Multi-bucket wheel machine cooperative operation scheduling system and method based on swarm intelligence algorithm

Abstract

The application provides a multi-bucket wheel machine cooperative operation scheduling system and method based on a swarm intelligence algorithm, relates to the field of electronic information technology, acquires equipment information, stacking information and environmental information; inputs the equipment information, the stacking information and the environmental information into a first model to generate a constraint condition; performs global search based on an artificial bee colony algorithm introducing a harmony search, wherein, in each iteration, the solution is screened according to the constraint condition to obtain a group of feasible solutions; performs local search in the neighborhood of each feasible solution by using a mixed integer programming to obtain a global optimal solution; generates the working parameters of the equipment according to the global optimal solution to perform operation scheduling. The application improves the artificial bee colony algorithm by introducing the harmony search mechanism, enhances the global search capability and avoids premature convergence; meanwhile, the local optimization is performed in combination with the mixed integer programming to improve the accuracy of the solution, so that the algorithm can adapt to the stacking yard scheduling requirements of different scales and different constraint conditions and improve the overall operation efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electronic information, in particular to a multi-bucket wheel machine cooperative operation scheduling system and method based on a swarm intelligence algorithm. BACKGROUND

[0002] The bucket wheel machine is a high-efficiency continuous operation bulk material handling equipment, which is widely used in large bulk material yards such as ports, power plants and steel plants, and undertakes the tasks of material stacking, material taking and material transfer. During the material taking operation, the bucket wheel machine rotates to take material and outputs the material through the forward running belt conveyor, while the rotation and pitching mechanisms are adjusted to adjust the operation range; during the material stacking operation, the belt conveyor runs in reverse to deliver the material to the designated position, and the rotation and pitching mechanisms are adjusted to adjust the stacking trajectory.

[0003] A material stacking site usually arranges multiple bucket wheel machines for cooperative operation, and multiple bucket wheel machines usually need to be dynamically scheduled between multiple material piles to meet the material demand of different production links. However, during the scheduling process, spatial interference or motion conflict may occur between multiple bucket wheel machines, and as the automation and intelligence level of the material yard is improved, the cooperative operation demand of multiple bucket wheel machines is increasing, and how to optimize the scheduling strategy, avoid interference and improve the overall operation efficiency has become a key problem in the operation and management of the bulk material yard. SUMMARY

[0004] The present application relates to the technical field of electronic information, in particular to a multi-bucket wheel machine cooperative operation scheduling system and method based on a swarm intelligence algorithm.

[0005] In order to achieve the above-mentioned purpose, the technical scheme adopted by the present application is as follows:

[0006] Obtaining equipment information, material stacking information and environmental information;

[0007] Inputting the equipment information, material stacking information and environmental information into a first model to generate constraint conditions;

[0008] Performing global search based on an artificial bee colony algorithm with introduced harmony search, wherein a group of feasible solutions is obtained by screening the solutions according to the constraint conditions in each iteration;

[0009] Performing local search in the neighborhood of each feasible solution by using mixed integer programming to obtain a global optimal solution;

[0010] Generating the working parameters of the equipment according to the global optimal solution to perform operation scheduling.

[0011] The present application relates to the technical field of electronic information, in particular to a multi-bucket wheel machine cooperative operation scheduling system and method based on a swarm intelligence algorithm.

[0012] The first module is used for acquiring equipment information, stockpile information and environment information;

[0013] The second module is used for inputting the equipment information, the stockpile information and the environment information into a first model to generate a constraint condition;

[0014] The third module is used for performing global search based on an artificial bee colony algorithm with introduction of a harmony search mechanism, wherein a set of feasible solutions is obtained by screening solutions according to the constraint condition in each iteration;

[0015] The fourth module is used for performing local search in a neighborhood of each feasible solution by using a mixed integer programming to obtain a global optimal solution;

[0016] The fifth module is used for generating working parameters of the equipment according to the global optimal solution to perform job scheduling.

