An intelligent scheduling method based on improved BAS algorithm
By improving the BAS algorithm, combining the MES system and PLC data acquisition station, and optimizing the antenna distance and step length of the longhorn beetle foraging behavior simulation, the accuracy problem of smart factory scheduling in the absence of historical data was solved, and efficient and accurate production line task allocation was achieved.
Patent Information
- Application Number
- CN202211084703.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-09-06
- Publication Date
- 2025-09-23
- Estimated Expiration
- 2042-09-06
AI Technical Summary
Existing smart factory scheduling methods lack accuracy in the absence of historical data and rely on large amounts of historical data for training, leading to scheduling inaccuracies.
An improved BAS algorithm is used to pre-process order and equipment information, obtain data using the MES system and PLC data acquisition station, and combine the foraging behavior simulation of longhorn beetles in the BAS algorithm to optimize the antenna distance and step length, construct a fitness function, and perform optimal production line allocation.
It improves the accuracy and efficiency of smart factory production scheduling, reduces dependence on historical data, simplifies the model training process, and ensures accurate scheduling within different data ranges.
Smart Images
Figure CN115409397B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of control and scheduling of intelligent factories, and in particular to an intelligent scheduling method based on an improved BAS algorithm. Background Art
[0002] Smart manufacturing has gradually become a hot topic of social concern. As industrial manufacturing evolves towards digitalization, intelligence, and diversification, smart manufacturing technology has become a highly valued area within the industry. Against this backdrop, smart factories are effectively supporting the implementation of smart manufacturing.
[0003] Unlike traditional mechanized factories, smart factories, with their increasing digitalization and expanded production scale, generate vastly more data during the production process. This makes scheduling even more challenging, and the efficient and accurate scheduling of smart factory production lines is crucial. Scheduling issues and their associated knowledge, including how to rationally plan product processing sequences, efficiently process production data, and effectively integrate inventory, production, and transportation, are crucial challenges in modern factory management and production. Currently, domestic and international scholars have focused on leveraging datasets and past experience to allocate factory tasks, employing intelligent algorithms such as convolutional neural networks, multi-agent systems, and agents. While these approaches have demonstrated promising results for smart factory scheduling, they require extensive historical data for training to ensure model accuracy. Exceeding the dataset limits can lead to scheduling inaccuracies. Summary of the Invention
[0004] The purpose of the present invention is to overcome the deficiencies of the prior art and to provide an intelligent scheduling method based on an improved BAS algorithm, which can accurately and quickly complete the optimal scheduling allocation for smart factory production.
[0005] In order to achieve the above object, the technical solution adopted by the present invention is: an intelligent scheduling method based on an improved BAS algorithm, comprising the following steps:
[0006] (1) Obtain order information and equipment information;
[0007] (2) Marking order information and equipment information: Marking orders according to their urgency; classifying equipment according to the production of orders with different urgency levels;
[0008] (3) Allocate orders to production equipment according to the BAS algorithm, evaluate each allocation scheme and iterate to find the optimal scheduling allocation method.
[0009] Step (1) includes: using the MES system to obtain customer order information; using the PLC data acquisition station to receive equipment status information and environmental information.
[0010] The step (1) also includes pre-processing the order information and device information to eliminate abnormal data.
[0011] The step (2) comprises:
[0012] Classify order information into B1, B2, B3, etc. according to the urgency of the order;
[0013] Classify the production line equipment information into A1, A2, A3...;
[0014] Each type of production line equipment corresponds to the urgency of the orders that can be produced; and the higher the production urgency of the order, the higher the production line cost and the shorter the time.
[0015] The BAS algorithm is an improved BAS algorithm that updates the antenna distance and step size to schedule and allocate orders.
[0016] Improving the BAS algorithm includes the following steps:
[0017] Step a: When the beetle is at any position, the direction of its head is random. The normalization formula of the beetle's direction vector is:
[0018]
[0019] Where rand(D,1) represents the randomly generated scheduling instruction function, and D is the dimension of the space;
[0020] Step b: After obtaining the direction vector, define the positions of the beetle's left and right antennae as follows:
[0021]
[0022]
[0023] In the formula, the number of orders with different urgency levels is used as the initial position of the left and right antennae of the beetle l with x r ; Take the number of orders at each level at time m as the center of mass position x m , the difference in the number of orders of each level of the left and right antennas at time m is used as the distance d between the two antennas m .
