operation after the first charging operation is moved to the AGV k′ the r-th ′ after the (r + n)-th charging operation.

[0203] Repeat the above steps until the iteration count limit is reached.

[0204] In this embodiment, the system can be used through a visualization interface to display scheduling schemes, AGV status, charging facility occupancy, etc., enabling workshop managers to intuitively understand the production status and facilitating real-time monitoring and adjustment.

[0205] Combined with Figure 1 The process of the adaptive large neighborhood search method is further described as follows:

[0206] 1. Start: The starting point of the algorithm, initializing all necessary parameters and variables.

[0207] 2. Initialization: Generate an initial feasible solution to prepare for the iterative process of the algorithm.

[0208] 3. Main iteration starts: Enter the iterative loop of the algorithm, with each iteration aiming to improve the current solution.

[0209] 4. Select operators: Based on the adaptive mechanism, select destruction and repair operators suitable for the current solution.

[0210] 5. Apply the destruction operator: Random job or charging operation removal: Randomly select a job or charging operation to remove from the current solution to break the existing structure.

[0211] 6. Apply the repair operator: Random / greedy / regret job or charging operation insertion: According to the processing time and energy consumption of the job, use the adaptive selection mechanism to re-insert the job into the job sequence of the AGV.

[0212] 7. Local search: Further optimize the quality of the solution through local search operators such as splitting and moving operators.

[0213] 8. Evaluate the solution quality: Evaluate the total completion time of the current solution and other possible metrics.

[0214] 9. Simulated annealing acceptance criterion: Use the simulated annealing method to decide whether to accept the new solution to balance exploration and exploitation.

[0215] 10. Update the best solution: If the new solution is better than the current best solution, update the best solution record.

[0216] 11. Iteration termination condition: Check whether the algorithm has reached the preset iteration count or meets other termination conditions.

[0217] 12. Output the result: Output the best solution found by the algorithm for further analysis or practical application.

[0218] 13. End: The algorithm completes all iterations and ends the operation.

[0219] Combined with computational experiments on 729 randomly generated instances, the performance of the adaptive large neighborhood search algorithm in the present invention is further illustrated:

[0220] 1. Instance generation: According to the number of AGVs, the number of transportation operations, and the number of charging facilities, randomly generate experimental instances of different scales.

[0221] 2. Parameter setting: Set the key parameters for the adaptive large neighborhood search algorithm, including the number of iterations, the number of segment iterations, the size of the destruction operation, the initial temperature control parameter of simulated annealing, and the cooling rate.

[0222] 3. Initial solution generation: Use a three-stage heuristic method to generate an initial solution: The first stage: Allocate transportation operations to AGVs s ; The second stage: Further allocate transportation operations to charging operations; The third stage: Schedule charging operations on a limited number of charging facilities.

[0223] 4. Execution of the adaptive large neighborhood search algorithm: Apply destruction and repair operators to explore the solution space. Use an adaptive mechanism to dynamically select operators. Decide whether to accept a new solution through the acceptance criterion of simulated annealing. Use a local search mechanism to improve the current best solution.

[0224] 5. Performance evaluation: Compare the performance of the adaptive large neighborhood search algorithm with the CPLEX solver, including the quality of the solution (objective value) and the computational time. Analyze the experimental results of instances of different scales to evaluate the efficiency and effectiveness of the algorithm.

[0225] 6. Sensitivity analysis: Adjust the key parameters (the number of charging facilities, setup time, and job energy consumption ratio) and observe the impact on the solution quality.

[0226] 7. Result analysis: Analyze the experimental results, including the convergence speed, stability of the algorithm, and adaptability to instances of different scales. Through performance analysis, such as the time-objective graph, evaluate the performance of the algorithm in practical applications.

[0227] Through the method of the present invention, the adaptive large neighborhood search algorithm can effectively explore the solution space, generate high-quality initial solutions, and improve the solution in each iteration, thus providing an effective solution for the AGV scheduling problem.

[0228] In the instance test, by increasing the number of charging facilities, the job completion time is reduced from 341.5 seconds to 284.5 seconds, a decrease of about 17%. In multiple experiments and tests, the solution presented by the present invention shows high stability and can maintain excellent performance in instances of different scales and complexities.

