Production scheduling optimization method and system based on multi-objective mixing

Abstract

The invention discloses a production scheduling optimization method and system based on multi-objective mixing, and the method comprises the steps: designing a first objective function for minimizing the late shift subsidy and a second objective function for minimizing the energy consumption cost, and determining a constraint condition; initializing a particle swarm, and calculating fitness values of each particle in the particle swarm corresponding to the first objective function and the second objective function; storing the non-dominated solution of each particle through an external storage library, calculating an entropy change value, a probability threshold value, a probability value and a random value of the external storage library, and determining a global optimal particle; updating each particle according to the global optimal particle, the entropy change value, the probability threshold and the constraint condition to obtain a non-dominated solution set; and selecting a compromise solution in the non-dominated solution set as a production scheduling optimization result through an approximate ideal solution sorting method. The method can reduce the subjectivity of weight setting, does not need repeated trial and error, improves the production efficiency and intelligent level of automobile production scheduling, and can be widely applied to the technical field of production scheduling.

Description

Technical Field

[0001] The present application relates to the technical field of production scheduling, and in particular to a production scheduling optimization method and system based on multi-objective mixing. Background Art

[0002] At present, the production scheduling system adopted by most automobile manufacturers is still based on the linear weighted method, that is, multiple objectives are combined into a single objective function through preset weights for optimization. This method has obvious defects. On the one hand, the setting of weights often depends on the subjective judgment of decision makers and lacks scientific basis, which easily leads to the result being biased towards a single objective and ignoring the importance of other objectives; on the other hand, once the user is not satisfied with the optimization result or wants to explore more possibilities, it is necessary to repeatedly adjust the weight and re-run the optimization process, which undoubtedly increases the complexity and time-consumingness of decision-making. Summary of the invention

[0003] In order to solve the above technical problems, the purpose of the present invention is to provide a more efficient production scheduling optimization method and system based on multi-objective mixing.

[0004] To achieve the above purpose, one aspect of an embodiment of the present application proposes a production scheduling optimization method based on multi-objective mixing, comprising the following steps:

[0005] Design a first objective function for minimizing the late shift subsidy and a second objective function for minimizing the energy consumption cost, and determine the constraints;

[0006] Initializing a particle swarm, and calculating a fitness value of each particle in the particle swarm corresponding to the first objective function and the second objective function;

[0007] The non-dominated solutions of each particle are stored in an external repository, and an entropy change value, a probability threshold, a probability value, and a random value of the external repository are calculated, and a global optimal particle is determined according to the probability value and the random value;

[0008] updating each of the particles according to the global best particle, the entropy change value, the probability threshold and the constraint condition, and when a preset termination condition is reached, storing the non-dominated solution corresponding to the updated particle in the external storage repository to obtain a non-dominated solution set;

[0009] A compromise solution in the non-dominated solution set is selected as a production scheduling optimization result according to the fitness value through an approximate ideal solution sorting method.

[0010] In some embodiments, the first objective function for minimizing the late shift subsidy and the second objective function for minimizing the energy consumption cost are designed to specifically include:

[0011] Obtain target output, number of overtime workers, subsidy time period, energy consumption per shift, production status in each preset time period, and peak-valley electricity prices corresponding to each peak and valley period, including the electricity prices corresponding to peak time period, normal time period, and valley time period;

[0012] Setting a production state array, and obtaining a cost array according to the peak and valley electricity prices and the subsidy time period;

[0013] Obtaining the first objective function according to the number of overtime workers, the production status array, and the cost array;

[0014] The second objective function is obtained according to the target output, the peak-valley electricity price, the energy consumption of a single vehicle per shift, and the production status in each preset time period.

[0015] In some embodiments, the determining of the constraint condition specifically includes:

[0016] Determine the start time and end time, and obtain the target production time;

[0017] The constraint condition is determined to be that the sum of the production states of each time period from the start time to the end time is equal to the target output production time.

[0018] In some embodiments, the initializing the particle swarm specifically includes:

[0019] Setting the initial position and initial velocity of each particle, and determining the individual extreme value and group extreme value of each particle;

[0020] Determine the inertia weight and acceleration constant.

[0021] In some embodiments, storing the non-dominated solutions of each particle in an external repository and calculating the entropy change value, probability threshold, probability value and random value of the external repository specifically includes:

[0022] Initializing the external storage repository, and storing each of the non-dominated solutions in the external storage repository;

[0023] Representing the position of each non-dominated solution in the target space by the fitness value corresponding to each particle, and obtaining the coordinate value of each non-dominated solution;

[0024] Calculate the potential value and density value of each of the non-dominated solutions according to the coordinate values;

[0025] The entropy change value, probability threshold, probability value and random value of the external storage repository are calculated.

[0026] In some embodiments, determining the global best particle according to the probability value and the random value specifically includes:

[0027] Determine a minimum potential value according to each of the potential values, and determine a minimum density value according to each of the density values;

[0028] When the random value is greater than the probability value, determining the particle corresponding to the minimum potential value as the global best particle;

[0029] When the random value is greater than the probability value, the particle corresponding to the minimum density value is determined as the global best particle.

[0030] In some embodiments, the updating of each particle according to the global best particle, the entropy change value, the probability threshold and the constraint condition specifically includes:

[0031] When the absolute value of the entropy change value is greater than or equal to the probability threshold, each of the particles is updated according to the global best particle and the constraint condition by using a multi-objective particle swarm optimization algorithm;

[0032] When the absolute value of the entropy change value is less than the probability threshold, each particle is updated according to the global best particle and the constraint condition by a multi-objective evolutionary algorithm.