[0024] Step c: Taking production line utilization efficiency and profitability as the objectives, construct a fitness function, which is as follows:
[0025]
[0026]
[0027]
[0028] In the formula, M represents the number of production lines, T represents the total time for producing and processing devices, C represents the total production income, and C cost is the total cost, λ1 and λ2 are the weights of production line utilization efficiency and rate of return in the objective function respectively; N1, N2, N3 are the number of production tasks when the production urgency is B1, B2, B3 respectively; O1, O2, O3 are the number of production task processes when the production urgency is B1, B2, B3 respectively; s, n, o are the serial numbers of production lines, production tasks and processes respectively, and x i (i=1,2,3) represents the execution time of the oth process of task n on production line s for order urgency i; T i (i=1, 2, 3) represents the latest end time of all processes for an order with order urgency i, i.e., the maximum completion time. C1, C2, and C3 represent the costs of production machines on different production lines, with smaller numbers indicating higher costs. C4 represents other comprehensive costs. The larger J(x) is, the higher the utilization efficiency and profitability of the smart factory production line.
[0029] Step d: Calculate the odor concentration J(x) of the antennae on both sides of the longhorn beetle based on the fitness function J(x) l ) and J(x r ), that is, the level of the production line, compares the fitness of the longhorn beetle's left and right antennae to determine the next move. To simulate the longhorn beetle's detection mechanism, the following position update iterative model is generated:
[0030]
[0031] Where sign is the sign function, and the number of new orders at each level at time m is used as the search step length δ m .
[0032] Step e: In practical applications, the distance difference d between the two whiskers m and the step length δ of the longicorn m The update rules are as follows:
[0033] d m =d e ·d m-1
[0034] δ m =δ e ·δ m-1
[0035] Where, d eis the attenuation coefficient of the distance between the two antennae, and the difference in the number of orders of each level of the antennae at time m-1 is taken as the distance d between the two antennae at the previous moment m-1 ,δ e is the attenuation coefficient of the step length, and the number of new orders at each level at time m-1 is used as the search step length δ of the previous moment m-1 .
[0036] Step f: To solve the problem that the BAS algorithm can only update one antenna distance and step length in each iteration, which easily causes the system to fall into local optimality, improvements are made on the basis of step e. According to the existing attenuation coefficient, 10 groups of order quantity differences at each level of antennas and the number of new orders at each level are generated each time, that is, 10 groups of antenna distance and step length data. By weighted calculation of the data set, the optimal antenna distance and step length are obtained, and the data are used as the optimal antenna distance and step length values for the current position.
[0037]
[0038]
[0039] Step g: Update the antenna distance and step size to determine whether the optimal or maximum number of iterations has been reached. If so, the iteration ends and returns the optimal allocation quantity, that is, the data of each type of production line equipment corresponding to each urgency order. Otherwise, return to step b.
[0040] The order information is sent to the customer order workstation, and the PLC data acquisition station is used to receive the equipment status information and environmental information. Each PLC data acquisition station and customer order workstation uploads the data to the data layer.
[0041] The data processing center obtains order classification and production line equipment information from the data layer and executes step (3) to output the number of orders of each urgency level that are allocated to each type of production line equipment.
[0042] The production scheduling and allocation results output by the data processing center are sent to the production line control system. After receiving the allocation command, the production line control system begins to schedule production as required and transmits the relevant production status to the display layer. The console of the display layer will display the relevant information of the perception layer, data layer and scheduling layer in the form of a monitoring platform, and provide corresponding interfaces for manual intervention.
[0043] The advantages of the present invention are: 1. Based on the pre-analysis of the perception layer data, duplicate or erroneous order information is eliminated, thereby improving the efficiency of the data layer in processing the perception layer data;
[0044] 2. The BAS algorithm is used to find the optimal number of production line allocation tasks, achieving target optimization simply and quickly. This avoids the problem of other intelligent algorithms using large amounts of data to train models, resulting in allocation results being heavily dependent on historical data.
