A multi-resource collaborative intelligent workshop equipment configuration optimization method

Abstract

The present invention discloses a multi-resource collaborative intelligent workshop equipment configuration optimization method and system, specifically including: S1, constructing a simulation system that simulates the actual production environment; S2, establishing a mathematical model for optimizing equipment configuration and setting performance index constraint parameter conditions; S3, using the gray wolf optimization algorithm to calculate the mathematical model, using the simulation system in the gray wolf optimization algorithm to obtain performance indicators, updating the population and position of the gray wolf optimization algorithm according to the performance indicators, and obtaining the iterated gray wolf optimization algorithm after a set number of iterations; S4, obtaining the optimized equipment configuration according to the solution of the iterated gray wolf optimization algorithm. Compared with the traditional technology, this method uses the gray wolf optimization algorithm to solve the mathematical model, uses the simulation system to solve the fitness value and updates the iterative gray wolf optimization algorithm, and realizes the minimization of the investment cost of workshop equipment under the premise of meeting the maximum production cycle and the constraints of the order delivery time and the minimum system output rate.

Description

Technical Field

[0001] The present invention relates to the field of manufacturing system planning and design, and more specifically, to a multi-resource collaborative intelligent workshop equipment configuration optimization method. Background Art

[0002] Developing intelligent manufacturing in the traditional manufacturing industry is an important support for promoting the quality and efficiency of the traditional manufacturing industry and promoting the transformation of China's manufacturing industry from large to strong. Personalized customization is an important feature of intelligent manufacturing. With the continuous advancement of the intelligent manufacturing strategy, the intelligent transformation of factories has become the main way for the current manufacturing industry to promote the transformation of intelligent manufacturing. The intelligent workshop is an important link in realizing the smart factory and then realizing intelligent manufacturing. For the planning and design of the intelligent workshop, since highly automated and intelligent production equipment is expensive, how to configure these equipment resources in the intelligent workshop to ensure the expected production capacity at the lowest cost and deliver orders on time and quickly is a problem that needs to be solved in the planning and design of customized production workshops.

[0003] In the process of customized production, the arrival time of workpieces, process path and processing time, material transportation time, and operation time of industrial robots are all uncertain. Therefore, it is impossible to solve the resource allocation optimization problem of such production workshops through traditional deterministic mathematical programming models, which brings great challenges to the planning and design of production workshops. Multi-resource collaborative intelligent workshops usually require the collaborative operation of multiple types of equipment resources to make the system operate normally. For example, the handling process, loading and unloading process and processing process of workpieces need to be carried out under the condition that multiple resources such as AGV, robots and processing machine tools are available at the same time, that is, resource collaborative constraints. Therefore, the planning and design optimization of multi-resource collaborative intelligent workshops needs to consider the collaborative relationship between equipment with the minimum equipment configuration cost. Grey Wolf Optimizer (GWO) is a swarm intelligence optimization algorithm proposed by Mirjalili et al. in 2014. It is an optimization search algorithm developed based on the inspiration of gray wolf packs when they hunt prey. The Grey Wolf Optimizer has the characteristics of strong optimization ability, few parameters and easy implementation.

[0004] The prior art discloses an AGV optimization scheduling method based on simulated annealing particle swarm, which uses a particle swarm algorithm to initialize the population, determine the fitness function, calculate the fitness value of the particles, and combines the simulated annealing algorithm to update the speed and position of the particles, determine the fitness value of the new particles, and finally obtain the optimization result. The defect of this solution is that it does not perform calculations based on the actual production environment, cannot truly characterize the equipment configuration system, does not solve the problem of unstable efficiency of the production system, and cannot minimize costs.

[0005] To this end, in combination with the above characteristics and defects in the prior art, this application proposes a multi-resource collaborative intelligent workshop equipment configuration optimization method and system. Summary of the invention

[0006] The present invention provides a multi-resource collaborative intelligent workshop equipment configuration optimization method and system, which can minimize the investment cost of AGV and manipulator under the premise of meeting the maximum production cycle, with the order delivery time and the minimum system output rate as constraints.

[0007] The primary purpose of the present invention is to solve the above technical problems. The technical solutions of the present invention are as follows:

[0008] The first aspect of the present invention provides a multi-resource collaborative intelligent workshop equipment configuration optimization method, the method comprising the following steps:

[0009] S1. Build a simulation system that simulates the actual production environment.

[0010] S2. Establish a mathematical model for optimizing the number and types of equipment in the simulation system and set performance indicator constraint parameter conditions.

[0011] S3. Use the gray wolf optimization algorithm to calculate the mathematical model, use the simulation system in the gray wolf optimization algorithm to obtain performance indicators, update the population and position of the gray wolf optimization algorithm according to the optimal performance indicators, and after a set number of iterations, obtain the iterated gray wolf optimization algorithm.

[0012] S4. Obtain optimized equipment configuration based on the solution of the mathematical model using the iterated Grey Wolf Optimization Algorithm.

[0013] Furthermore, the step S1 includes:

[0014] S11. Construct a simulation system according to the actual production environment, wherein the simulation system includes a task pool, a raw material warehouse, a processing unit, a finished product warehouse and a track. The processing unit includes a front buffer, a processing center and a rear buffer connected in sequence. The track is the operating path of the AGV and the robot. The task pool issues tasks to the AGV and the robot.

[0015] S12. The working steps of the simulation system are as follows: AGV transports the raw materials from the raw material warehouse to the front buffer zone of each processing unit, the robot loads the raw materials from the front buffer zone to the processing center, the processing center processes the raw materials, the robot unloads the finished products processed by the processing center to the rear buffer zone, and the AGV transports the finished products stored in the rear buffer zone to the finished product warehouse.

[0016] S13, the simulation system follows the following rules: the arrival of workpieces in the system follows the Poisson distribution process, the arrival of each workpiece is an independent event, and the mean arrival rate is λ; the system determines whether the raw material warehouse is full before the workpiece arrives. If the raw material warehouse is full, the system will refuse the workpiece to enter; each processing unit has only one processing device, which can only allow a single workpiece to be processed, and the processing is subject to the parameter R t Negative exponential distribution; the processing equipment needs a robot to load or unload materials before processing, and the loading and unloading time of each time is independent of each other. The loading and unloading time is subject to the parameter R m The negative exponential distribution of the system processing unit follows the first-come-first-served rule, and the blocking of each processing unit obeys the post-blocking mechanism. The front and rear buffers of each processing unit are both limited and have the same capacity. When the AGV loads workpieces in the raw material warehouse, it follows the load balancing principle to decide which processing center to go to. In order to ensure the system output capacity, the time it takes for the equipment to transport the workpiece is not considered, and the overall system output capacity is determined by the smallest process.

