Flow shop energy consumption optimization scheduling method and system based on improved genetic algorithm

Abstract

The invention discloses a flow shop energy consumption optimization scheduling method based on an improved genetic algorithm. The method comprises the following steps: S1, obtaining a plurality of orders, and obtaining a corresponding first scheduling scheme, a second scheduling scheme and a third scheduling scheme based on an improved LR rule, a total power consumption minimum principle and a random generation principle; s2, obtaining corresponding chromosomes based on corresponding scheduling schemes, and generating a corresponding first population, a second population and a third population; s3, genetic manipulation is carried out based on each population to generate corresponding new chromosomes, and corresponding updated populations are obtained; s4, based on all new chromosomes, selecting according to a rule to obtain a best solution; s5, searching based on the best solution to obtain an optimal solution, and adding the optimal solution to all the updated populations to obtain corresponding final populations; s6, performing chromosome exchange according to corresponding objective functions based on all final populations to obtain corresponding non-dominated solutions; and S7, obtaining a final non-dominated solution set based on all non-dominated solutions. And the production efficiency and the resource utilization rate of the whole flow workshop are improved.

Description

Technical Field

[0001] The invention relates to the technical field of workshop energy consumption optimization scheduling, and more specifically to a flow workshop energy consumption optimization scheduling method and system based on an improved genetic algorithm. Background Art

[0002] At present, with the continuous development of industrial modernization and intelligence, economic benefits and environmental benefits have become a comprehensive consideration for enterprises and the country, and have become a basic requirement. Especially in the production process of electrode foil, the formation process is particularly critical. Because it involves frequent charging and discharging processes, it usually accounts for 30%-40% of the total power consumption of the factory. The energy consumption is extremely large. The huge energy consumption problem has a great impact on the energy consumption and cost of the entire factory. Therefore, the use of intelligent algorithms to optimize production scheduling is particularly important for accurately controlling energy consumption, reducing production costs, and promoting the sustainable development of the manufacturing industry.

[0003] Scheduling of electrode foil forming flow shop is a permutation flow shop scheduling problem. Chen Shuilin et al. proposed a multi-objective grey wolf optimization algorithm (MOGWO) to optimize the completion time and total delay of permutation flow shop scheduling; Li Haoran et al. proposed a multi-objective discrete artificial bee colony algorithm with the maximum completion time and total delay as the goal; Su Jiantao et al. quickly solved the rescheduling problem of machine failure by integrating multiple rescheduling methods and using equipment failure simulation to verify the rescheduling method, effectively solving the machine failure problem of mixed flow shop; Ruiz R et al. optimized the processing completion time of the flow shop by using the hyper-heuristic algorithm. The above existing technologies take the maximum completion time, total delay, etc. as the optimization objectives, while relatively few studies consider energy consumption and shutdown conversion losses; at the same time, the existing technologies use a relatively simple initialization strategy and use a randomly generated initialization method; and most algorithms are only applied to the theoretical level of scheduling problems, without considering practical problems such as shutdown switching losses in the actual production process.

[0004] Therefore, how to ensure the stability and applicability of the entire assembly line workshop in the actual workshop production scheduling problem is an urgent problem that technical personnel in this field need to solve. Summary of the invention

[0005] In view of this, the present invention provides a method and system for optimizing energy consumption scheduling of an assembly line workshop based on an improved genetic algorithm, which solves the problems of low search efficiency, single solution population and poor convergence of most algorithms in assembly line workshop scheduling, thereby improving the production efficiency and resource utilization of the entire assembly line workshop.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] The energy consumption optimization scheduling method of flow shop based on improved genetic algorithm includes:

[0008] S1 obtains multiple orders of the target flow shop and obtains the corresponding first scheduling plan, second scheduling plan and third scheduling plan based on the improved LR rule, the principle of minimum total power consumption and the principle of random generation;

[0009] S2 obtains the corresponding chromosome based on the corresponding scheduling scheme and generates the corresponding first population, second population and third population;

[0010] S3 performs genetic operations on each population to generate corresponding new chromosomes and obtain corresponding updated populations;

[0011] S4 selects two of the new chromosomes as the best solutions according to the rules based on the new chromosomes generated by all populations;

[0012] S5 obtains the optimal solution based on the best solution search and adds it to all the update populations respectively to obtain the corresponding final population;

[0013] S6 performs chromosome exchange based on all the final populations according to the corresponding objective function to obtain corresponding non-dominated solutions;

[0014] S7 obtains a final non-dominated solution set based on all the non-dominated solutions.

[0015] Preferably, the first scheduling scheme is obtained based on the improved LR rule, specifically including:

[0016] Calculate the completion time of each order on each production line;

[0017] Get a comprehensive score based on the stated completion time;

[0018] Sort all the orders in descending order based on the comprehensive scores, select the first R orders and distribute them to the production lines of R different voltage partitions in sequence;

[0019] The remaining orders calculate their index values ​​with the production line;

[0020] Select the order with the smallest index value and assign it to the production line with the shortest average completion time.

[0021] The above process is repeated until all remaining orders are assigned to corresponding production lines, thereby obtaining the first scheduling solution based on the improved LR rule.