[0045] 3. In view of the problem that the attenuation coefficient of the longhorn beetle is fixed when updating the antennae distance and step length, and only one set of data can be generated at a time, and it is impossible to determine whether the current data is the optimal antennae distance and step length, the present invention proposes an optimization algorithm for improving the antennae distance and step length of the longhorn beetle. By weighted processing of the antennae distance and step length, the optimal antennae distance and step length at the position can be obtained, thereby providing the optimal antennae distance and step length parameters for the optimization of the overall algorithm. BRIEF DESCRIPTION OF THE DRAWINGS
[0046] The following is a brief description of the contents and symbols in the drawings of the present invention:
[0047] Figure 1 This is the overall framework diagram of the smart factory scheduling system
[0048] Figure 2 This is the flowchart of the improved longicorn beard algorithm DETAILED DESCRIPTION
[0049] The specific implementation of the present invention will be further explained in detail below by describing the best embodiment with reference to the accompanying drawings.
[0050] In order to solve the technical problems existing in the production scheduling process of smart factories, the present invention provides a scheduling method based on an improved BAS algorithm. The present invention will be further described below with reference to the accompanying drawings and embodiments.
[0051] like Figure 1 As shown, a factory intelligent scheduling method mainly includes the following steps:
[0052] Step 1: Use smart sensors installed on the underlying mechanical equipment to obtain the equipment's operating status and environmental status, and use the MES system to obtain customer order information;
[0053] Step 2: Use the PLC data acquisition station to receive equipment status information and environmental information, and send the order information to the customer order workstation;
[0054] Step 3: Each PLC data acquisition station and customer order workstation undergoes data preprocessing, removes abnormal data, and sends the remaining data to the data layer;
[0055] Step 4: Use a clustering algorithm to record customer order information as B1, B2, and B3, where smaller numbers indicate more urgent customer orders. Record production line equipment information as A1, A2, and A3, where smaller numbers indicate higher-level production lines, higher costs, and shorter production times. Define A1 as being able to complete any task in B1, B2, and B3, A2 as being able to complete any task in B2 and B3, and A3 as being able to complete only tasks in B3. Count the number of tasks in each state and upload the data to the data processing center at the scheduling layer.
[0056] Step 5: The scheduling layer first calls the improved BAS algorithm to simulate the smart factory order allocation process by using the longhorn beetle's continuous search for food in three-dimensional space. The iterative results are then delivered to the corresponding production line.
[0057] Step 6: The three levels of production lines receive the allocation commands from the upper-level intelligent controller, start scheduling production as required, and pass the relevant production status to the display layer;
[0058] Step 7: The console of the display layer will display the relevant information of the perception layer, data layer, and scheduling layer in the form of a monitoring platform, so that relevant personnel can obtain factory information in a timely and accurate manner, and provide corresponding interfaces for manual intervention to achieve human-computer friendly interactive applications and experience.
[0059] like Figure 2 , for the improved BAS algorithm used in step 5, specifically including the following steps:
[0060] Step a: When the beetle is at any position, the direction of its head is random. The normalization formula of the beetle's direction vector is:
[0061]
[0062] Where rand(D,1) represents the randomly generated scheduling instruction function, and D is the dimension of the space.
[0063] Step b: After obtaining the direction vector, define the number of orders for the left and right antennae of the beetle to obtain different levels as follows:
[0064]
[0065]
[0066] In the formula, the number of orders with different urgency levels is used as the initial position of the left and right antennae of the beetle l with x r ; Take the number of orders at each level at time m as the center of mass position x m , the difference in the number of orders of each level of the left and right antennas at time m is used as the distance d between the two antennas m .
[0067] Step c: Taking production line utilization efficiency and profitability as the objectives, construct a fitness function, which is as follows:
[0068]
[0069]
[0070]
[0071] In the formula, M represents the number of production lines, T represents the total time for producing and processing devices, C represents the total production income, and C cost is the total cost, λ1 and λ2 are the weights of production line utilization efficiency and rate of return in the objective function respectively; N1, N2, N3 are the number of production tasks when the production urgency is B1, B2, B3 respectively; O1, O2, O3 are the number of production task processes when the production urgency is B1, B2, B3 respectively; s, n, o are the serial numbers of production lines, production tasks and processes respectively, and x i (i=1,2,3) represents the execution time of the oth process of task n on production line s for order urgency i; T i (i=1, 2, 3) represents the latest end time of all processes for an order with order urgency i, i.e., the maximum completion time. C1, C2, and C3 represent the costs of production machines on different production lines, with smaller numbers indicating higher costs. C4 represents other comprehensive costs. The larger J(x) is, the higher the utilization efficiency and profitability of the smart factory production line.