[0033] In some embodiments, the updating of each particle according to the global best particle and the constraint condition by using a multi-objective particle swarm optimization algorithm specifically includes:

[0034] Update the initial velocity of each particle according to the global best particle, the individual extreme value, the group extreme value, the inertia weight, and the acceleration constant;

[0035] Updating the initial position of each particle according to the initial position and the updated initial velocity;

[0036] Wherein, the updated initial position satisfies the constraint condition.

[0037] In some embodiments, the non-dominated solution set includes multiple candidate non-dominated solutions, and the method of selecting a compromise solution in the non-dominated solution set as a production scheduling optimization result according to the fitness value by using the approximate ideal solution sorting method specifically includes:

[0038] Normalizing the fitness values ​​corresponding to the candidate non-dominated solutions to obtain a normalized decision matrix;

[0039] Determining weights of the first objective function and the second objective function, and obtaining a weighted normalized decision matrix according to the weights and the normalized decision matrix;

[0040] A positive ideal solution and a negative ideal solution are calculated according to the fitness value;

[0041] Calculating the Euclidean distance from each of the candidate non-dominated solutions to the positive ideal solution to obtain a first Euclidean distance, and calculating the Euclidean distance from each of the candidate non-dominated solutions to the negative ideal solution to obtain a second Euclidean distance;

[0042] Calculating a relative proximity of each of the candidate non-dominated solutions according to the first Euclidean distance and the second Euclidean distance;

[0043] The candidate non-dominated solution corresponding to the largest relative proximity is determined as a compromise solution, and the production scheduling optimization result is obtained.

[0044] To achieve the above purpose, another aspect of the embodiment of the present application proposes a production scheduling optimization system based on multi-objective mixing, including:

[0045] An objective function design module, used to design a first objective function for minimizing the late shift subsidy and a second objective function for minimizing the energy consumption cost, and determine constraint conditions;

[0046] A particle evaluation module, used for initializing a particle swarm and calculating a fitness value of each particle in the particle swarm corresponding to the first objective function and the second objective function;

[0047] An external repository updating module, used for storing the non-dominated solutions of each particle through an external repository, calculating the entropy change value, probability threshold, probability value and random value of the external repository, and determining the global best particle according to the probability value and the random value;

[0048] A particle updating module, used for updating each of the particles according to the global best particle, the entropy change value, the probability threshold and the constraint condition, and when a preset termination condition is reached, storing the non-dominated solution corresponding to the updated particle in the external storage repository to obtain a non-dominated solution set;

[0049] The compromise solution selection module is used to select a compromise solution in the non-dominated solution set as a production scheduling optimization result according to the fitness value by using an approximate ideal solution sorting method.

[0050] The beneficial effects of the present invention are as follows: the production scheduling optimization method and system based on multi-objective mixing of the present invention first designs a first objective function for minimizing the night shift subsidy and a second objective function for minimizing the energy consumption cost, and determines the constraints, and then initializes the particle swarm, calculates the fitness value of each particle in the particle swarm corresponding to the first objective function and the second objective function, stores the non-dominated solutions of each particle through an external repository, and calculates the entropy change value, probability threshold, probability value and random value of the external repository, determines the global best particle according to the probability value and the random value, and then updates each particle according to the global best particle, entropy change value, probability threshold and constraints. When the preset termination condition is reached, the non-dominated solution corresponding to the updated particle is stored in the external repository to obtain a non-dominated solution set, and finally selects a compromise solution in the non-dominated solution set according to the fitness value through the approximate ideal solution sorting method as the production scheduling optimization result. The present invention can optimize multiple objectives at the same time without pre-setting objective weights. Instead, it automatically searches for and presents a series of non-dominated solutions through the Pareto optimal concept, and then selects a compromise solution in the non-dominated solution set as the production scheduling optimization result through the approximate ideal solution sorting method. This can reduce the subjectivity of weight setting and eliminates the need for repeated trial and error, thereby improving the production efficiency and intelligence level of automobile production scheduling. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solution in the embodiments of the present invention, the following introduction is made to the drawings required for use in the embodiments of the present invention. It should be understood that the drawings introduced below are only for the convenience of clearly describing some embodiments of the technical solution of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0052] Figure 1 A flowchart of a production scheduling optimization method based on multi-objective mixing provided by an embodiment of the present invention;

[0053] Figure 2 A flowchart of a production scheduling optimization method based on multi-objective mixing provided by an embodiment of the present invention;

[0054] Figure 3 A schematic diagram of the structure of a production scheduling optimization system based on multi-objective mixing provided by an embodiment of the present invention;

[0055] Figure 4 A schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below in conjunction with the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present application. They are only examples of devices and methods consistent with some aspects of the embodiments of the present application as detailed in the attached claims.

[0057] It is understood that the terms "first", "second", etc. used in this application can be used to describe various concepts in this article, but unless otherwise specified, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another concept. For example, without departing from the scope of the embodiment of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if" and "if" as used herein can be interpreted as "at the time of" or "when" or "in response to determination".

[0058] The terms "at least one", "multiple", "each", "any", etc. used in this application, at least one includes one, two or more, multiple includes two or more, each refers to each of the corresponding multiple, and any refers to any one of the multiple.

[0059] At present, the production scheduling system adopted by most automobile manufacturers is still based on the linear weighted method, that is, multiple objectives are combined into a single objective function through preset weights for optimization. This method has obvious defects. On the one hand, the setting of weights often depends on the subjective judgment of decision makers and lacks scientific basis, which easily leads to the result being biased towards a single objective and ignoring the importance of other objectives; on the other hand, once the user is not satisfied with the optimization result or wants to explore more possibilities, it is necessary to repeatedly adjust the weight and re-run the optimization process, which undoubtedly increases the complexity and time-consumingness of decision-making.