[0017] The application has the following beneficial effects:

[0018] The application improves the artificial bee colony algorithm by introducing the harmony search mechanism, enhances the global search capability and avoids premature convergence; meanwhile, the local optimization is performed by combining the mixed integer programming to improve the accuracy of the solution, so that the algorithm can adapt to the stockyard scheduling requirements of different scales and different constraint conditions, can quickly generate a high-quality scheduling scheme under complex constraint conditions, optimizes the collaborative job order and path of the multi-bucket wheel machine, reduces the equipment idle running and waiting time, and improves the overall job efficiency.

[0019] Other features and advantages of the present application will be illustrated in the following description, and some will become apparent from the description, or will be understood by those skilled in the art from the description, or will be understood by implementing the embodiments of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope, and for those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0021] Figure 1 The flow chart of the multi-bucket wheel machine collaborative job scheduling method based on the swarm intelligence algorithm in the embodiments of the present application;

[0022] Figure 2 The structural diagram of the multi-bucket wheel machine collaborative job scheduling system based on the swarm intelligence algorithm in the embodiments of the present application;

[0023] In the figure, the symbols are as follows: 100 - first module; 200 - second module; 300 - third module; 310 - first unit; 320 - second unit; 330 - third unit; 331 - first subunit; 332 - second subunit; 333 - third subunit; 340 - fourth unit; 350 - fifth unit; 360 - sixth unit; 370 - seventh unit; 400 - fourth module; 500 - fifth module. DETAILED DESCRIPTION

[0024] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts fall within the scope of protection of the present application.

[0025] It should be noted that: similar reference numerals and letters represent similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0026] Embodiment 1

[0027] Referring to Figure 1 The embodiment provides a multi-bucket wheel machine cooperative operation scheduling method based on a swarm intelligence algorithm, which comprises steps S100, S200, S300, S400 and S500.

[0028] S100, obtaining device information, stacking information and environmental information;

[0029] The device information comprises device type information, device fault information, device main component temperature information and device current position information.

[0030] The stacking information comprises stacking position, stacking geometric parameter, material type, priority identifier and corresponding stacking rule (layering requirement, slope limitation, anti-collapse safety angle).

[0031] The environmental information comprises temperature, humidity, wind speed, rainfall and dust concentration; wherein, the factors such as temperature, humidity, wind speed and rainfall will affect the operation time and height of the device, and the area with too high dust concentration will be divided into an operation prohibited area, and the specific division threshold is related to the stacking type, and the concentration will be reduced after watering.

[0032] S200, inputting the device information, stockpile information and environment information into a first model to generate a constraint condition;

[0033] The first model is constructed in advance according to historical data and expert knowledge. On the one hand, hard constraints are defined by expert knowledge, such as safe distance between devices, stockpile height constraint under corresponding weather, working height constraint under certain wind speed and humidity, maximum temperature limit of main components, working radius constraint of bucket wheel machine, maximum working time of device under certain temperature, etc.

[0034] In addition to the conventional physical constraints, due to the different geological structures of each stockpile yard and the change in obstacle distribution, it is necessary to disable part of the devices or divide temporary exclusion zones at some time, so historical data and corresponding disable information are also collected to construct a corresponding disable list as one of the constraint conditions.

[0035] The first model can combine a preset rule table with a pre-trained neural network. The rule table is used to query hard constraints that cannot be violated. The neural network is constructed by training historical data to generate other constraint items according to the input information.

[0036] S300, global search based on artificial bee colony algorithm with introduction of harmony search, wherein a set of feasible solutions is obtained by screening the solutions according to the constraint condition in each iteration, and the specific process is as follows:

[0037] S310, randomly generating an initial population, searching based on the initial population to obtain a feasible solution and storing it in a harmony memory bank;

[0038] The encoding of the initial population needs to include task allocation of each device, execution task order, device working speed and task start time, etc., which includes discrete variables and continuous variables. This step adopts a matrix mixing method for encoding, and each row of the matrix corresponds to a device. Each row includes task sequence, working speed and task start time in turn.