[0017] The parameters of the simulation system are: robot type i, i = 1, 2, 3, ...,; the unit price of the i-th robot v i (10,000 yuan); the number of manipulators of type i x i ; Various types of manipulator quantity configuration vector X, X = {x i}; AGV type j, j = 1, 2, 3, ...,; AGV unit price μ of the i-th type j (10,000 yuan); the number of AGVs of type i y j ; Various AGV quantity configuration vector Y, Y = {y j}; total equipment investment cost Q (10,000 yuan); average system output rate θ (pieces / min); average system production cycle Γ (min); mathematical expectation of random function E{·}; average speed V of AGV running a circle without waiting; AGV capacity C; system output rate l; raw material arrival rate λ (pieces / min); buffer capacity B1 before processing unit; buffer capacity B2 after processing unit; random element ξ.

[0018] Furthermore, the track is the running path of the AGV and the robot, wherein the running path of the AGV is a bidirectional single-loop path, and the running path of the robot is a bidirectional multiple-loop path.

[0019] Among them, due to the limitation of buffer space capacity and delivery deadline, the front and rear buffers in the processing unit are limited buffers. Due to the large size and heavy weight of the workpiece, a robot is required to load and unload.

[0020] Furthermore, the AGV loading quantity is determined by the AGV capacity, the existing workpieces in the raw material warehouse and the remaining capacity of the destination, and its mathematical expression is:

[0021]

[0022] Among them, V c is the number of workpieces loaded by AGV, n is the number of processing units, There are W workpieces in the raw material warehouse. is the remaining capacity W of the front buffer of the kth processing unit.

[0023] The mathematical expression of the overall system output capability is:

[0024]

[0025] Where l is the system output rate, x i is the number of configurations of the i-th type of manipulator, R t is the machining rate of the machining center, R m The working speed of the robot for loading and unloading.

[0026] Furthermore, the workpieces are generated according to Poisson distribution and are of the same type and enter the raw material warehouse. After the AGV accepts the handling task, it goes to the raw material warehouse to carry V c The AGV moves the workpieces to the front buffer with the least workpieces according to the load balancing principle.

[0027] Among them, there are two types of tasks that AGV can handle after the transportation is completed, namely:

[0028] (1) When That is, the urgency of finished product delivery is higher than that of raw material handling, so the AGV goes to the rear buffer with the most finished products in the processing unit to deliver the finished products, where r is the fitness value of the AGV reverse driving transfer probability, There are W workpieces in the back buffer of the kth processing unit.

[0029] (2) When That is, the urgency of shipping finished products out of the warehouse is lower than that of transporting raw materials. The AGV drives in the opposite direction and returns to the raw material warehouse to transport the workpiece to the front buffer zone.

[0030] The purpose of adding the fitness value r is to prevent the AGV from frequently reversing and causing the next AGV to wait.

[0031] Furthermore, when the machining center needs to load or unload materials, it will send an application to the task pool; the robot follows the first-come-first-served rule of the task pool and selects the shortest circular track to handle the loading or unloading task.

[0032] Furthermore, the mathematical model for optimizing the number and types of devices in the simulation system described in step S2 is expressed in the following mathematical form:

[0033]

[0034] Among them, (X * , Y * ) is the optimal vector set of the manipulator and AGV, which is used to minimize the total cost; the performance index constraint parameter conditions set are:

[0035] E{Γ(X, Y:ξ)}≤Γ max

[0036] E{θ(X, Y:ξ)}≥θ min

[0037] X, Y∈N +

[0038] Among them, E{Γ(X, Y:ξ)} is the average production cycle, and the conditions are set so that the simulation production cycle Γ max is smaller than the actual average production cycle of the system, which is used to ensure that the delivery deadline is met; E{θ(X, Y:ξ)} is the average output rate setting condition so that the simulation average output rate θ min Greater than the average output rate of the system, used to ensure that the system has sufficient capacity; X, Y∈N + Indicates that the X vector and the Y vector are positive integers.

[0039] Among them, since E{Γ(X, Y:ξ)} and E{θ(X, Y:ξ)} cannot be expressed in mathematical closed form with the robot X and the number of AGVs Y, the typical nonlinear integer programming method cannot solve the problem. Therefore, the simulation system is used to obtain the two performance indicators of average production cycle and system output rate, and then the gray wolf algorithm is embedded to calculate and optimize the equipment configuration to minimize the equipment configuration cost.

[0040] Furthermore, the step S3 is specifically as follows:

[0041] S31. Initialize the wolf pack, use integer coding to randomly generate the initial solution, divide the feasible solution sequence into two parts: the number of the i-th type of manipulator and the number of the j-th type of AGV, define the gray wolf individual as Z(i), the number of each type of manipulator as X, the number of each type of AGV as Y, and the wolf pack as

[0042] S32. Import the encoded wolf individuals into the simulation model to solve the fitness value. The parameters of each wolf individual are run in the simulation system for several times and then the average value is taken to obtain two performance indicators, namely the system output rate θ and the average production cycle T, and the fitness value H is obtained by solving. A hierarchy of gray wolves is established, which is divided into four levels from high to low: α, β, δ, ω. The fitness values ​​H of the wolf individuals are sorted from small to large. The first three wolves with the smallest fitness are α, β, and δ wolves, and the rest are ω wolves.

[0043] Among them, the smaller the fitness value, the better the performance index obtained.

[0044] S33. Update the parameters of the gray wolf optimization algorithm according to the number of iterations, and update the gray wolf population; update the convergence factor a, coefficient vectors A and C that decrease linearly with the number of iterations according to the current number of iterations; adopt a population update mechanism that eliminates the last N individuals after each iteration.

[0045] S34. Update the position of the gray wolf. The gray wolf individual adjusts its own position according to the positions of α, β, and δ, and introduces a dynamic proportional weight method to reduce the probability of the algorithm falling into the local optimum and complete an iteration.

[0046] Among them, the principle of the gray wolf individuals adjusting their own positions according to the positions of α, β, and δ is that when the gray wolf finds the target, the α wolf will command the β wolf and the δ wolf to surround the target.

[0047] S35, repeat steps S32 and S33, and save the current gray wolf optimization algorithm after reaching the set maximum number of iterations.

[0048] Among them, as the number of iterations increases, the alpha , beta , and delta alphas dynamically adjust the positions of the wolf pack, effectively increasing the probability of finding the global optimal solution.

[0049] Furthermore, the mathematical expression of step S3 is:

[0050] Z(i)=[X|Y]=[x1,x2,…,x n |y1,y2,…,y n ]

[0051] X={x i}

[0052] Y={y j}

[0053]

[0054] Among them, lnum represents the number of gray wolves in the wolf pack.

[0055] Furthermore, the mathematical expression of the fitness value H of the i-th gray wolf in S32 is:

[0056] H i =T i -θ i

[0057] Furthermore, the mathematical expression of the parameters of the gray wolf optimization algorithm and the gray wolf population described in S33 is:

[0058] A=2a·r1-a

[0059] C=2·r2

[0060]

[0061] N=0.3*lnum

[0062] Among them, t represents the number of current iterations, r1 and r2 are random elements in the value range [0, 1], and a decreases from 2 to 0.

[0063] Among them, if the value of N is too large, the algorithm is difficult to converge; if the value of N is too small, the number of new solutions generated is small, which reduces the algorithm's optimization ability. Using N = 0.3*lnum can not only ensure population diversity, but also make it easy for the algorithm to converge and optimize.