[0060] Multi-Objective Particle Swarm Optimization (MOPSO) is an algorithm based on particle swarm optimization. The MOPSO algorithm simulates the behavior of bird flocks foraging, and each particle represents a possible solution. By continuously updating the speed and position of the particles, the particle swarm gradually approaches the optimal solution. In MOPSO, in order to deal with multi-objective problems, mechanisms such as external archiving and crowding distance are used. The multi-objective evolutionary algorithm (NSGA-II) is an algorithm based on genetic algorithms. Through mechanisms such as non-dominated sorting and crowding distance, NSGA-II maintains the diversity and superiority of the population during the optimization process.

[0061] However, both NSGA-II and MOPSO have their own limitations when dealing with complex multi-objective optimization problems alone. NSGA-II may have a trade-off between solution set distribution and convergence speed, while MOPSO may face the problems of premature convergence and insufficient solution set diversity.

[0062] To this end, an embodiment of the present invention proposes a production scheduling optimization method based on multi-objective mixing. First, a first objective function for minimizing the night shift subsidy and a second objective function for minimizing the energy consumption cost are designed, and constraints are determined. Then, a particle swarm is initialized, and the fitness value of each particle in the particle swarm corresponding to the first objective function and the second objective function is calculated. The non-dominated solution of each particle is stored in an external repository, and the entropy change value, probability threshold, probability value and random value of the external repository are calculated. The global best particle is determined according to the probability value and the random value. Then, each particle is updated according to the global best particle, entropy change value, probability threshold and constraints. When the preset termination condition is reached, the non-dominated solution corresponding to the updated particle is stored in the external repository to obtain a non-dominated solution set. Finally, a compromise solution in the non-dominated solution set is selected according to the fitness value by the approximate ideal solution sorting method as the production scheduling optimization result. The present invention can optimize multiple objectives at the same time without pre-setting objective weights. Instead, it automatically searches for and presents a series of non-dominated solutions through the Pareto optimal concept, and then selects a compromise solution in the non-dominated solution set as the production scheduling optimization result through the approximate ideal solution sorting method. This can reduce the subjectivity of weight setting and eliminates the need for repeated trial and error, thereby improving the production efficiency and intelligence level of automobile production scheduling.

[0063] Reference Figure 1 , Figure 1 The present invention provides a flowchart of a production scheduling optimization method based on multi-objective mixing. The present invention provides a production scheduling optimization method based on multi-objective mixing, and the method includes steps S101 to S105:

[0064] S101. Design a first objective function for minimizing the night shift subsidy and a second objective function for minimizing the energy consumption cost, and determine the constraints;

[0065] Further as an optional implementation, the step of designing a first objective function for minimizing the night shift subsidy and a second objective function for minimizing the energy consumption cost can be specifically divided into the following steps S1011 to S1014:

[0066] S1011. Obtain the target output, the number of overtime workers, the subsidy time period, the energy consumption per vehicle per shift, the production status within each preset time period, and the peak-valley electricity prices corresponding to each peak and valley. The peak-valley electricity prices include the electricity prices corresponding to the peak period, the normal period, and the valley period;

[0067] S1012. Set up a production status array and obtain a cost array based on the peak-valley electricity prices and the subsidy time period;

[0068] In some optional embodiments, the peak period is 9:00 - 12:00 and 19:00 - 22:00; the normal period is 0:00 - 8:00; the valley period is 8:00 - 9:00, 12:00 - 19:00, and 22:00 - 24:00. The peak-valley electricity price corresponding to the peak period is 0.9659 yuan, the peak-valley electricity price corresponding to the normal period is 0.5854 yuan, and the peak-valley electricity price corresponding to the valley period is 0.2927 yuan. Set the night shift subsidy time period to 0:00 - 8:00 and 18:00 - 24:00, with a subsidy of 10 yuan per hour and 4 overtime workers. The energy consumption per vehicle per shift is calculated by dividing the total energy consumption within the shift time by the shift output.

[0069] Set the production status array A′ as A′ = [a′0,…,a′ 23 , representing 24 time periods, each period being 1 hour.

[0070] The cost array C is

[0071] The production status within each preset time period refers to the statistical number of non-zero elements in different time periods A′ after scheduling. Among them, the total number of all non-zero elements in the time periods 9 < i <= 11 and 19 < i <= 21 is k1; the total number of all non-zero elements in the time periods 0 <= i <= 7 is k2; the total number of all non-zero elements in the time periods i = 8, 12 <= i <= 18, and 22 <= i <= 23 is k3.

[0072] Those skilled in the art can understand that the target output, the number of overtime workers, the subsidy time period, the energy consumption per vehicle per shift, the preset time period, and the peak-valley electricity prices corresponding to each peak and valley can all be set according to actual production requirements.

[0073] S1013, obtaining a first objective function according to the number of overtime workers, the production status array, and the cost array;

[0074] S1014, obtaining a second objective function according to the target output, peak and valley electricity prices, energy consumption of a single vehicle per shift, and the production status in each preset time period.

[0075] Specifically, let K be the target output and V be the production efficiency. is the number of overtime workers, B is the energy consumption of a single vehicle per shift, assuming V, B is a constant that does not change.

[0076] The time required to complete the production task, Is a positive integer.

[0077] Then, the first objective function f1(A′) for minimizing the late shift subsidy can be determined as follows:

[0078]

[0079] The second objective function f2(A′,K) for minimizing energy consumption cost is shown as follows:

[0080] f2(A′,K)=K*B*(0.9659k1+0.5854k2+0.2927k3)

[0081] Among them, a′ i is the production status (0 or 1) in the i-th period after scheduling, c i is the energy consumption cost in the i-th period.