[0039] Artificial bee colony algorithm (ABC) is a swarm intelligence optimization algorithm based on the behavior of honey bees collecting nectar. It simulates the behavior of a honey bee colony searching for nectar sources (solutions), which contains three types of bee roles:

[0040] Employed Bees: responsible for developing known nectar sources (local search).

[0041] Onlooker Bees: further develop high-quality nectar sources according to roulette selection (global exploration).

[0042] Scout Bees: abandon low-quality food sources and search randomly for new food sources (avoid local optima).

[0043] Wherein, the employed bees search around the initial food sources to find better food sources, the employed bees return the information of food sources and fitness to the onlooker bees, the onlooker bees determine the selection probability according to the fitness of each food source, each onlooker bee selects a food source, the selected food source generates a new solution and updates in the same way as the employed bees, if the quality of a food source does not improve after n iterations, the food source is abandoned, the onlooker peak becomes the scout peak, and a new food source is randomly selected in the global.

[0044] For each solution searched by the artificial bee colony algorithm, the maximum working time, working height, minimum distance between devices and maximum load of each device corresponding to the solution can be obtained, so that it can be judged whether the working information corresponding to the solution violates the constraint condition, if so, the solution is infeasible.

[0045] S320, in each iteration, for each variable in each feasible solution, determine whether to select a corresponding variable from the harmony memory to replace it to obtain a new solution with a first probability; if not, randomly select in the value range of the variable to generate a random new solution;

[0046] Harmony search (HS) is a heuristic global search algorithm, which realizes optimization through the following steps:

[0047] Initialization of harmony memory: a certain number of initial solutions (harmonies) are generated and stored in the harmony memory (HM).

[0048] Generating new harmony: determine whether to select a variable from the HM according to the harmony memory selection probability (HMCR), if selected, it can be fine-tuned (adjustment probability is PAR, adjustment amplitude is bandwidth BW), otherwise, a variable is randomly generated in the solution space.

[0049] Update the harmony memory: if the newly generated harmony is better than the worst harmony in the HM, replace it.

[0050] Iterative optimization: repeat the above process until a predetermined number of iterations is reached.

[0051] The present application combines harmony search into the artificial bee colony algorithm, constantly updates the harmony memory, and retains better solutions;

[0052] Employing bees search around the center of the updated harmony memory in each iteration. The traditional observer bee stage selects high-quality solutions by roulette selection, which may over-focus on local areas. The randomness and diversity of HS can make up for the defect of ABC falling into local optimum. The improvisation mechanism of harmony memory (randomly replacing variables) can quickly generate diverse solutions and quickly explore the solution space, with low dependence on the quality of initial solutions.

[0053] The first probability is the harmony memory value probability (HMCR). In the early exploration stage, a small first probability value (such as 0.6-0.7) can be set to quickly explore the solution space, and a larger first probability (such as 0.95) can be set in the later stage to highlight the local development ability of ABC to refine the solution in the later stage.

[0054] S330, adjust each variable of the new solution by Kent mapping with a second probability. If the adjusted new solution is better than the worst solution in the harmony memory, replace the worst solution.

[0055] The second probability is the adjustment probability PAR. After replacing the variables in the solution, the probability PAR is used to determine whether to adjust the variables. The variables in the solution are divided into discrete variables and continuous variables, and the adjustment methods for the two types of variables are slightly different. The adjustment method is as follows:

[0056] S331, construct an adaptive bandwidth function based on the number of iterations, population diversity, and convergence state;

[0057] Population diversity is the standard deviation of the current population fitness. If the population diversity is high, a larger bandwidth should be maintained to encourage exploration. If the diversity is low, the bandwidth should be reduced to enhance local development. As the number of iterations progresses, the bandwidth should also be gradually reduced to enhance local development.