[0064] Furthermore, the mathematical expression of the gray wolf individual in S34 adjusting its own position according to the positions of α, β, and δ is:

[0065] D α =C1·Z α -Z

[0066] D β =C2·Z β -Z

[0067] D δ =C3·Z δ -Z

[0068] Z1=Z α -A1·(D α )

[0069] Z2=Zx-A2·(D β )

[0070] Z3=Z δ -A3·(D δ )

[0071] Among them, D α , D β , D δ Represents the distance vector between individuals α, β, δ and the gray wolf ω, Z α , Z β , Z δ Represents the current position of α, β, δ wolves; C1, C2, C3 represent vectors with random values ​​in the range [0, 1], Z is the current position of the gray wolf, and Z1, Z2, Z3 are the position vectors of the gray wolf.

[0072] Furthermore, the mathematical expression of the dynamic proportional weight in S34 is:

[0073]

[0074]

[0075]

[0076]

[0077]

[0078]

[0079] Z(t+1)=W1×σ1+W2×σ2+W3×σ3

[0080] Among them, σ1, σ2, σ3 are the proportional weights of the position vector; W1, W2, W3 are the learning proportions of the gray wolf for the α, β, and δ wolf positions respectively, and Z(t+1) is the updated gray wolf position.

[0081] Among them, the purpose of introducing the dynamic proportional weight method is to solve the problem that in the traditional GWO, the gray wolf position update adopts the average position of the three wolves α, β, and δ, and the proportional weight is always equal and unchanged, which leads to the problem that when the α wolf in the algorithm is the local optimal, the other gray wolves will continue to approach the α wolf and fall into the local optimal state, which can significantly reduce the probability of the algorithm falling into the local optimal state.

[0082] Furthermore, the step S4 is specifically as follows: the iterative grey wolf algorithm is used to solve the mathematical model of the number and type of equipment in the optimization simulation system, and the obtained grey wolf position is the optimal solution for the number and type of equipment.

[0083] A second aspect of the present invention provides a multi-resource collaborative intelligent workshop equipment configuration optimization system, including a memory and a processor, wherein the memory includes a multi-resource collaborative intelligent workshop equipment configuration optimization program, and when the multi-resource collaborative intelligent workshop equipment configuration optimization program is executed by the processor, the following steps are implemented:

[0084] S1. Build a simulation system that simulates the actual production environment.

[0085] S2. Establish a mathematical model for optimizing the number and types of equipment in the simulation system and set performance indicator constraint parameter conditions.

[0086] S3. Use the gray wolf optimization algorithm to calculate the mathematical model, use the simulation system in the gray wolf optimization algorithm to obtain performance indicators, update the population and position of the gray wolf optimization algorithm according to the optimal performance indicators, and after a set number of iterations, obtain the iterated gray wolf optimization algorithm.

[0087] S4. Obtain optimized equipment configuration based on the solution of the mathematical model using the iterated Grey Wolf Optimization Algorithm.

[0088] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0089] The present invention provides a multi-resource collaborative intelligent workshop equipment configuration optimization method, which uses the Gray Wolf Optimization Algorithm to solve the mathematical model of equipment configuration optimization, and uses the simulation system to solve the fitness value and update the iterative Gray Wolf Optimization Algorithm, thereby minimizing the investment cost of workshop equipment under the premise of meeting the maximum production cycle and with the order delivery time and the minimum system output rate as constraints. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] Figure 1 The present invention is a flowchart of a multi-resource collaborative intelligent workshop equipment configuration optimization method.

[0091] Figure 2 It is a schematic diagram of the simulation system in the present invention.

[0092] Figure 3 It is a schematic diagram of a group of processing units in the present invention.

[0093] Figure 4 The figure is a flow chart of the gray wolf optimization algorithm embedded in the simulation system of the present invention.

[0094] Figure 5 This is a schematic diagram of initializing a wolf pack in one embodiment of the present invention.

[0095] Figure 6 It is a schematic diagram of a multi-resource collaborative intelligent workshop equipment configuration optimization system of the present invention. DETAILED DESCRIPTION

[0096] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0097] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0098] Example 1

[0099] like Figure 1-Figure 4 As shown, the present invention provides a multi-resource collaborative intelligent workshop equipment configuration optimization method, the method comprising the following steps:

[0100] S1. Build a simulation system that simulates the actual production environment.

[0101] S2. Establish a mathematical model for optimizing the number and types of equipment in the simulation system and set performance indicator constraint parameter conditions.

[0102] S3. Use the gray wolf optimization algorithm to calculate the mathematical model, use the simulation system in the gray wolf optimization algorithm to obtain performance indicators, update the population and position of the gray wolf optimization algorithm according to the optimal performance indicators, and after a set number of iterations, obtain the iterated gray wolf optimization algorithm.

[0103] S4. Obtain optimized equipment configuration based on the solution of the mathematical model using the iterated Grey Wolf Optimization Algorithm.

[0104] Furthermore, the step S1 includes:

[0105] S11. Construct a simulation system according to the actual production environment, wherein the simulation system includes a task pool, a raw material warehouse, a processing unit, a finished product warehouse and a track. The processing unit includes a front buffer, a processing center and a rear buffer connected in sequence. The track is the operating path of the AGV and the robot. The task pool issues tasks to the AGV and the robot.

[0106] S12, such as Figure 2 As shown, the working steps of the simulation system are: AGV transports the raw materials from the raw material warehouse to the front buffer zone of each processing unit, the robot loads the raw materials from the front buffer zone to the processing center, the processing center processes the raw materials, the robot unloads the finished products processed by the processing center to the rear buffer zone, and the AGV transports the finished products stored in the rear buffer zone to the finished product warehouse.

[0107] S13, the simulation system follows the following rules: the arrival of workpieces in the system follows the Poisson distribution process, the arrival of each workpiece is an independent event, and the mean arrival rate is λ; the system determines whether the raw material warehouse is full before the workpiece arrives. If the raw material warehouse is full, the system will refuse the workpiece to enter; each processing unit has only one processing device, which can only allow a single workpiece to be processed, and the processing is subject to the parameter R t Negative exponential distribution; the processing equipment needs a robot to load or unload materials before processing, and the loading and unloading time of each time is independent of each other. The loading and unloading time is subject to the parameter R m The negative exponential distribution of the system processing unit follows the first-come-first-served rule, and the blocking of each processing unit obeys the post-blocking mechanism. The front and rear buffers of each processing unit are both limited and have the same capacity. When the AGV loads workpieces in the raw material warehouse, it follows the load balancing principle to decide which processing center to go to. In order to ensure the system output capacity, the time it takes for the equipment to transport the workpiece is not considered, and the overall system output capacity is determined by the smallest process.