[0082] When the specified time period of the production task is completed and the task output is confirmed, the first objective function and the second objective function can be simplified to obtain:

[0083]

[0084] f2(A′,K)=K*(0.9659k1+0.5854k2+0.2927k3)

[0085] For example, there is a production task that takes 5 hours. Assume that the production task must be completed between 8:00 and 19:00. That is, only 5 elements can be selected between 8<=i<=18 and replaced with 1. One example is A′=[a′0,…,a′ 23 ]=[0,0,0,0,0,0,0,0,1,1,1,0,0,0,1,0,1,0,0,0,0,0,0,0], we need to calculate how to choose which element to replace with 1 so that the unsimplified first objective function The value of the second objective function Min(K*B*(0.9659k1+0.5854k2+0.2927k3)) is the smallest, where It can be obtained through simple calculation.

[0086] As an optional implementation, the step of determining the constraint condition may be specifically divided into the following steps S1015 and S1016:

[0087] S1015. Determine the start time and end time, and obtain the target production time;

[0088] S1016. Determine the constraint condition that the sum of the production status of each time period from the start time to the end time is equal to the target output production time.

[0089] Specifically, in order to ensure that the production task is completed within a specific time window, it is necessary to clarify that this time window is limited by the start time and end time of the production task. First, define the time period variable: In the 24-hour production scheduling, define the variable of each time period as a′ i , where i represents the time period number. If the i-th time period is scheduled for production, then a′ i =1; otherwise a′ i = 0. Then define the application time window constraint: for all i belonging to the ath time period to the bth time period (i.e. a≤i≤b), the time period variable a′ i , we must ensure that exactly The time period is marked as 1 to complete the production task.

[0090] To ensure that only Time periods are selected for production, and the constraints are determined as follows:

[0091]

[0092] In choosing this When selecting a time period, not only should the completion of the production task be met, but also the dual objectives of the night shift subsidy cost and the energy cost should be considered, that is, priority should be given to those time periods with lower costs and lower energy costs. In addition, attention should be paid to the non-zero element counts of time periods 9 to 11 and 19 to 21, the non-zero element counts of time periods 0 to 7, and the non-zero element counts of time periods 8, 12 to 18, and 22 to 23, which will affect the second objective function used to minimize the energy consumption cost.

[0093] S102, initializing a particle swarm, and calculating the fitness value of each particle in the particle swarm corresponding to a first objective function and a second objective function;

[0094] Furthermore, for each particle, the fitness values ​​of the two objective functions f1(A′) and f2(A′,K) are calculated, and then the particles are ranked using non-dominated sorting, and the crowding distance of each particle is calculated.

[0095] As an optional implementation, the step of initializing the particle swarm can be specifically divided into the following steps S1021 and S1022:

[0096] S1021, setting the initial position and initial velocity of each particle, and determining the individual extreme value and group extreme value of each particle;

[0097] S1022. Determine the inertia weight and acceleration constant.

[0098] Specifically, first set the initial position of the particle and initial velocity Where i is the index of the particle and g is the number of iterations. Indicates whether production is carried out in each period, that is, Initial Speed Control the change of particle positions and initialize them to random values ​​close to 0. Then determine the individual extremum and the group extremum. The initial position of each particle is regarded as its individual extremum, because this is the optimal position when the particle starts searching. The group extremum is usually the best of the individual extremums of all particles in the group. In the initialization stage, the group extremum can be set to the position of any particle, and then updated according to the objective function value in each iteration.

[0099] The acceleration constant is used to control the acceleration of particles moving toward their individual and group extremes, and the inertia weight is used to control the speed and direction of particles moving in the search space.

[0100] S103, storing the non-dominated solutions of each particle in an external repository, calculating the entropy change value, probability threshold, probability value and random value of the external repository, and determining the global best particle according to the probability value and the random value;

[0101] It should be noted that in multi-objective optimization, the external repository (REP) is used to save non-dominated solutions, that is, those solutions that are not dominated by other solutions in all objective functions. The update of the external repository (REP) is crucial to maintain the diversity and convergence of the solution set.

[0102] As an optional implementation, the step of storing the non-dominated solutions of each particle in an external repository and calculating the entropy change value, probability threshold, probability value and random value of the external repository can be specifically divided into the following steps S1031 to S1034:

[0103] S1031, initializing an external storage repository, and storing each non-dominated solution in the external storage repository;

[0104] Specifically, an empty external repository (REP) is created, and then the non-dominated solutions corresponding to each particle are stored in the external repository (REP). The external repository (REP) is empty at the beginning and will be dynamically filled during the algorithm operation.

[0105] S1032, representing the position of each non-dominated solution in the target space by the fitness value corresponding to each particle, and obtaining the coordinate value of each non-dominated solution;

[0106] Specifically, the PCCS coordinate value is calculated for each non-dominated solution in the external repository (REP), and the coordinate value is used to measure the position of the non-dominated solution in the objective function space.

[0107] For each solution x k , calculate its coordinate value C by the following formula mk :

[0108]

[0109] Among them, f mk is the solution x k The value of the mth objective function, f min and f max are the minimum and maximum values ​​of the objective function respectively.

[0110] S1033, calculating the potential value and density value of each non-dominated solution according to the coordinate value;

[0111] Specifically, the density value and potential value are calculated for each non-dominated solution in the external repository (REP). The density value is used to measure the degree of spatial crowding around the solution, and the potential value is used to measure the location of the solution in the target space.

[0112] Density value ρ(x k ) is calculated by the following formula:

[0113]

[0114] Among them, d PC (x k ,x l ) is the solution x k and x l The parallel cell distance (PCD) is calculated as follows:

[0115]

[0116] Potential value φ(x k ) is calculated by the following formula:

[0117]

[0118] S1034. Calculate the entropy change value, probability threshold, probability value and random value of the external storage repository.