[0058] Convergence state refers to detecting whether it has fallen into local optimum. If the change in fitness value of the last k generations is too small, it is considered to be in a state of convergence stagnation, and the bandwidth needs to be increased to jump out.

[0059] Let the number of iterations be t, the population diversity be D t , and the convergence state be C t . Consider the three adjustment terms to construct the adaptive bandwidth function:

[0060] BW t =B0×α(t)×β(D t ))×γ(C t );

[0061] Where B0 is the initial bandwidth;

[0062] α(t) is the iteration number decay term:

[0063]

[0064] where λ is the decay rate, usually taken as [1, 5]; t max is the preset maximum number of iterations;

[0065] β(D t ) is a population diversity adjustment term:

[0066]

[0067] where η is an adjustment coefficient, taken as [0, 1]; D min and D max are the historical minimum and maximum values of population diversity.

[0068] γ(C t ) is a convergence state adjustment term:

[0069]

[0070] where κ is a steepness coefficient; C th is a convergence threshold (an empirical value, such as 10 -5 ).

[0071] S332, based on the adaptive bandwidth function, dynamically adjusting the bandwidth parameter of the sound search, and determining an adjustment range based on the bandwidth parameter;

[0072] According to the adaptive bandwidth function, the bandwidth parameter BW t of the current iteration is calculated.

[0073] Based on the bandwidth parameter, an adjustment range is determined:

[0074] For continuous variables: the adjustment range is BW t ×N(-1, 1).

[0075] For discrete variables: the adjustment range is round(BW t ×N(-1, 1)); Round represents rounding. It should be noted that discrete variables generally have only a few consecutive integer values. If the integer values are directly added or subtracted from the discrete variables, it may exceed the solution space range of the discrete variables. Therefore, the adjustment range of the discrete variable defined in the present application refers to the number of moving bits to the left and right of the discrete variable. For example, a discrete variable takes values of {1, 2, 3, 4}, the adjustment range is ±1, and the adjustment range is 3 and 1 based on the value 4 (when the value is at the head / tail position, the tail / head value is considered as its adjacent value).

[0076] It should be noted that the initial bandwidth B0 of different variables is different and needs to be set respectively.

[0077] S333, determining whether to make adjustment according to the second probability, if yes, generating a random number by using Kent mapping in the adjustment range, and adding the random number to the variable.

[0078] Kent mapping is a pseudo-random number generation method based on chaotic dynamics, and the random number generated by Kent mapping has the characteristic of uniform distribution. The application generates a random number sequence in advance by using Kent mapping, and obtains a random number set by disordering the random number sequence;

[0079] In the iteration of harmony search, each time a random number is needed, the first random number is extracted from the random number set and mapped to the required range by taking modulo operation or hash mapping, and after extraction, the used random number is deleted from the random number set to ensure the randomness of each adjustment.

[0080] S340, dividing the solution space into multiple grid regions, and statistically calculating the improvement score of each region based on the distribution of historical optimal solutions;

[0081] The number of historical optimal solution distribution in each region is counted, and the ratio of the number of historical optimal solutions in the region to the total number of historical optimal solutions is taken as the improvement score of the region. The greater the improvement score, the greater the possibility of the optimal solution appearing in the region.

[0082] S350, constructing a fitness function based on total operation time, energy consumption weighting and device load variance;

[0083] The optimization goal of the application is to minimize the total operation time and energy consumption while balancing the load of each device to ensure the use performance of the device. The energy consumption can be calculated according to the operation time and load information of each device. The sum of the operation time, energy consumption weighting and device load variance is weighted to obtain a smaller fitness value.