[0108] The parameters of the simulation system are: robot type i, i = 1, 2, 3, ...,; the unit price of the i-th robot vi (10,000 yuan); the number of manipulators of type i x i ; The number of various manipulators is configured as vector X, X = {x i}; AGV type j, j = 1, 2, 3, ...,; AGV unit price μ of the i-th type j (10,000 yuan); the number of AGVs of type i y j ; Various AGV quantity configuration vector Y, Y = {y j}; total equipment investment cost Q (10,000 yuan); average system output rate θ (pieces / min); average system production cycle Γ (min); mathematical expectation of random function E{·}; average speed V of AGV running a circle without waiting; AGV capacity C; system output rate l; raw material arrival rate λ (pieces / min); buffer capacity B1 before processing unit; buffer capacity B2 after processing unit; random element ξ.

[0109] Furthermore, the track is the running path of the AGV and the robot, wherein the running path of the AGV is a bidirectional single-loop path, and the running path of the robot is a bidirectional multiple-loop path.

[0110] Among them, due to the limitation of buffer space capacity and delivery deadline, the front and rear buffers in the processing unit are limited buffers. Due to the large size and heavy weight of the workpiece, a robot is required to load and unload.

[0111] Furthermore, the AGV loading quantity is determined by the AGV capacity, the existing workpieces in the raw material warehouse and the remaining capacity of the destination, and its mathematical expression is:

[0112]

[0113] Among them, V c is the number of workpieces loaded by AGV, n is the number of processing units, There are W workpieces in the raw material warehouse. is the remaining capacity W of the front buffer of the kth processing unit.

[0114] The mathematical expression of the overall system output capability is:

[0115]

[0116] Where l is the system output rate, x i is the number of configurations of the i-th type of manipulator, R t is the machining rate of the machining center, R m The working speed of the robot for loading and unloading.

[0117] Furthermore, the workpieces are generated according to Poisson distribution and are of the same type and enter the raw material warehouse. After the AGV accepts the handling task, it goes to the raw material warehouse to carry V c The AGV moves the workpieces to the front buffer with the least workpieces according to the load balancing principle.

[0118] Among them, there are two types of tasks that AGV can handle after the transportation is completed, namely:

[0119] (1) When That is, the urgency of finished product delivery is higher than that of raw material handling, so the AGV goes to the rear buffer with the most finished products in the processing unit to deliver the finished products, where r is the fitness value of the AGV reverse driving transfer probability, There are W workpieces in the back buffer of the kth processing unit.

[0120] (2) When That is, the urgency of shipping finished products out of the warehouse is lower than that of transporting raw materials. The AGV drives in the opposite direction and returns to the raw material warehouse to transport the workpiece to the front buffer zone.

[0121] The purpose of adding the fitness value r is to prevent the AGV from frequently reversing and causing the next AGV to wait.

[0122] Furthermore, when the machining center needs to load or unload materials, it will send an application to the task pool; the robot follows the first-come-first-served rule of the task pool and selects the shortest circular track to handle the loading or unloading task.

[0123] Furthermore, the mathematical model for optimizing the number and types of devices in the simulation system described in step S2 is expressed in the following mathematical form:

[0124]

[0125] Among them, (X * , Y * ) is the optimal vector set of the manipulator and AGV, which is used to minimize the total cost; the performance index constraint parameter conditions set are:

[0126] E{Γ(X, Y:ξ)}≤Γ max

[0127] E{θ(X, Y:ξ)}≥θ min

[0128] X, Y∈N +

[0129] Among them, E{Γ(X, Y:ξ)} is the average production cycle, and the conditions are set so that the simulation production cycle Γ maxis smaller than the actual average production cycle of the system, which is used to ensure that the delivery deadline is met; E{θ(X, Y:ξ)} is the average output rate setting condition so that the simulation average output rate θ min Greater than the average output rate of the system, used to ensure that the system has sufficient capacity; X, Y∈N + Indicates that the X vector and the Y vector are positive integers.

[0130] Among them, since E{Γ(X, Y:ξ)} and E{θ(X, Y:ξ)} cannot be expressed in mathematical closed form with the robot X and the number of AGVs Y, the typical nonlinear integer programming method cannot solve the problem. Therefore, the simulation system is used to obtain the two performance indicators of average production cycle and system output rate, and then the gray wolf algorithm is embedded to calculate and optimize the equipment configuration to minimize the equipment configuration cost.

[0131] Further, such as Figure 4 As shown, the step S3 is specifically as follows:

[0132] S31. Initialize the wolf pack, use integer coding to randomly generate the initial solution, divide the feasible solution sequence into two parts: the number of the i-th type of manipulator and the number of the j-th type of AGV, define the gray wolf individual as Z(i), the number of each type of manipulator as X, the number of each type of AGV as Y, and the wolf pack as

[0133] S32. Import the encoded wolf individuals into the simulation model to solve the fitness value. The parameters of each wolf individual are run in the simulation system for several times and then the average value is taken to obtain two performance indicators, namely the system output rate θ and the average production cycle T, and the fitness value H is obtained by solving. A hierarchy of gray wolves is established, which is divided into four levels from high to low: α, β, δ, ω. The fitness values ​​H of the wolf individuals are sorted from small to large. The first three wolves with the smallest fitness are α, β, and δ wolves, and the rest are ω wolves.

[0134] Among them, the smaller the fitness value, the better the performance index obtained.

[0135] S33. Update the parameters of the gray wolf optimization algorithm according to the number of iterations, and update the gray wolf population; update the convergence factor a, coefficient vectors A and C that decrease linearly with the number of iterations according to the current number of iterations; adopt a population update mechanism that eliminates the last N individuals after each iteration.

[0136] S34. Update the position of the gray wolf. The gray wolf individual adjusts its own position according to the positions of α, β, and δ, and introduces a dynamic proportional weight method to reduce the probability of the algorithm falling into the local optimum and complete an iteration.

[0137] Among them, the principle of the gray wolf individuals adjusting their own positions according to the positions of α, β, and δ is that when the gray wolf finds the target, the α wolf will command the β wolf and the δ wolf to surround the target.

[0138] S35, repeat steps S32 and S33, and save the current gray wolf optimization algorithm after reaching the set maximum number of iterations.

[0139] Among them, as the number of iterations increases, the alpha , beta , and delta alphas dynamically adjust the positions of the wolf pack, effectively increasing the probability of finding the global optimal solution.

[0140] Furthermore, the mathematical expression of step S3 is:

[0141] Z(i)=[X|Y]=[x1,x2,…,x n |y1,y2,…,y n ]

[0142] X={x i}

[0143] Y={y j}

[0144]

[0145] Among them, lnum represents the number of gray wolves in the wolf pack.

[0146] Furthermore, the mathematical expression of the fitness value H of the i-th gray wolf in S32 is:

[0147] H i =T i -θ i

[0148] Furthermore, the mathematical expression of the parameters of the gray wolf optimization algorithm and the gray wolf population described in S33 is:

[0149] A=2a·r1-a

[0150] C=2·r2

[0151]

[0152] N=0.3*lnum

[0153] Among them, t represents the number of current iterations, r1 and r2 are random elements in the value range [0, 1], and a decreases from 2 to 0.