[0119] Specifically, the entropy change ΔH of the external storage (REP) g The entropy H of the external repository (REP) used to evaluate the current state of the algorithm g Calculated by the following formula:

[0120]

[0121] in, is the number of non-dominated solutions in the cell at row k and column m in the external repository (REP) at iteration g.

[0122] Then the entropy change value ΔH g It can be calculated by the following formula:

[0123] ΔH g =H g -H g-1

[0124] The probability threshold δ is used to distinguish the state of the algorithm, and its calculation formula is shown as follows:

[0125] δ=-2MK1Mlog(K1M)

[0126] Where M is the number of objective functions and K1M is the number of non-dominated solutions in the external repository (REP).

[0127] As an optional implementation, the step of determining the global best particle according to the probability value and the random value can be specifically divided into the following steps S1035 to S1037:

[0128] S1035, determining a minimum potential value according to each potential value, and determining a minimum density value according to each density value;

[0129] S1036. When the random value is greater than the probability value, determine the particle corresponding to the minimum potential value as the global best particle;

[0130] S1037. When the random value is greater than the probability value, determine the particle corresponding to the minimum density value as the global best particle.

[0131] Specifically, using the current entropy change value ΔH g and probability threshold δ to decide whether to select the global best particle. g | is greater than δ, then the probability value of selection is P rg By 0.5-|ΔH g|decision; if |ΔH g | is equal to δ, then the probability value P rg is 0.5; if |ΔH g | is less than δ, then the probability value P rg By 0.5+|ΔH g |decision, as shown in the following formula:

[0132]

[0133] When the randomly generated random value is greater than the probability value P rg , at this time, the algorithm is considered to be in a convergence state (i.e., close to finding a better solution set), and the particle with the smallest potential value in the external repository (REP) is selected as the global best particle; when the randomly generated random value is less than the probability value P rg ,At this time, the algorithm is considered to be in a state where diversity needs to be increased, then the particle with the smallest density value is selected as the global best particle, and the genetic operator of the multi-objective evolutionary algorithm (NSGA-II) is applied to improve the diversity, thereby maintaining the diversity of the solution set and preventing premature convergence.

[0134] S104, updating each particle according to the global best particle, entropy change value, probability threshold and constraint conditions, and when a preset termination condition is reached, storing the non-dominated solution corresponding to the updated particle in an external storage repository to obtain a non-dominated solution set;

[0135] As an optional implementation, the step of updating each particle according to the global best particle, entropy change value, probability threshold and constraint condition can be specifically divided into the following steps S1041 and S1042:

[0136] S1041, when the absolute value of the entropy change value is greater than or equal to the probability threshold, each particle is updated according to the global best particle and constraint conditions through a multi-objective particle swarm optimization algorithm;

[0137] S1042. When the absolute value of the entropy change value is less than the probability threshold, each particle is updated according to the global best particle and constraint conditions through a multi-objective evolutionary algorithm.

[0138] Specifically, by calculating the entropy change ΔH of the external storage (REP) gTo evaluate the state of the algorithm. If the absolute value of the entropy change is greater than or equal to the probability threshold, the algorithm is in a convergence state, and the standard multi-objective particle swarm optimization (MOPSO) method is used to update the velocity and position of the particles. If the absolute value of the entropy change is less than the probability threshold, the algorithm is in a diversity state, and it is necessary to improve the diversity of solutions. At this time, the genetic operator of the multi-objective evolutionary algorithm (NSGA-II) is used, such as simulated binary crossover and polynomial mutation, to perform genetic operations on particles, and calculate the potential value of the new solution, sort it in ascending order according to the potential value, and finally retain the top N solutions as the next generation, thereby improving the diversity of particles.

[0139] As an optional implementation, the step of updating each particle according to the global best particle and the constraint condition by using the multi-objective particle swarm optimization algorithm can be specifically divided into the following steps S1043 and S1044:

[0140] S1043, updating the initial velocity of each particle according to the global best particle, individual extreme value, group extreme value, inertia weight and acceleration constant;

[0141] S1044, updating the initial position of each particle according to the initial position and the updated initial velocity;

[0142] Among them, the updated initial position satisfies the constraint conditions.

[0143] It should be noted that in the particle update phase, particles will update their positions and velocities based on individual extreme values, group extreme values, and learning factors (i.e., acceleration constants). The global best particle, as part of the group extreme value, is used to guide the search direction of the particle swarm and help the algorithm converge to a high-quality solution faster.

[0144] Specifically, the speed of each particle Update according to the following formula:

[0145]

[0146] Among them, w is the inertia weight, which indicates the dependence of the controlled particle on the past velocity, c1 and c2 are the individual learning factor and the group learning factor (i.e., the acceleration constant), which respectively control the tendency of the particle to move toward the individual extreme value and the group extreme value, r1 and r2 are two random numbers in the interval [0,1], which are used to introduce randomness and prevent the particle from falling into the local optimum, pib egst represents individual extreme values, ggbes represents group extreme values, represents the initial velocity of the particle, Indicates the initial position of the particle.

[0147] According to the updated initial velocity, the position of each particle Update according to the following formula:

[0148]

[0149] The updated position needs to satisfy the constraints. If it exceeds the feasible range, it needs to be corrected accordingly.