[0072] Step d: Calculate the odor concentration J(x) of the antennae on both sides of the longhorn beetle based on the fitness function J(x) l ) and J(x r ), that is, the level of the production line, compares the fitness of the longhorn beetle's left and right antennae to determine the next move. To simulate the longhorn beetle's detection mechanism, the following position update iterative model is generated:
[0073]
[0074] Where sign is the sign function, and the number of new orders at each level at time m is used as the search step length δ m .
[0075] Step e: In practical applications, the distance difference d between the two whiskers m and the step length δ of the longicorn m The update rules are as follows:
[0076] d m =d e ·d m-1
[0077] δm =δ e ·δ m-1
[0078] Where, d e is the attenuation coefficient of the distance between the two antennae, and the difference in the number of orders of each level of the antennae at time m-1 is taken as the distance d between the two antennae at the previous moment m-1 ,δ e is the attenuation coefficient of the step length, and the number of new orders at each level at time m-1 is used as the search step length δ of the previous moment m-1 .
[0079] Step f: To solve the problem that the BAS algorithm can only update one antenna distance and step length in each iteration, which easily causes the system to fall into local optimality, improvements are made on the basis of step e. According to the existing attenuation coefficient, 10 groups of order quantity differences at each level of antennas and the number of new orders at each level are generated each time, that is, 10 groups of antenna distance and step length data. By weighted calculation of the data set, the optimal antenna distance and step length are obtained, and the data are used as the optimal antenna distance and step length values for the current position.
[0080]
[0081]
[0082] Step g: Update the antenna distance and step size to determine whether the optimal or maximum number of iterations is reached. If so, the iteration ends and the optimal allocation quantity is returned. Otherwise, return to step b.
[0083] The intelligent factory scheduling system of the present invention can effectively improve the efficiency and accuracy of factory logistics scheduling and is simple and easy to implement. Specific advantages include:
[0084] 1. Based on the pre-analysis of the perception layer data, duplicate or erroneous order information is eliminated, which improves the efficiency of the data layer in processing the perception layer data;
[0085] 2. The BAS algorithm is used to find the optimal number of production line allocation tasks, achieving target optimization simply and quickly. This avoids the problem of other intelligent algorithms using large amounts of data to train models, resulting in allocation results being heavily dependent on historical data.
[0086] 3. In view of the problem that the attenuation coefficient of the longhorn beetle is fixed when updating the antennae distance and step length, and only one set of data can be generated at a time, and it is impossible to determine whether the current data is the optimal antennae distance and step length, the present invention proposes an optimization algorithm for improving the antennae distance and step length of the longhorn beetle. By weighted processing of the antennae distance and step length, the optimal antennae distance and step length at the position can be obtained, thereby providing the optimal antennae distance and step length parameters for the optimization of the overall algorithm.
[0087] Obviously, the specific implementation of the present invention is not limited to the above-mentioned methods. As long as various non-substantial improvements are made using the method concept and technical solution of the present invention, they are all within the scope of protection of the present invention.