[0154] Among them, if the value of N is too large, the algorithm is difficult to converge; if the value of N is too small, the number of new solutions generated is small, which reduces the algorithm's optimization ability. Using N = 0.3*lnum can not only ensure population diversity, but also make it easy for the algorithm to converge and optimize.

[0155] Furthermore, the mathematical expression of the gray wolf individual in S34 adjusting its own position according to the positions of α, β, and δ is:

[0156] D α =C1·Z α -Z

[0157] D β =C2·Z β -Z

[0158] D δ =C3·Z δ -Z

[0159] Z1=Z α -A1·(D α )

[0160] Z2=Z β -A2·(D β )

[0161] Z3=Z δ -A3·(D δ )

[0162] Among them, D α , D β , D δ Represents the distance vector between individuals α, β, δ and the gray wolf ω, Z α , Z β , Z δ Represents the current position of α, β, δ wolves; C1, C2, C3 represent vectors with random values ​​in the range [0, 1], Z is the current position of the gray wolf, and Z1, Z2, Z3 are the position vectors of the gray wolf.

[0163] Furthermore, the mathematical expression of the dynamic proportional weight in S34 is:

[0164]

[0165]

[0166]

[0167]

[0168]

[0169]

[0170] Z(t+1)=W1×σ1+W2×σ2+W3×σ3

[0171] Among them, σ1, σ2, σ3 are the proportional weights of the position vector; W1, W2, W3 are the learning proportions of the gray wolf for the α, β, and δ wolf positions respectively, and Z(t+1) is the updated gray wolf position.

[0172] Among them, the purpose of introducing the dynamic proportional weight method is to solve the problem that in the traditional GWO, the gray wolf position update adopts the average position of the three wolves α, β, and δ, and the proportional weight is always equal and unchanged, which leads to the problem that when the α wolf in the algorithm is the local optimal, the other gray wolves will continue to approach the α wolf and fall into the local optimal state, which can significantly reduce the probability of the algorithm falling into the local optimal state.

[0173] Furthermore, the step S4 is specifically as follows: the iterative grey wolf algorithm is used to solve the mathematical model of the number and type of equipment in the optimization simulation system, and the obtained grey wolf position is the optimal solution for the number and type of equipment.

[0174] Example 2

[0175] Based on the above embodiment 1, combined with Figure 5 ,This embodiment describes in detail the process of initializing the wolf pack.

[0176] In a specific embodiment, Figure 5 As shown, integer coding is used to randomly generate the initial solution, encoding the 4-digit integer 2153 as the wolf pack individuals, where the first two digits encode 21 as the number configuration of each type of manipulator, and the last two digits encode 53 as the number configuration of each type of AGV.

[0177] In a specific embodiment, when the encoded wolf pack individuals are imported into the simulation model to solve the fitness value, each individual parameter is averaged after running the simulation model 1000 times and simulating 20 times, and two performance indicators, system output rate θ and average production cycle T, are obtained, and the fitness value H is solved.

[0178] Example 3

[0179] like Figure 6 As shown, the present invention also provides a multi-resource collaborative intelligent workshop equipment configuration optimization system, including a memory and a processor, wherein the memory includes a multi-resource collaborative intelligent workshop equipment configuration optimization program, and the multi-resource collaborative intelligent workshop equipment configuration optimization program is executed by the processor to implement the following steps:

[0180] S1. Build a simulation system that simulates the actual production environment.

[0181] S2. Establish a mathematical model for optimizing the number and types of equipment in the simulation system and set performance indicator constraint parameter conditions.

[0182] S3. Use the gray wolf optimization algorithm to calculate the mathematical model, use the simulation system in the gray wolf optimization algorithm to obtain performance indicators, update the population and position of the gray wolf optimization algorithm according to the optimal performance indicators, and after a set number of iterations, obtain the iterated gray wolf optimization algorithm.

[0183] S4. Obtain optimized equipment configuration based on the solution of the mathematical model using the iterated Grey Wolf Optimization Algorithm.

[0184] Furthermore, the step S1 includes:

[0185] S11. Construct a simulation system according to the actual production environment, wherein the simulation system includes a task pool, a raw material warehouse, a processing unit, a finished product warehouse and a track. The processing unit includes a front buffer, a processing center and a rear buffer connected in sequence. The track is the operating path of the AGV and the robot. The task pool issues tasks to the AGV and the robot.

[0186] S12. The working steps of the simulation system are as follows: AGV transports the raw materials from the raw material warehouse to the front buffer zone of each processing unit, the robot loads the raw materials from the front buffer zone to the processing center, the processing center processes the raw materials, the robot unloads the finished products processed by the processing center to the rear buffer zone, and the AGV transports the finished products stored in the rear buffer zone to the finished product warehouse.

[0187] S13, the simulation system follows the following rules: the arrival of workpieces in the system follows the Poisson distribution process, the arrival of each workpiece is an independent event, and the mean arrival rate is λ; the system determines whether the raw material warehouse is full before the workpiece arrives. If the raw material warehouse is full, the system will refuse the workpiece to enter; each processing unit has only one processing device, which can only allow a single workpiece to be processed, and the processing is subject to the parameter R t Negative exponential distribution; the processing equipment needs a robot to load or unload materials before processing, and the loading and unloading time of each time is independent of each other. The loading and unloading time is subject to the parameter R m The negative exponential distribution of the system processing unit follows the first-come-first-served rule, and the blocking of each processing unit obeys the post-blocking mechanism. The front and rear buffers of each processing unit are both limited and have the same capacity. When the AGV loads workpieces in the raw material warehouse, it follows the load balancing principle to decide which processing center to go to. In order to ensure the system output capacity, the time it takes for the equipment to transport the workpiece is not considered, and the overall system output capacity is determined by the smallest process.

[0188] The parameters of the simulation system are: robot type i, i = 1, 2, 3, ...,; the unit price of the i-th robot v i (10,000 yuan); the number of manipulators of type i x i ; Various types of manipulator quantity configuration vector X, X = {x i}; AGV type j, j = 1, 2, 3, ...,; AGV unit price μ of the i-th type j (10,000 yuan); the number of AGVs of type i y j ; Various AGV quantity configuration vector Y, Y = {y j}; total equipment investment cost Q (10,000 yuan); average system output rate θ (pieces / min); average system production cycle Γ (min); mathematical expectation of random function E{·}; average speed V of AGV running a circle without waiting; AGV capacity C; system output rate l; raw material arrival rate λ (pieces / min); buffer capacity B1 before processing unit; buffer capacity B2 after processing unit; random element ξ.

[0189] Furthermore, the track is the running path of the AGV and the robot, wherein the running path of the AGV is a bidirectional single-loop path, and the running path of the robot is a bidirectional multiple-loop path.

[0190] Among them, due to the limitation of buffer space capacity and delivery deadline, the front and rear buffers in the processing unit are limited buffers. Due to the large size and heavy weight of the workpiece, a robot is required to load and unload.