[0084] S360, in the observation bee phase, calculating the similarity of each solution to the historical optimal solution, and calculating the fitness value of the solution according to the fitness function;

[0085] In the traditional method, each observation bee selects a high-quality honey source for further development according to the search results of the employed bee, that is, the fitness of each honey source is calculated to obtain the probability of selecting the honey source. In the application, the fitness of the solution and its similarity to the historical optimal solution are considered at the same time, so that the observation bee is more inclined to select a solution close to the historical optimal solution.

[0086] S370, obtaining the improvement score of the region where the solution is located, calculating the weighted sum of the fitness and the similarity value of the solution according to the improvement score, and the observation bee selects the solution according to the weighted sum.

[0087] Specifically, the region where the current solution is located is considered in this step, and if the improvement score of the region where the current solution is located is high, the priority of the similarity value can be increased, that is, the weight of the similarity value is increased.

[0088] In the standard ABC, the employed bees perform neighborhood search around the current solution, and the search step is randomly selected within a fixed range, which may result in low search efficiency (the step is too large to skip the optimal solution, or too small to converge slowly). In order to balance the global exploration and local development ability, the application adopts a strategy of dynamically adjusting the search step in the employed peak stage:

[0089] The search step distribution corresponding to the successfully updated solutions in the current generation is counted;

[0090] According to the search step distribution, a kernel density estimation model is established; and is represented by φ, and a Gaussian kernel function is specifically used:

[0091]

[0092] Wherein, W is the total number of steps of the successfully updated solutions, and φ i is the i-th step, h = W -1 / 5 ·σ, σ is the sample standard deviation, and u is the sample mean; K(u) is a probability density function of a standard normal distribution; f(φ) is the estimated probability density of the step φ.

[0093] According to the kernel density estimation model, the search step parameter of the observing bee in the next iteration is adjusted. That is, a new step is extracted from the kernel density estimation model, and the value in the high probability density region is preferentially selected.

[0094] S400, a mixed integer programming is used to perform local search in each feasible solution neighborhood to obtain a global optimal solution;

[0095] Divide the integer variables and continuous variables in each solution; construct the corresponding boundary constraints for each variable;

[0096] The value of the integer variable is determined by the branch and bound method; first, the integer variable is relaxed to a continuous variable, a continuous optimization problem is constructed, the node with the minimum lower bound is preferentially processed, the integer variable whose fractional part is closest to 0.5 in the relaxed solution is preferentially selected as the branch variable, the selected variable is bisected to generate left and right child nodes, the lower bound of the node and the current optimal solution are compared, and pruning is performed.

[0097] After fixing the integer variable, the interior point method is used to optimize the continuous variable to obtain a global optimal solution.

[0098] S500, generating the working parameters of the equipment according to the global optimal solution to perform job scheduling.

[0099] Embodiment 2

[0100] Referring toFigure 2 The embodiment provides a multi-bucket wheel machine cooperative operation scheduling system based on a swarm intelligence algorithm, which comprises the following modules.

[0101] A first module 100 is configured to acquire device information, stockpile information and environment information.

[0102] A second module 200 is configured to input the device information, the stockpile information and the environment information into a first model to generate a constraint condition.

[0103] A third module 300 is configured to perform global search based on an artificial bee colony algorithm with introduction of a harmony search, wherein a feasible solution is obtained by screening the solution according to the constraint condition in each iteration.

[0104] A fourth module 400 is configured to perform local search in a neighborhood of each feasible solution by using a mixed integer programming to obtain a global optimal solution.

[0105] A fifth module 500 is configured to generate a working parameter of the device according to the global optimal solution to perform operation scheduling.

[0106] As an optional implementation, the third module 300 comprises the following units.

[0107] A first unit 310 is configured to randomly generate an initial population, perform search based on the initial population, obtain a feasible solution and store the feasible solution in a harmony memory.

[0108] A second unit 320 is configured to, in each iteration, determine whether a corresponding variable is selected from the harmony memory to replace the corresponding variable to obtain a new solution by replacement at a first probability for each variable in each feasible solution; if not, a random new solution is generated by randomly selecting in a value range of the variable.