[0150] Furthermore, we set the maximum number of iterations T max After each particle update, check whether the current number of iterations has reached the maximum number of iterations T max , when the maximum number of iterations T is reached max Or if the algorithm converges, the non-dominated solutions corresponding to the updated particles are added to the external repository (i.e., the external repository is updated) to obtain a non-dominated solution set, otherwise, the particle update step is returned to continue the iteration. The threshold δ is used to determine whether the algorithm finally converges. When the entropy change value ΔH of the external repository (REP) is g If it is greater than or equal to the threshold δ, it is determined that the algorithm converges at this time. The calculation formula of the threshold δ is as follows:

[0151]

[0152] S105. Select a compromise solution in the non-dominated solution set as the production scheduling optimization result according to the fitness value through the approximate ideal solution sorting method.

[0153] As an optional implementation, the non-dominated solution set includes multiple candidate non-dominated solutions. The step of selecting a compromise solution in the non-dominated solution set as the production scheduling optimization result according to the fitness value by the approximate ideal solution sorting method can be specifically divided into the following steps S1051 to S1056:

[0154] S1051, normalizing the fitness values ​​corresponding to the candidate non-dominated solutions to obtain a normalized decision matrix;

[0155] S1052, determining the weights of the first objective function and the second objective function, and obtaining a weighted normalized decision matrix according to the weights and the normalized decision matrix;

[0156] Specifically, first, the objective function values ​​(i.e., fitness values) of all candidate non-dominated solutions in the external repository (REP) need to be normalized. Assuming that there are m candidate non-dominated solutions in the external repository (REP), and each candidate non-dominated solution has n objective function values, the normalized decision matrix is ​​as follows:

[0157]

[0158] Among them, σ ijis the fitness value corresponding to the jth objective function of the ith candidate non-dominated solution, and the objective function includes the first objective function and the second objective function designed above.

[0159] Next, we calculate the weighted normalized decision matrix H, where each objective function has a weight w j , these weights are given in advance, reflecting the importance of each goal. The weighted normalized decision matrix H is shown as follows:

[0160]

[0161] S1053, obtaining a positive ideal solution and a negative ideal solution according to the fitness value;

[0162] Specifically, the positive ideal solution (PIS) is a vector consisting of the maximum value of each objective function, and the negative ideal solution (NIS) is a vector consisting of the minimum value of each objective function. + and negative ideal solution H - As shown below:

[0163]

[0164] S1054, calculating the Euclidean distance from each candidate non-dominated solution to the positive ideal solution to obtain a first Euclidean distance, and calculating the Euclidean distance from each candidate non-dominated solution to the negative ideal solution to obtain a second Euclidean distance;

[0165] Specifically, for each candidate non-dominated solution in the non-dominated solution set, its first Euclidean distance to the positive ideal solution (PIS) is calculated by the following formula: and the second Euclidean distance to the negative ideal solution (NIS)

[0166]

[0167] S1055, calculating the relative proximity of each candidate non-dominated solution according to the first Euclidean distance and the second Euclidean distance;

[0168] S1056. Determine the candidate non-dominated solution corresponding to the maximum relative proximity as a compromise solution, and obtain the production scheduling optimization result.

[0169] Specifically, for each candidate non-dominated solution, the ratio of its Euclidean distance to the positive ideal solution (PIS) to the sum of its Euclidean distances to the positive ideal solution (PIS) and the negative ideal solution (NIS) is calculated to measure the closeness of the solution to the ideal solution, and the relative closeness is obtained. As shown below:

[0170]

[0171] Finally, according to the relative proximity Sort all candidate non-dominated solutions in descending order. The highest relative closeness It indicates that the candidate non-dominated solution is closest to the ideal solution, so the candidate non-dominated solution is determined to be a compromise solution, and the production scheduling optimization result is obtained.

[0172] In summary, the processing flow of the production scheduling optimization method based on multi-objective mixing in the embodiment of the present invention is as follows: Figure 2 As shown:

[0173] The first step is to initialize the parameters, variables, the first objective function and the second objective function;

[0174] Step 2: Determine the constraints [a, b] and

[0175] Step 3: Calculate the fitness values ​​f1 and f2 of the first objective function and the second objective function, and rank the particles using non-dominated sorting;

[0176] Step 4: Store the non-dominated solutions in the external repository (REP), sort the non-dominated solutions in the external repository (REP), and select the top N non-dominated solutions to return;

[0177] Step 5: Calculate the entropy change value, probability threshold, probability value and random value, and determine the size of the random value and probability value;

[0178] Step 6: When the calculated random value is greater than the probability value, the particle with the minimum potential value in the external repository (REP) is selected as the global best particle, and the particle is updated using the multi-objective particle swarm optimization (MOPSO); when the calculated random value is less than the probability value, the particle with the minimum density value is selected as the global best particle, and the genetic operator of the multi-objective evolutionary algorithm (NSGA-II) is used to improve the diversity of particle updates;

[0179] Step 7: Update the particles and ensure that the updated particle speed and position meet the constraints;

[0180] Step 8: When the preset maximum number of iterations or other termination conditions are reached, the non-dominated solutions in the external repository (REP) are output to obtain a non-dominated solution set;

[0181] Step 9: Calculate the weighted normalized decision matrix;

[0182] Step 10: Calculate the positive ideal solution and the negative ideal solution;

[0183] Step 11: Calculate the Euclidean distance of each candidate non-dominated solution to the positive ideal solution and the negative ideal solution;

[0184] Step 12: Calculate relative proximity;

[0185] Step 13: Sort the candidate non-dominated solutions according to their relative proximity, and select the candidate non-dominated solution with the highest relative proximity as the compromise solution.