[0229] The adaptive large neighborhood search algorithm performs excellently in terms of computational time, especially in large-scale instances. For example, in an instance with 80 transportation operations, 10 AGVs, and 5 charging facilities, the average computational time is only 29.05 seconds, demonstrating efficient computational capabilities.

[0230] Through the sensitivity analysis of parameters such as the number of charging facilities, charging setup time, and job energy consumption ratio, the present invention shows good robustness and can cope with various changes in the production environment.

[0231] By introducing a flexible charging strategy in the present invention, the charging operations of AGVs are reasonably arranged, avoiding unnecessary charging waiting time, thus reducing the overall transportation operation completion time; in the sensitivity analysis of the change in the number of charging facilities, compared with one charging facility, two charging facilities not only reduce the number of charging times but also balance the transportation operation distribution of AGVs, further saving energy consumption; under the conditions of limited charging facilities and charging setup time, the present invention effectively improves the resource utilization rate and reduces the idle time of AGVs through intelligent scheduling. The decision support system provided by the present invention automatically generates a scheduling plan without frequent manual intervention, reducing the operation complexity. At the same time, the system supports flexible adjustment of the scheduling plan to adapt to different production requirements.

[0232] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An AGV scheduling method considering charging facility resource constraints and flexible charging strategies, characterized in that: include: The decision variables are transportation operation allocation variables, charging timing and duration variables, the task order of transportation operations on the same AGV, the AGV charging order on each charging facility, the allocation variables of AGVs at charging facilities, and the completion time of all transportation operations and charging operations; the objective function is to minimize the maximum completion time of all transportation operations; and the transportation operation scheduling constraints, battery energy flow balance constraints, and charging scheduling constraints are used as constraints to construct a mixed integer linear programming model. The mixed integer linear programming model is solved by an adaptive large domain search algorithm to obtain an AGV scheduling solution; The method of solving the mixed integer linear programming model by using an adaptive large neighborhood search algorithm includes: An initial feasible solution is generated through a three-stage heuristic method; By introducing destruction operators and repair operators, the search range of the initial feasible solution space is expanded; A local search mechanism is introduced to update the optimal solution obtained through iterative search.

2. According to claim 1, an AGV scheduling method considering charging facility resource constraints and flexible charging strategies is characterized in that: The transport operation allocation variables, charging timing and duration variables, the task sequence of transport operations on the same AGV, the AGV charging sequence on each charging facility, the AGV allocation variables on the charging facility, and the completion time of all transport operations and charging operations include: v jkr :If transport operation j is assigned to AGV k After the rth charging operation, v jkr is equal to 1, otherwise it is 0; w kr :If AGV k After the rth charging operation is performed, w kr is equal to 1, otherwise it is 0; δ kr :AGV k The charging duration of the rth charging operation; Δ kr :AGV k The remaining power before the rth charging operation is performed; y krs :If AGV k The rth charging operation of is performed on charging facility s, then y krs is equal to 1, otherwise it is 0; x krk′r ,: If AGV k The rth charging operation and AGV k′ The r′th charging operation is assigned to the same charging facility, and the AGV k The rth charging operation in AGV k′ Before the r′th charging operation, x krk′r′ is equal to 1, otherwise it is 0; C kr :AGV k The completion time of the rth charging operation.

3. According to claim 1, an AGV scheduling method considering charging facility resource constraints and flexible charging strategies is characterized in that: The transportation operation scheduling constraints include: Each transport operation is assigned to an AGV k After the rth charging operation: in jkr ≤w kr When AGV k After the rth charging operation is performed, the AGV is assigned k The number of transport operations j is greater than or equal to 1: When AGV k When the r-1th charging operation is not performed, the rth charging operation is also not performed: In kr ≤in k,r-1 。 4. The AGV scheduling method according to claim 1 that considers charging facility resource constraints and flexible charging strategies, characterized in that: The battery energy flow balance constraint includes: AGV k The remaining power before the rth charging operation is equal to the AGV k The remaining power before the r-1th charging operation plus the power of the r-1th charging operation minus the energy consumption of the transportation operation between the r-1th charging operation and the rth charging operation: To impose a battery capacity limit on each charging operation: td kr +D kr ≤b AGV k The remaining power is non-negative: D kr ≥0 When AGV k If the rth charging operation is not performed, the charging duration of the rth charging operation is 0: AGV k Initially it is in a charged state, and the initial power is equal to the AGV k Battery capacity: w k1 ＝1 D k1 =b AGV k The rth charging operation must be assigned to a charging facility s: Among them, J: the set of transportation operations, represented by indexes i, j; K: the set of AGVs, represented by index k; R: the set of charging operations, represented by index r; S: the set of charging facilities, represented by index s.