[0109] A third unit 330 is configured to adjust each variable of the new solution by replacement by using a Kent mapping at a second probability, and if the adjusted new solution is better than a worst solution in the harmony memory, the worst solution is replaced.

[0110] As an optional implementation, the third unit comprises the following sub-units.

[0111] A first sub-unit 331 is configured to construct an adaptive bandwidth function based on an iteration number, population diversity and a convergence state.

[0112] A second sub-unit 332 is configured to dynamically adjust a bandwidth parameter of the harmony search based on the adaptive bandwidth function, and determine an adjustment range based on the bandwidth parameter.

[0113] A third sub-unit 333 is configured to determine whether to perform adjustment according to the second probability, and if yes, a random number is generated by using the Kent mapping in the adjustment range, and the random number is added to the variable.

[0114] As an optional implementation, the third module 300 comprises:

[0115] The fourth unit 340 is configured to divide the solution space into a plurality of grid regions, and statistically calculate an improvement score of each region based on the distribution of the historical optimal solution;

[0116] The fifth unit 350 is configured to construct a fitness function based on the total operation time, energy consumption weight and device load variance;

[0117] The sixth unit 360 is configured to calculate the similarity of each solution to the historical optimal solution in the observation bee stage, and calculate the fitness value of the solution according to the fitness function;

[0118] The seventh unit 370 is configured to obtain the improvement score of the region where the solution is located, calculate the weighted sum of the fitness and similarity values of the solution according to the improvement score, and select and update the solution according to the weighted sum by the observation bee.

[0119] It should be noted that, in the present document, relational terms such as“first” and“second”, and the like, are used solely to distinguish one entity or action from another entity or action, without necessarily requiring or implying any actual such relationship or order between or among the entities or actions. Moreover, the terms“comprises”,“comprising”, or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by“comprises a...” does not, without more constraints, exclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0120] The above description is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for scheduling collaborative operations of multiple bucket turbines based on swarm intelligence algorithms, characterized in that, include: Obtain equipment information, material stockpile information, and environmental information; The equipment information, material stacking information, and environmental information are input into the first model to generate constraints. A global search is performed based on an artificial bee colony algorithm that incorporates harmony search, where solutions are filtered according to constraints in each iteration to obtain a set of feasible solutions; The global search based on the artificial bee colony algorithm with the introduction of harmony search includes: dividing the solution space into multiple grid regions and statistically analyzing the improvement score of each region based on the distribution of historical best solutions; A fitness function is constructed based on total operation time, energy consumption weighting, and equipment load variance. During the observation phase, the similarity between each solution and the historical best solution is calculated, and the fitness value of the solution is calculated according to the fitness function. Obtain the improvement score of the region where the solution is located, calculate the weighted sum of the fitness and similarity values ​​of the solution based on the improvement score, and observe the bees to select and update the solution based on the weighted sum; By employing mixed-integer programming to perform a local search in the neighborhood of each feasible solution, the global optimal solution can be obtained. The operating parameters of the equipment are generated based on the global optimal solution, and job scheduling is performed.

2. The multi-bucket turbine cooperative operation scheduling method based on swarm intelligence algorithm according to claim 1, characterized in that, Global search based on the artificial bee colony algorithm incorporating harmony search includes: An initial population is randomly generated, and a search is performed based on the initial population to obtain feasible solutions and store them in the harmony memory bank. In each iteration, for each variable in each feasible solution, it is determined with a first probability whether to select a corresponding variable from the harmony memory bank to replace it and obtain a new solution; if not, a new solution is generated by randomly selecting from the value range of the variable. Each variable that replaces the new solution is adjusted with a second probability using the Kent map. If the adjusted new solution is better than the worst solution in the harmony memory, then the worst solution is replaced.