Claims
1. An intelligent scheduling method based on an improved BAS algorithm, characterized by: The steps include: (1) Obtain order information and equipment information; (2) Marking order information and equipment information: Marking orders according to their urgency; classifying equipment according to the production of orders with different urgency levels; (3) Allocate orders to production equipment according to the BAS algorithm, evaluate each allocation scheme and iterate to find the optimal scheduling allocation method; Improving the BAS algorithm includes the following steps: Step a: When the beetle is at any position, the direction of its head is random. The normalization formula of the beetle's direction vector is: Where rand(D,1) represents the randomly generated scheduling instruction function, and D is the dimension of the space; Step b: After obtaining the direction vector, define the positions of the beetle's left and right antennae as follows: In the formula, the number of orders with different urgency levels is used as the initial position of the left and right antennae of the beetle l with x r ; Take the number of orders at each level at time m as the center of mass position x m , the difference in the number of orders of each level of the left and right antennas at time m is used as the distance d between the two antennas m ; Step c: Taking production line utilization efficiency and profitability as the objectives, construct a fitness function, which is as follows: In the formula, M represents the number of production lines, T represents the total time for producing and processing devices, C represents the total production income, and C cost is the total cost, λ1 and λ2 are the weights of production line utilization efficiency and rate of return in the objective function respectively; N1, N2, N3 are the number of production tasks when the production urgency is B1, B2, B3 respectively; O1, O2, O3 are the number of production task processes when the production urgency is B1, B2, B3 respectively; s, n, o are the serial numbers of production lines, production tasks and processes respectively, and x i (i=1,2,3) represents the execution time of the oth process of task n on production line s for order urgency i; T i (i=1,2,3) represents the latest end time of all processes for order urgency i, i.e., the maximum completion time; C1, C2, and C3 represent the costs of production machines on different production lines, respectively. Larger numbers indicate higher costs. C4 represents other comprehensive costs. In this case, the larger J(x) is, the higher the utilization efficiency and profitability of the smart factory production line. Step d: Calculate the odor concentration J(x) of the antennae on both sides of the longhorn beetle based on the fitness function J(x) l ) and J(x r ), that is, the level of the production line, compares the fitness of the left and right antennae of the longhorn beetle and determines the next move position; Position update iterative model: Where sign is the sign function, and the number of new orders at each level at time m is used as the search step length δ m ; Step e: The distance difference d between the two whiskers m and the step length δ of the longicorn m The update rules are as follows: d m =d e ·d m-1 d m =d e ·d m-1 Where, d e is the attenuation coefficient of the distance between the two antennae, and the difference in the number of orders of each level of the antennae at time m-1 is taken as the distance d between the two antennae at the previous moment m-1 ,δ e is the attenuation coefficient of the step length, and the number of new orders at each level at time m-1 is used as the search step length δ of the previous moment m-1 ; Step f: Obtain the optimal antenna distance and step length by weighted calculation of the data set, and use the data as the optimal antenna distance and step length value for the current position; Step g: Update the antenna distance and step size to determine whether the optimal or maximum number of iterations has been reached. If so, the iteration ends and returns the optimal allocation quantity, that is, the data of each type of production line equipment corresponding to each urgency order. Otherwise, return to step b.
2. The intelligent scheduling method based on the improved BAS algorithm according to claim 1, characterized in that: Step (1) includes: using the MES system to obtain customer order information; using the PLC data acquisition station to receive equipment status information and environmental information.
3. The intelligent scheduling method based on the improved BAS algorithm according to claim 2, characterized in that: The step (1) also includes pre-processing the order information and device information to eliminate abnormal data.
4. The intelligent scheduling method based on the improved BAS algorithm according to claim 1, characterized in that: The step (2) comprises: Classify order information into B1, B2, B3, etc. according to the urgency of the order; Classify the production line equipment information into A1, A2, A3...; Each type of production line equipment corresponds to the urgency of the orders that can be produced; and the higher the production urgency of the order, the higher the production line cost and the shorter the time.
5. The intelligent scheduling method based on the improved BAS algorithm according to claim 1, characterized in that: The order information is sent to the customer order workstation, and the PLC data acquisition station is used to receive the equipment status information and environmental information. Each PLC data acquisition station and customer order workstation uploads the data to the data layer.
6. The intelligent scheduling method based on the improved BAS algorithm according to claim 5, characterized in that: The data processing center obtains order classification and production line equipment information from the data layer and executes step (3) to output the number of orders of each urgency level that are allocated to each type of production line equipment.
7. The intelligent scheduling method based on the improved BAS algorithm according to claim 6, characterized in that: The production scheduling and allocation results output by the data processing center are sent to the production line control system. After receiving the allocation command, the production line control system begins to schedule production as required and transmits the relevant production status to the display layer. The console of the display layer will display the relevant information of the perception layer, data layer and scheduling layer in the form of a monitoring platform, and provide corresponding interfaces for manual intervention.