[0186] The above is an explanation of the production scheduling optimization method based on multi-objective hybrid according to the embodiment of the present invention. It can be recognized that, compared with the production scheduling method in the prior art, the embodiment of the present invention is based on a hybrid meta-heuristic algorithm, and proposes a multi-objective optimization framework for realizing the production scheduling optimization method as described above, which can optimize multiple production targets at the same time, and there is no need to pre-set the target weights, but automatically find and present a series of non-dominated solutions through the Pareto optimal concept, and then select the compromise solution in the non-dominated solution set as the production scheduling optimization result through the approximate ideal solution sorting method, which can reduce the subjectivity of weight setting, and enable decision makers to flexibly choose according to actual needs, without repeated trial and error, thereby improving the production efficiency and intelligence level of automobile production scheduling.

[0187] Reference Figure 3 The embodiment of the present invention further provides a production scheduling optimization system based on multi-objective mixing, including:

[0188] An objective function design module, used to design a first objective function for minimizing the late shift subsidy and a second objective function for minimizing the energy consumption cost, and determine constraint conditions;

[0189] A particle evaluation module is used to initialize the particle swarm and calculate the fitness value of each particle in the particle swarm corresponding to the first objective function and the second objective function;

[0190] An external repository update module is used to store the non-dominated solutions of each particle through the external repository, calculate the entropy change value, probability threshold, probability value and random value of the external repository, and determine the global best particle according to the probability value and random value;

[0191] The particle update module is used to update each particle according to the global best particle, entropy change value, probability threshold and constraint conditions. When the preset termination condition is reached, the non-dominated solution corresponding to the updated particle is stored in the external storage to obtain the non-dominated solution set;

[0192] The compromise solution selection module is used to select the compromise solution in the non-dominated solution set as the production scheduling optimization result according to the fitness value by using the approximate ideal solution sorting method.

[0193] The contents of the above-mentioned production scheduling optimization method embodiment based on multi-objective mixing are all applicable to the present production scheduling optimization system embodiment based on multi-objective mixing. The functions specifically implemented by the present production scheduling optimization system embodiment based on multi-objective mixing are the same as those of the above-mentioned production scheduling optimization method embodiment based on multi-objective mixing, and the beneficial effects achieved are also the same as those achieved by the above-mentioned production scheduling optimization method embodiment based on multi-objective mixing.

[0194] The embodiment of the present invention further provides an electronic device, the electronic device comprising: a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for realizing connection and communication between the processor and the memory, and when the program is executed by the processor, the above-mentioned production scheduling optimization method based on multi-objective mixing is realized. The electronic device can be any intelligent terminal including a tablet computer, a vehicle-mounted computer, etc.

[0195] like Figure 4 FIG. 1 is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention. Figure 4 , an embodiment of the present invention provides an electronic device, including:

[0196] The processor 1001 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present invention.

[0197] The memory 1002 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 1002 can store an operating system and other application programs. When the technical solution provided in the embodiments of this specification is implemented by software or firmware, the relevant program code is stored in the memory 1002, and the processor 1001 calls and executes the production scheduling optimization method based on multi-objective mixing of the embodiment of the present invention;

[0198] Input / output interface 1003, used to implement information input and output;

[0199] The communication interface 1004 is used to realize the communication interaction between the device and other devices. The communication can be realized through a wired manner (such as USB, network cable, etc.) or a wireless manner (such as mobile network, WIFI, Bluetooth, etc.);

[0200] A bus 1005 , which transmits information between various components of the device (e.g., the processor 1001 , the memory 1002 , the input / output interface 1003 , and the communication interface 1004 );

[0201] The processor 1001 , the memory 1002 , the input / output interface 1003 and the communication interface 1004 are connected to each other in communication within the device via the bus 1005 .

[0202] An embodiment of the present invention also provides a storage medium, which is a computer-readable storage medium used for computer-readable storage. The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the above-mentioned production scheduling optimization method based on multi-objective mixing.

[0203] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0204] The embodiment of the present invention also discloses a computer program product or a computer program, wherein the computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes Figure 1 The method shown.

[0205] In some selectable embodiments, the function / operation mentioned in the block diagram may not occur in the order mentioned in the operation diagram. For example, depending on the function / operation involved, the two boxes shown in succession can actually be executed substantially simultaneously or the above-mentioned boxes can sometimes be executed in reverse order. In addition, the embodiment presented and described in the flow chart of the present invention is provided by way of example, for the purpose of providing a more comprehensive understanding of technology. The disclosed method is not limited to the operation and logic flow presented herein. Selectable embodiments are expected, wherein the order of various operations is changed and the sub-operation of a part for which is described as a larger operation is performed independently.

[0206] In addition, although the present invention is described in the context of functional modules, it should be understood that, unless otherwise specified to the contrary, one or more of the above-mentioned functions and / or features can be integrated into a single physical device and / or software module, or one or more functions and / or features can be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the present invention. More specifically, in view of the properties, functions and internal relationships of the various functional modules in the device disclosed herein, the actual implementation of the module will be understood within the conventional skills of the engineer. Therefore, those skilled in the art can implement the present invention set forth in the claims without excessive experimentation using ordinary techniques. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present invention, which is determined by the full scope of the appended claims and their equivalents.

[0207] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the above methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0208] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0209] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0210] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

[0211] The above is a specific description of the preferred implementation of the present invention, but the present invention is not limited to the embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A production scheduling optimization method based on multi-objective mixing, characterized in that: The following steps are involved: Design a first objective function for minimizing the late shift subsidy and a second objective function for minimizing the energy consumption cost, and determine the constraints; Initializing a particle swarm, and calculating a fitness value of each particle in the particle swarm corresponding to the first objective function and the second objective function; The non-dominated solutions of each particle are stored in an external repository, and an entropy change value, a probability threshold, a probability value, and a random value of the external repository are calculated, and a global optimal particle is determined according to the probability value and the random value; updating each of the particles according to the global best particle, the entropy change value, the probability threshold and the constraint condition, and when a preset termination condition is reached, storing the non-dominated solution corresponding to the updated particle in the external storage repository to obtain a non-dominated solution set; A compromise solution in the non-dominated solution set is selected as a production scheduling optimization result according to the fitness value through an approximate ideal solution sorting method.