[0191] Furthermore, the AGV loading quantity is determined by the AGV capacity, the existing workpieces in the raw material warehouse and the remaining capacity of the destination, and its mathematical expression is:

[0192]

[0193] Among them, V c is the number of workpieces loaded by AGV, n is the number of processing units, There are W workpieces in the raw material warehouse. is the remaining capacity W of the front buffer of the kth processing unit.

[0194] The mathematical expression of the overall system output capability is:

[0195]

[0196] Where l is the system output rate, x i is the number of configurations of the i-th type of manipulator, R t is the machining rate of the machining center, R m The working speed of the robot for loading and unloading.

[0197] Furthermore, the workpieces are generated according to Poisson distribution and are of the same type and enter the raw material warehouse. After the AGV accepts the handling task, it goes to the raw material warehouse to carry V c The AGV moves the workpieces to the front buffer with the least workpieces according to the load balancing principle.

[0198] Among them, there are two types of tasks that AGV can handle after the transportation is completed, namely:

[0199] (1) When That is, the urgency of finished product delivery is higher than that of raw material handling, so the AGV goes to the rear buffer with the most finished products in the processing unit to deliver the finished products, where r is the fitness value of the AGV reverse driving transfer probability, There are W workpieces in the back buffer of the kth processing unit.

[0200] (2) When That is, the urgency of shipping finished products out of the warehouse is lower than that of transporting raw materials. The AGV drives in the opposite direction and returns to the raw material warehouse to transport the workpiece to the front buffer zone.

[0201] The purpose of adding the fitness value r is to prevent the AGV from frequently reversing and causing the next AGV to wait.

[0202] Furthermore, when the machining center needs to load or unload materials, it will send an application to the task pool; the robot follows the first-come-first-served rule of the task pool and selects the shortest circular track to handle the loading or unloading task.

[0203] Furthermore, the mathematical model for optimizing the number and types of devices in the simulation system described in step S2 is expressed in the following mathematical form:

[0204]

[0205] Among them, (X * , Y * ) is the optimal vector set of the manipulator and AGV, which is used to minimize the total cost; the performance index constraint parameter conditions set are:

[0206] E{Γ(X, Y:ξ)}≤Γ max

[0207] E{θ(X, Y:ξ)}≥θ min

[0208] X, Y∈N +

[0209] Among them, E{Γ(X, Y:ξ)} is the average production cycle, and the conditions are set so that the simulation production cycle Γ max is smaller than the actual average production cycle of the system, which is used to ensure that the delivery deadline is met; E{θ(X, Y:ξ)} is the average output rate setting condition so that the simulation average output rate θ min Greater than the average output rate of the system, used to ensure that the system has sufficient capacity; X, Y∈N + Indicates that the X vector and the Y vector are positive integers.

[0210] Among them, since E{Γ(X, Y:ξ)} and E{θ(X, Y:ξ)} cannot be expressed in mathematical closed form with the robot X and the number of AGVs Y, the typical nonlinear integer programming method cannot solve the problem. Therefore, the simulation system is used to obtain the two performance indicators of average production cycle and system output rate, and then the gray wolf algorithm is embedded to calculate and optimize the equipment configuration to minimize the equipment configuration cost.

[0211] Furthermore, the step S3 is specifically as follows:

[0212] S31. Initialize the wolf pack, use integer coding to randomly generate the initial solution, divide the feasible solution sequence into two parts: the number of the i-th type of manipulator and the number of the j-th type of AGV, define the gray wolf individual as Z(i), the number of each type of manipulator as X, the number of each type of AGV as Y, and the wolf pack as

[0213] S32. Import the encoded wolf individuals into the simulation model to solve the fitness value. The parameters of each wolf individual are run in the simulation system for several times and then the average value is taken to obtain two performance indicators, namely the system output rate θ and the average production cycle T, and the fitness value H is obtained by solving. A hierarchy of gray wolves is established, which is divided into four levels from high to low: α, β, δ, ω. The fitness values ​​H of the wolf individuals are sorted from small to large. The first three wolves with the smallest fitness are α, β, and δ wolves, and the rest are ω wolves.

[0214] Among them, the smaller the fitness value, the better the performance index obtained.

[0215] S33. Update the parameters of the gray wolf optimization algorithm according to the number of iterations, and update the gray wolf population; update the convergence factor a, coefficient vectors A and C that decrease linearly with the number of iterations according to the current L number of iterations; adopt a population update mechanism that eliminates the last N individuals after each iteration.

[0216] S34. Update the position of the gray wolf. The gray wolf individual adjusts its own position according to the positions of α, β, and δ, and introduces a dynamic proportional weight method to reduce the probability of the algorithm falling into the local optimum and complete an iteration.

[0217] Among them, the principle of the gray wolf individuals adjusting their own positions according to the positions of α, β, and δ is that when the gray wolf finds the target, the α wolf will command the β wolf and the δ wolf to surround the target.

[0218] S35, repeat steps S32 and S33, and save the current gray wolf optimization algorithm after reaching the set maximum number of iterations.

[0219] Among them, as the number of iterations increases, the alpha , beta , and delta alphas dynamically adjust the positions of the wolf pack, effectively increasing the probability of finding the global optimal solution.

[0220] Furthermore, the mathematical expression of step S3 is:

[0221] Z(i)=[X|Y]=[x1,x2,…,x n |y1,y2,…,y n ]

[0222] X={x i}

[0223] Y={y j}

[0224]

[0225] Among them, lnum represents the number of gray wolves in the wolf pack.

[0226] Furthermore, the mathematical expression of the fitness value H of the i-th gray wolf in S32 is:

[0227] H i =T i -θ i

[0228] Furthermore, the mathematical expression of the parameters of the gray wolf optimization algorithm and the gray wolf population described in S33 is:

[0229] A=2a·r1-a

[0230] C=2·r2

[0231]

[0232] N=0.3*lnum

[0233] Among them, t represents the number of current iterations, r1 and r2 are random elements in the value range [0, 1], and a decreases from 2 to 0.

[0234] Among them, if the value of N is too large, the algorithm is difficult to converge; if the value of N is too small, the number of new solutions generated is small, which reduces the algorithm's optimization ability. Using N = 0.3*lnum can not only ensure population diversity, but also make it easy for the algorithm to converge and optimize.

[0235] Furthermore, the mathematical expression of the gray wolf individual in S34 adjusting its own position according to the positions of α, β, and δ is:

[0236] D α =C1·Z α -Z

[0237] D β =C2·Z β -Z

[0238] D δ=C3·Z δ -Z

[0239] Z1=Z α -A1·(D α )

[0240] Z2=Z β -A2·(D β )

[0241] Z3=Z δ -A3·(D δ )

[0242] Among them, D α , D β , D δ Represents the distance vector between individuals α, β, δ and the gray wolf ω, Z α , Z β , Z δ Represents the current position of α, β, δ wolves; C1, C2, C3 represent vectors with random values ​​in the range [0, 1], Z is the current position of the gray wolf, and Z1, Z2, Z3 are the position vectors of the gray wolf.