3. The multi-bucket turbine cooperative operation scheduling method based on swarm intelligence algorithm according to claim 2, characterized in that, The variables of the new solution are adjusted with a third probability using the Kent map, including: An adaptive bandwidth function is constructed based on the number of iterations, population diversity, and convergence state. The bandwidth parameters of the harmony search are dynamically adjusted based on the adaptive bandwidth function, and the adjustment range is determined based on the bandwidth parameters. Whether to make an adjustment is determined based on the second probability. If so, within the adjustment range, a random number is generated using the Kent mapping and added to the variable.

4. The multi-bucket turbine cooperative operation scheduling method based on swarm intelligence algorithm according to claim 1, characterized in that, Global search is performed using an artificial bee colony algorithm that incorporates harmony search, including: Statistically analyze the search step size distribution corresponding to the successfully updated solutions in the current generation; Based on the search step size distribution, a kernel density estimation model is established; Adjust the search step size parameter of the observation bee for the next iteration based on the kernel density estimation model.

5. The multi-bucket turbine cooperative operation scheduling method based on swarm intelligence algorithm according to claim 1, characterized in that, Mixed-integer programming is used to perform a local search in the neighborhood of each feasible solution to obtain the global optimum, including: Divide the integer variables and continuous variables in each solution; Determine the value of an integer variable using the branch and bound method; After fixing the integer variables, the interior point method is used to optimize the continuous variables and obtain the global optimal solution.

6. A multi-bucket turbine collaborative operation scheduling system based on swarm intelligence algorithm, characterized in that, include: The first module is used to acquire equipment information, material stacking information, and environmental information; The second module is used to input the equipment information, material stacking information and environmental information into the first model to generate constraint conditions; The third module is used for global search based on the artificial bee colony algorithm with introduced harmony search. In each iteration, the solutions are filtered according to the constraints to obtain a set of feasible solutions. The third module includes: The fourth unit is used to divide the solution space into multiple grid regions and statistically analyze the improvement score of each region based on the distribution of historical best solutions. The fifth unit is used to construct a fitness function based on total operating time, energy consumption weighting, and equipment load variance; Unit 6 is used during the observation bee phase to calculate the similarity between each solution and the historical best solution, and at the same time to calculate the fitness value of the solution based on the fitness function. The seventh unit is used to obtain the improvement score of the region where the solution is located, calculate the weighted sum of the fitness and similarity values ​​of the solution based on the improvement score, and observe the bees to select and update the solution based on the weighted sum; The fourth module is used to perform a local search in the neighborhood of each feasible solution using mixed integer programming to obtain the global optimal solution; The fifth module is used to generate the operating parameters of the equipment based on the global optimal solution and to perform job scheduling.

7. The multi-bucket turbine collaborative operation scheduling system based on swarm intelligence algorithm according to claim 6, characterized in that, The third module includes: The first unit is used to randomly generate an initial population, search based on the initial population, obtain feasible solutions, and store them in the harmony memory bank. The second unit is used in each iteration to determine, with a first probability, whether to select a corresponding variable from the harmony memory bank to replace it for each variable in each feasible solution to obtain a new alternative solution; if not, a new random solution is generated by randomly selecting from the value range of the variable. The third unit is used to adjust each variable of the new solution with a second probability through the Kent map. If the adjusted new solution is better than the worst solution in the harmony memory, then the worst solution is replaced.

8. The multi-bucket turbine collaborative operation scheduling system based on swarm intelligence algorithm according to claim 7, characterized in that, The third unit includes: The first subunit is used to construct an adaptive bandwidth function based on the number of iterations, population diversity, and convergence state. The second subunit is used to dynamically adjust the bandwidth parameters of the harmony search based on the adaptive bandwidth function, and to determine the adjustment range based on the bandwidth parameters; The third subunit is used to determine whether to make an adjustment based on the second probability. If so, within the adjustment range, a random number is generated using the Kent map and the random number is added to the variable.

Images