2. The production scheduling optimization method based on multi-objective mixing according to claim 1 is characterized in that: The first objective function designed to minimize the night shift subsidy and the second objective function designed to minimize the energy consumption cost specifically include: Obtain target output, number of overtime workers, subsidy time period, energy consumption per shift, production status in each preset time period, and peak-valley electricity prices corresponding to each peak and valley period, including the electricity prices corresponding to peak time period, normal time period, and valley time period; Setting a production state array, and obtaining a cost array according to the peak and valley electricity prices and the subsidy time period; Obtaining the first objective function according to the number of overtime workers, the production status array, and the cost array; The second objective function is obtained according to the target output, the peak-valley electricity price, the energy consumption of a single vehicle per shift, and the production status in each preset time period.

3. The production scheduling optimization method based on multi-objective mixing according to claim 1 is characterized in that: The determining constraint conditions specifically include: Determine the start time and end time, and obtain the target production time; The constraint condition is determined to be that the sum of the production states of each time period from the start time to the end time is equal to the target output production time.

4. The production scheduling optimization method based on multi-objective mixing according to claim 1 is characterized in that: The initialization of the particle swarm specifically includes: Setting the initial position and initial velocity of each particle, and determining the individual extreme value and group extreme value of each particle; Determine the inertia weight and acceleration constant.

5. The production scheduling optimization method based on multi-objective mixing according to claim 4 is characterized in that: The storing of the non-dominated solutions of the particles in the external repository and calculating the entropy change value, probability threshold, probability value and random value of the external repository specifically includes: Initializing the external storage repository, and storing each of the non-dominated solutions in the external storage repository; Representing the position of each non-dominated solution in the target space by the fitness value corresponding to each particle, and obtaining the coordinate value of each non-dominated solution; Calculate the potential value and density value of each of the non-dominated solutions according to the coordinate values; The entropy change value, probability threshold, probability value and random value of the external storage repository are calculated.

6. The production scheduling optimization method based on multi-objective mixing according to claim 5 is characterized in that: Determining the global best particle according to the probability value and the random value specifically includes: Determine a minimum potential value according to each of the potential values, and determine a minimum density value according to each of the density values; When the random value is greater than the probability value, determining the particle corresponding to the minimum potential value as the global best particle; When the random value is greater than the probability value, the particle corresponding to the minimum density value is determined as the global best particle.

7. The production scheduling optimization method based on multi-objective mixing according to claim 5 is characterized in that: The updating of each particle according to the global best particle, the entropy change value, the probability threshold and the constraint condition specifically includes: When the absolute value of the entropy change value is greater than or equal to the probability threshold, each of the particles is updated according to the global best particle and the constraint condition by using a multi-objective particle swarm optimization algorithm; When the absolute value of the entropy change value is less than the probability threshold, each particle is updated according to the global best particle and the constraint condition by a multi-objective evolutionary algorithm.

8. The production scheduling optimization method based on multi-objective mixing according to claim 7 is characterized in that: The updating of each particle according to the global best particle and the constraint condition by using the multi-objective particle swarm optimization algorithm specifically includes: Update the initial velocity of each particle according to the global best particle, the individual extreme value, the group extreme value, the inertia weight, and the acceleration constant; Updating the initial position of each particle according to the initial position and the updated initial velocity; Wherein, the updated initial position satisfies the constraint condition.

9. The production scheduling optimization method based on multi-objective mixing according to claim 1 is characterized in that: The non-dominated solution set includes a plurality of candidate non-dominated solutions, and the method of selecting a compromise solution in the non-dominated solution set as a production scheduling optimization result according to the fitness value by using the approximate ideal solution sorting method specifically includes: Normalizing the fitness values ​​corresponding to the candidate non-dominated solutions to obtain a normalized decision matrix; Determining weights of the first objective function and the second objective function, and obtaining a weighted normalized decision matrix according to the weights and the normalized decision matrix; A positive ideal solution and a negative ideal solution are calculated according to the fitness value; Calculating the Euclidean distance from each candidate non-dominated solution to the positive ideal solution to obtain a first Euclidean distance, and calculating the Euclidean distance from each candidate non-dominated solution to the negative ideal solution to obtain a second Euclidean distance; Calculating a relative proximity of each of the candidate non-dominated solutions according to the first Euclidean distance and the second Euclidean distance; The candidate non-dominated solution corresponding to the largest relative proximity is determined as a compromise solution, and the production scheduling optimization result is obtained.

10. A production scheduling optimization system based on multi-objective mixing, characterized in that: include: An objective function design module, used to design a first objective function for minimizing the late shift subsidy and a second objective function for minimizing the energy consumption cost, and determine constraint conditions; A particle evaluation module, used for initializing a particle swarm and calculating a fitness value of each particle in the particle swarm corresponding to the first objective function and the second objective function; An external repository updating module, used for storing the non-dominated solutions of each particle through an external repository, calculating the entropy change value, probability threshold, probability value and random value of the external repository, and determining the global best particle according to the probability value and the random value; A particle updating module, used for updating each of the particles according to the global best particle, the entropy change value, the probability threshold and the constraint condition, and when a preset termination condition is reached, storing the non-dominated solution corresponding to the updated particle in the external storage repository to obtain a non-dominated solution set; The compromise solution selection module is used to select a compromise solution in the non-dominated solution set as a production scheduling optimization result according to the fitness value by using an approximate ideal solution sorting method.

Images