[0243] Furthermore, the mathematical expression of the dynamic proportional weight in S34 is:

[0244]

[0245]

[0246]

[0247]

[0248]

[0249]

[0250] Z(t+1)=W1×σ1+W2×σ2+W3×σ3

[0251] Among them, σ1, σ2, σ3 are the proportional weights of the position vector; W1, W2, W3 are the learning proportions of the gray wolf for the α, β, and δ wolf positions respectively, and Z(t+1) is the updated gray wolf position.

[0252] Among them, the purpose of introducing the dynamic proportional weight method is to solve the problem that in the traditional GWO, the gray wolf position update adopts the average position of the three wolves α, β, and δ, and the proportional weight is always equal and unchanged, which leads to the problem that when the α wolf in the algorithm is the local optimal, the other gray wolves will continue to approach the α wolf and fall into the local optimal state, which can significantly reduce the probability of the algorithm falling into the local optimal state.

[0253] Furthermore, the step S4 is specifically as follows: the iterative grey wolf algorithm is used to solve the mathematical model of the number and type of equipment in the optimization simulation system, and the obtained grey wolf position is the optimal solution for the number and type of equipment.

[0254] The icons describing the structural positional relationships in the accompanying drawings are for illustrative purposes only and should not be construed as limitations on this patent.

[0255] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A multi-resource collaborative intelligent workshop equipment configuration optimization method, characterized in that: The following steps are involved: S1. Construct a simulation system that simulates the actual production environment; the simulation system specifically comprises: constructing a simulation system according to the actual production environment, the simulation system comprising a task pool, a raw material warehouse, a processing unit, a finished product warehouse and a track, the processing unit comprising a front buffer, a processing center and a rear buffer connected in sequence, the track being the running path of the AGV and the manipulator; the task pool issues tasks to the AGV and the manipulator; The working steps of the simulation system are as follows: AGV transports the raw materials from the raw material warehouse to the front buffer of each processing unit, the manipulator loads the raw materials from the front buffer to the processing center, the processing center processes the raw materials, the manipulator unloads the finished products processed by the processing center to the rear buffer, and the AGV transports the finished products stored in the rear buffer to the finished product warehouse; The simulation system follows the following rules: the arrival of workpieces in the system follows a Poisson distribution process, the arrival of each workpiece is an independent event, and the mean arrival rate is λ; the system determines whether the raw material warehouse is full before the workpiece arrives. If the raw material warehouse is full, the system will refuse the workpiece to enter; each processing unit has only one processing device, which can only allow a single workpiece to be processed, and the processing is subject to the parameter R t Negative exponential distribution; the processing equipment needs a robot to load or unload materials before processing, and the loading and unloading time of each time is independent of each other. The loading and unloading time is subject to the parameter R m The negative exponential distribution of the system processing unit follows the first-come-first-served rule, and the blocking of each processing unit is subject to the post-blocking mechanism. The front and rear buffers of each processing unit are both limited and have the same capacity. When the AGV loads workpieces in the raw material warehouse, it follows the load balancing principle to decide which processing center to go to. In order to ensure the system output capacity, the time it takes for the equipment to transport the workpiece is not considered, and the overall system output capacity is determined by the process with the shortest processing time. S2. Establish a mathematical model for optimizing the number and types of devices in the simulation system and set performance index constraint parameter conditions; the mathematical model for optimizing the number and types of devices in the simulation system is expressed in the following mathematical form: Among them, (X * ,Y * ) is the optimal vector set of the manipulator and AGV, which is used to minimize the total cost; the performance index constraint parameter conditions set are: E{Γ(X,Y:ξ)}≤Γ max E{θ(X,Y:ξ)}≥θ min X,Y∈N + Among them, E{Γ(X,Y:ξ)} is the average production cycle, and the conditions are set so that the simulation production cycle Γ max is smaller than the actual average production cycle of the system to ensure that the delivery deadline is met; E{θ(X,Y:ξ)} is the average output rate setting condition so that the simulation average output rate θ min Greater than the average output rate of the system, used to ensure that the system has sufficient capacity; X, Y∈N + Indicates that the X vector and the Y vector are positive integers; S3, using the gray wolf optimization algorithm to calculate the mathematical model, using the simulation system in the gray wolf optimization algorithm to obtain the performance index, updating the population and position of the gray wolf optimization algorithm according to the optimal performance index, and obtaining the gray wolf optimization algorithm after a set number of iterations; S4. The solution obtained by the iterative grey wolf optimization algorithm to solve the mathematical model is decoded to obtain the optimized device configuration.

2. According to the multi-resource collaborative intelligent workshop equipment configuration optimization method of claim 1, it is characterized in that: The track is the running path of the AGV and the robot, wherein the running path of the AGV is a bidirectional single-loop path, and the running path of the robot is a bidirectional multiple-loop path.

3. The multi-resource collaborative intelligent workshop equipment configuration optimization method according to claim 1 is characterized in that: The AGV loading quantity is determined by the AGV capacity, the existing workpieces in the raw material warehouse and the remaining capacity at the destination, and its mathematical expression is: Among them, V c is the number of workpieces loaded by AGV, n is the number of processing units, There are W workpieces in the raw material warehouse. is the remaining capacity W of the front buffer of the kth processing unit; The mathematical expression of the overall system output capability is: Where l is the system output rate, x i is the number of configurations of the i-th type of manipulator, R t is the machining rate of the machining center, R m The working speed of the robot for loading and unloading.

4. The multi-resource collaborative intelligent workshop equipment configuration optimization method according to claim 3 is characterized in that: Workpieces are generated according to Poisson distribution and are of the same type and enter the raw material warehouse. After the AGV accepts the handling task, it goes to the raw material warehouse to carry V c The AGV moves the workpiece to the front buffer with the least current workpiece according to the load balancing principle. After the workpiece is moved, the AGV can handle two tasks: (1) When That is, the urgency of finished product delivery is higher than that of raw material handling, so the AGV goes to the rear buffer with the most finished products in the processing unit to deliver the finished products, where r is the fitness value of the AGV reverse driving transfer probability, There are W workpieces in the back buffer of the kth processing unit; (2) When That is, the urgency of finished product delivery is lower than that of raw material transportation, and the AGV drives in the opposite direction back to the raw material warehouse to transport the workpiece to the front buffer zone; When the machining center needs to load or unload materials, it will send an application to the task pool; the robot follows the first-come-first-served rule of the task pool and selects the shortest circular track to handle the loading or unloading task.