5. The AGV scheduling method according to claim 1 considering charging facility resource constraints and flexible charging strategies is characterized in that: The charging scheduling constraints include: If AGV k The rth charging operation and AGV k′ The r'th charging operation of is assigned to the same charging facility s, and r' is performed after r, then the completion time of r' is greater than or equal to the completion time of r plus the charging time of r': C k′r′ ≥C kr +δ k′r′ -M(3-x krk′r′ -y krs -y k′r′s ) C kr ≥C k′r′ +δ kr -M(2+x krk′r′ -y krs -y k′r′s ) AGV k The completion time of the rth charging operation is greater than or equal to the completion time of the r-1th charging operation, the charging duration of the rth charging operation, the charging preparation time, and the AGV after the r-1th charging operation k The sum of the processing time for transport operations: AGV k The maximum completion time is greater than or equal to AGV k The completion time of the rth charging operation and the total processing time of the transport operation after the rth charging operation: C max ≥C kr +∑ j∈J v jkr p j 。 6. The AGV scheduling method according to claim 1 that considers charging facility resource constraints and flexible charging strategies, characterized in that: The three-stage heuristic method is used to generate an initial feasible solution, including: Phase 1: Use the branch and bound method to assign transportation tasks to AGVs; Phase 2: Using the Arc-flow method, each AGV is assigned transport and charging operations. Phase 3: Scheduling charging operations on a preset number of charging facilities through a first-come, first-charge scheduling rule.

7. The AGV scheduling method according to claim 1, wherein: The method of expanding the search range of the initial feasible solution space by introducing a destruction operator and a repair operator includes: Use a roulette wheel selection mechanism to select the destruction operator or the repair operator; The destruction operator destroys the current solution and explores new solution space; Repair Operator After the destruction operation, the repair operator is used to reconstruct the solution and restore the feasibility of the solution.

8. The AGV scheduling method according to claim 7, which takes into account the resource limitation of charging facilities and the flexible charging strategy, is characterized in that: The destruction operator includes: Random transport operations removed, critical transport operations removed, worst transport operations removed, random charging operations removed, critical charging operations removed, worst charging operations removed; The repair operators include: random transport operation insertion, greedy transport operation insertion, regret value transport operation insertion, random charging operation insertion, greedy charging operation insertion and regret value charging operation insertion.

9. The AGV scheduling method according to claim 1, which takes into account the resource limitation of charging facilities and the flexible charging strategy, is characterized in that: The local search mechanism includes a split operator and a move operator; The segmentation operator identifies the AGV with the longest charging completion time k , and the AGV k The rth charging operation has a waiting time; determine the AGV k The charging facility s where the rth charging operation is located is identified, and the AGV on the charging facility s is identified. k The previous charging operation of the rth charging operation is AGV k′ The r′th charging operation, the AGV k′ The r′th charging operation is divided into two charging operations r′1 and r′2, and the transportation operation allocated after the r′th charging operation is allocated after the divided r′1th charging operation and r′2th charging operation; The mobile operator first identifies the AGV with the longest charging completion time k , and the AGV k The rth charging operation has a waiting time; determine the AGV k The charging facility s where the rth charging operation is located is identified, and the AGV on the charging facility s is identified. k The previous charging operation of the rth charging operation is AGV k′ The r′th charging operation will be allocated to the AGV k′ The transport operation after the r′th charging operation moves to the AGV k′ After the r′+nth charging operation.

10. The AGV scheduling method according to claim 1, wherein: The input parameters of the mixed integer linear programming model include: p j : Processing time of transport operation j, e j : energy consumption of transport operation j, b: battery capacity of each AGV, τ: charging rate of each AGV, and λ: setup time of each charging operation.

Images