[0022] Preferably, the second scheduling scheme based on the principle of minimum total power consumption specifically includes:

[0023] Calculate the corresponding production time based on all the orders and their corresponding maximum vehicle speeds;

[0024] Sort all the orders in descending order based on the production time;

[0025] Based on the sorting results, the first two orders are selected and preferentially assigned to the production line with the lowest power consumption;

[0026] The remaining orders are allocated to the corresponding production lines so that the remaining orders are processed on the corresponding production lines with the minimum additional power consumption, thereby obtaining the second scheduling plan based on the principle of minimum total power consumption.

[0027] Preferably, S2 specifically includes:

[0028] Arrange the corresponding production line identifiers based on the order of the corresponding production lines in the first scheduling scheme, the second scheduling scheme, and the third scheduling scheme, and obtain a first arrangement result, a second arrangement result, and a third arrangement result accordingly;

[0029] Based on the first arrangement result, the second arrangement result and the third arrangement result, the first chromosome, the second chromosome and the third chromosome are correspondingly used;

[0030] The first population, the second population and the third population are correspondingly obtained based on the first chromosome, the second chromosome and the third chromosome.

[0031] Preferably, performing genetic operations on each population to generate corresponding new chromosomes specifically includes:

[0032] performing the genetic operation simultaneously based on the first population, the second population and the third population respectively;

[0033] The genetic operation includes: selection operation, crossover operation and mutation operation;

[0034] The selection operation is: randomly selecting two chromosomes from the corresponding population as selected chromosomes; selecting a selected chromosome with a low non-dominated level and a large crowding distance as a selected chromosome based on the selected chromosomes;

[0035] The crossover operation is as follows: based on the selected chromosome, a plurality of orders for production line crossover are randomly selected as crossover orders; based on the first position of each crossover order on the parent generation P1, the second position of the production line on the selected chromosome is found on the parent generation P2, and the first position and the second position are exchanged until all the crossover orders are crossed to generate corresponding crossover chromosomes;

[0036] The mutation operation is: randomly selecting two orders from the orders with position exchange in the cross chromosome as mutation orders; randomly selecting a production line other than the original processing line from the optional production lines of the mutation order to replace it, and another production line with electricity consumption less than the set value selected from the other mutation order to replace the original production line, to obtain the corresponding mutation chromosome;

[0037] The selected chromosome and the variant chromosome are used together as the new chromosome.

[0038] Preferably, the best solution acquisition method is:

[0039] Based on the new chromosomes in all updated populations, two new chromosomes are selected as the best solutions based on a comprehensive objective function;

[0040] The comprehensive objective function includes: minimizing the maximum completion time, minimizing the total power consumption and minimizing the shutdown switching loss.

[0041] Preferably, the optimal solution acquisition method is:

[0042] S41 initializes the taboo table and sets the best solution as the current solution and the initial global optimal solution;

[0043] S42 selects a plurality of candidate solutions based on all domain solutions of the current solution;

[0044] S43 determines whether the candidate solution is better than the initial global optimal solution. If so, the candidate solution is used as a selected candidate solution and added to the taboo table; otherwise, the optimal solution is selected from all the candidate solutions that are not tabooed as a selected candidate solution and added to the taboo table;

[0045] S44 determines whether all the candidate solutions are in the taboo table. If so, the best selected candidate solution is selected in the taboo table as the optimal solution; otherwise, returns to S42.

[0046] Preferably, S6 specifically includes:

[0047] Based on the chromosomes in the first final population, the second final population, and the third final population, calculations are performed based on minimizing completion time, minimizing total power consumption, and minimizing shutdown switching losses, respectively, to obtain corresponding first optimization target values, second optimization target values, and third optimization target values;

[0048] Sort the chromosomes in the corresponding final population in ascending order based on the corresponding optimization target value;

[0049] Based on the sorting results, the first quarter of chromosomes in each final population are transferred as transfer individuals to the other two final populations, and the corresponding fusion populations are obtained;

[0050] Performing a fast non-dominated sorting based on the fused population to obtain a first hierarchical result;

[0051] The top solution based on the first stratification result is defined as the upper elite solution;

[0052] Based on the chromosomes in the fusion population that have not entered the upper elite solution as discarded individuals, a corresponding discarded population is generated;

[0053] Perform non-dominated sorting based on the discarded population to obtain a second stratification result;

[0054] Based on the second stratification result, the top chromosome is selected as the priority solution;

[0055] Performing non-dominated operations based on the priority solution and the upper elite solution and replacing inferior solutions, obtaining the corresponding non-dominated solution;

[0056] The above process is repeated to obtain a first non-dominated solution, a second non-dominated solution and a third non-dominated solution based on the first final population, the second final population and the third final population respectively.

[0057] Preferably, S7 specifically includes:

[0058] Perform search and replacement based on the first non-dominated solution, the second non-dominated solution, and the third non-dominated solution to update a non-dominated solution set;

[0059] Calculating a fitness value based on the non-dominated solution set and the comprehensive objective function;

[0060] Determining whether the fitness value is greater than the population fitness value;

[0061] If yes, output the non-dominated solution set as the final non-dominated solution set;

[0062] Otherwise return to S3.