5. A multi-resource collaborative intelligent workshop equipment configuration optimization method according to any one of claims 1 to 4, characterized in that: The step S3 is specifically as follows: S31. Initialize the wolf pack, use integer coding to randomly generate the initial solution, divide the feasible solution sequence into two parts: the number of the i-th type of manipulator and the number of the j-th type of AGV, define the gray wolf individual as Z(i), the number of each type of manipulator as X, the number of each type of AGV as Y, and the wolf pack as S32, import the encoded wolf pack individuals into the simulation model to solve the fitness value, and take the average value of the parameters of each wolf pack individual after running the simulation several times in the simulation system, and obtain two performance indicators, namely, the system output rate θ and the average production cycle Τ, and solve to obtain the fitness value H; formulate a hierarchy system for gray wolves, which is divided into four levels from high to low: α, β, δ, ω, and sort the fitness values ​​H of the wolf pack individuals from small to large. The first three wolves with the smallest fitness are α, β, and δ wolves, and the rest are ω wolves; S33, updating the parameters of the gray wolf optimization algorithm according to the number of iterations, and updating the gray wolf population; updating the convergence factor a, coefficient vectors A and C that decrease linearly with the number of iterations according to the current number of iterations; adopting a population update mechanism that eliminates the last N individuals after each iteration; S34, update the position of the gray wolf. The gray wolf individual adjusts its own position according to the positions of α, β, and δ, and introduces a dynamic proportional weight method to reduce the probability of the algorithm falling into the local optimum, completing one iteration; S35, repeat steps S32 and S33, and save the current gray wolf optimization algorithm after reaching the set maximum number of iterations.

6. The multi-resource collaborative intelligent workshop equipment configuration optimization method according to claim 5 is characterized in that: The mathematical expression of step S3 is: Z(i)=[X|Y]=[x1,x2,…,x n |y1,y2,…,y n ] X={x i } And={and j } Where lnum represents the number of gray wolves in the wolf pack; The mathematical expression of the fitness value H of the i-th gray wolf in S32 is: H i =T i -θ i The mathematical expression of the parameters of the gray wolf optimization algorithm and the gray wolf population described in S33 is: A=2a·r1-a C＝2·r2 N=0.3*lnum Where t represents the number of iterations, r1 and r2 are random elements in the range [0,1], and a decreases from 2 to 0; The mathematical expression of the gray wolf individual adjusting its position according to the positions of α, β, and δ described in S34 is: D α =C1·Z α -Z D β =C2 Z β -WITH D δ =C3·Z δ -Z Z1=Z α -A1·(D α ) <h2 style=";text-align:left;direction:ltr">Z2 = Z<h2 style=";text-align:left;direction:ltr"> β <h2 style=";text-align:left;direction:ltr"> -A2·(D<h2 style=";text-align:left;direction:ltr"> β <h2 style=";text-align:left;direction:ltr"> ) <h2 style=";text-align:left;direction:ltr">Z3=Z<h2 style=";text-align:left;direction:ltr"> δ <h2 style=";text-align:left;direction:ltr"> -A3·(D<h2 style=";text-align:left;direction:ltr"> δ <h2 style=";text-align:left;direction:ltr"> ) Among them, D α , D β , D δ Represents the distance vector between individuals α, β, δ and the gray wolf ω, Z α , Z β , Z δ represents the current position of α, β, and δ wolves; C1, C2, and C3 represent vectors with random values ​​in the range [0,1]; Z is the current position of the gray wolf; and Z1, Z2, and Z3 are the position vectors of the gray wolf; The mathematical expression of the dynamic proportional weight in S34 is: Z(t+1)=W1×σ1+W2×σ2+W3×σ3 Among them, σ1, σ2, σ3 are the proportional weights of the position vector; W1, W2, W3 are the learning proportions of the gray wolf for the α, β, and δ wolf positions respectively, and Z(t+1) is the updated gray wolf position.

7. The multi-resource collaborative intelligent workshop equipment configuration optimization method according to claim 1 is characterized in that: The step S4 is specifically as follows: the iterative grey wolf algorithm is used to solve the mathematical model of the number and type of equipment in the optimization simulation system, and the obtained grey wolf position is the optimal solution for the number and type of equipment after decoding and splitting.

8. A multi-resource collaborative intelligent workshop equipment configuration optimization system, comprising a memory and a processor, wherein the memory comprises a multi-resource collaborative intelligent workshop equipment configuration optimization program, and when the multi-resource collaborative intelligent workshop equipment configuration optimization program is executed by the processor, the following steps are implemented: S1. Construct a simulation system that simulates the actual production environment; the simulation system specifically comprises: constructing a simulation system according to the actual production environment, the simulation system comprising a task pool, a raw material warehouse, a processing unit, a finished product warehouse and a track, the processing unit comprising a front buffer, a processing center and a rear buffer connected in sequence, the track being the running path of the AGV and the manipulator; the task pool issues tasks to the AGV and the manipulator; The working steps of the simulation system are as follows: AGV transports the raw materials from the raw material warehouse to the front buffer of each processing unit, the manipulator loads the raw materials from the front buffer to the processing center, the processing center processes the raw materials, the manipulator unloads the finished products processed by the processing center to the rear buffer, and the AGV transports the finished products stored in the rear buffer to the finished product warehouse; The simulation system follows the following rules: the arrival of workpieces in the system follows a Poisson distribution process, the arrival of each workpiece is an independent event, and the mean arrival rate is λ; the system determines whether the raw material warehouse is full before the workpiece arrives. If the raw material warehouse is full, the system will refuse the workpiece to enter; each processing unit has only one processing device, which can only allow a single workpiece to be processed, and the processing is subject to the parameter R t Negative exponential distribution; the processing equipment needs a robot to load or unload materials before processing, and the loading and unloading time of each time is independent of each other. The loading and unloading time is subject to the parameter R m The negative exponential distribution of the system processing unit follows the first-come-first-served rule, and the blocking of each processing unit is subject to the post-blocking mechanism. The front and rear buffers of each processing unit are both limited and have the same capacity. When the AGV loads workpieces in the raw material warehouse, it follows the load balancing principle to decide which processing center to go to. In order to ensure the system output capacity, the time it takes for the equipment to transport the workpiece is not considered, and the overall system output capacity is determined by the process with the shortest processing time. S2, establish a mathematical model for optimizing the number and type of equipment in the simulation system and set performance indicator constraint parameter conditions; S3, using the gray wolf optimization algorithm to calculate the mathematical model, using the simulation system in the gray wolf optimization algorithm to obtain performance indicators, updating the population and position of the gray wolf optimization algorithm according to the optimal performance indicators, and obtaining the gray wolf optimization algorithm after a set number of iterations; the mathematical model for optimizing the number and type of equipment in the simulation system, its mathematical expression is: in, (X * ,Y * ) is the optimal vector set of the manipulator and AGV, which is used to minimize the total cost; the performance index constraint parameter conditions set are: E{Γ(X,Y:ξ)}≤Γ max E{θ(X,Y:ξ)}≥θ min X,Y∈N + Among them, E{Γ(X,Y:ξ)} is the average production cycle, and the conditions are set so that the simulation production cycle Γ max is smaller than the actual average production cycle of the system to ensure that the delivery deadline is met; E{θ(X,Y:ξ)} is the average output rate setting condition so that the simulation average output rate θ min Greater than the average output rate of the system, used to ensure that the system has sufficient capacity; X, Y∈N + Indicates that the X vector and the Y vector are positive integers; S4. The solution obtained by the iterative grey wolf optimization algorithm to solve the mathematical model is decoded to obtain the optimized device configuration.

Images