[0063] The flow shop energy consumption optimization scheduling system based on improved genetic algorithm includes: scheme generation module, population generation module, population update module, best solution output module, final population generation module and result output module;

[0064] The scheme generation module is used to obtain multiple orders of the target flow shop and obtain the corresponding first scheduling scheme, second scheduling scheme and third scheduling scheme based on the improved LR rule, the total power consumption minimum principle and the random generation principle;

[0065] The population generation module is used to obtain the corresponding chromosome based on the corresponding scheduling scheme and generate the corresponding first population, second population and third population;

[0066] The population updating module is used to perform genetic operations on each population to generate corresponding new chromosomes and obtain corresponding updated populations;

[0067] The best solution output module is used to select two of the new chromosomes as the best solutions according to the rules based on the new chromosomes generated by all populations;

[0068] The final population generation module is used to obtain the optimal solution based on the best solution search and add it to all the updated populations respectively to obtain the corresponding final population;

[0069] The result output module is used to perform chromosome exchange based on all the final populations according to the corresponding objective function to obtain corresponding non-dominated solutions; and obtain a final non-dominated solution set based on all the non-dominated solutions.

[0070] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a method and system for optimizing the energy consumption of a flow shop based on an improved genetic algorithm, which has the following beneficial effects:

[0071] 1. This method adopts a diversified initialization strategy, using the improved LR rule and the minimum total power consumption to generate two relatively good individuals and use the two individuals to make the population explore the direction of the optimal solution. The remaining individuals are all randomly generated. This helps to cover the solution space more comprehensively, reduce the risk of the algorithm falling into the local optimum, and ensure the diversity of the population, thereby improving the quality of the initial solution and the quality of the population initialization, making the final solution set more diverse and more convergent.

[0072] 2. This method adopts the taboo search strategy to jump out of the local optimum and explore more solution spaces, thereby improving the global optimization effect of the algorithm; in addition, the algorithm incorporates the taboo search strategy as a mechanism to strengthen local exploration, allowing in-depth exploration of each member of the group; this strategy relies on its unique taboo principle to screen and adopt new solutions, which helps to break through the limitations of local optimality and enhance the global exploration ability of the algorithm. It improves the global search ability of the algorithm in complex solution space, enhances the depth of local exploration, and at the same time enhances the quality and efficiency of the algorithm's solution.

[0073] 3. This method uses the population coordination mechanism to balance the three optimization objectives, and combines it with the improved elite strategy. The poor solutions discarded in the fused population are cross-mutated to obtain a new population for fast non-dominated sorting. The solutions in the upper layer of the new population are non-dominated with the selected elite solutions to replace the poor solutions and retain the better individuals. Each solution balances the other two objectives while optimizing the current goal, thereby improving the production efficiency and resource utilization of the entire assembly line, making the solved scheduling plan more in line with the actual situation of the enterprise. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0075] Figure 1 A flow chart of the method for optimizing energy consumption of a flow shop based on an improved genetic algorithm provided by the present invention.

[0076] Figure 2 A flow chart of a method for obtaining a first scheduling solution based on an improved LR rule provided by the present invention.

[0077] Figure 3 A schematic diagram of encoding and decoding provided by the present invention.

[0078] Figure 4 A schematic diagram of the crossover operation provided by the present invention.

[0079] Figure 5 This is a schematic diagram of the variation operation provided by the present invention.

[0080] Figure 6 This is a flow chart of the optimal solution acquisition method provided by the present invention.

[0081] Figure 7 A schematic diagram of the structure of the flow shop energy consumption optimization scheduling system based on the improved genetic algorithm provided by the present invention. DETAILED DESCRIPTION

[0082] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0083] Example 1

[0084] like Figure 1 As shown, the embodiment of the present invention discloses a method for optimizing energy consumption of a flow shop based on an improved genetic algorithm, comprising:

[0085] S1 obtains multiple orders of the target flow shop and obtains the corresponding first scheduling plan, second scheduling plan and third scheduling plan based on the improved LR rule, the principle of minimum total power consumption and the principle of random generation;

[0086] S2 obtains the corresponding chromosome based on the corresponding scheduling scheme and generates the corresponding first population, second population and third population;

[0087] S3 performs genetic operations on each population to generate corresponding new chromosomes and obtain corresponding updated populations;

[0088] S4 selects two new chromosomes as the best solution based on the new chromosomes generated by all populations according to the rules;

[0089] S5 obtains the optimal solution based on the best solution search and adds it to all update populations to obtain the corresponding final population;

[0090] S6 performs chromosome exchange based on all final populations according to the corresponding objective function to obtain the corresponding non-dominated solution;

[0091] S7 obtains the final non-dominated solution set based on all non-dominated solutions.

[0092] Example 2

[0093] The embodiment of the present invention discloses a method for optimizing energy consumption of a flow shop based on an improved genetic algorithm, comprising:

[0094] A mathematical model for optimizing the energy consumption of assembly lines is established for subsequent optimization of the energy consumption of assembly lines. The mathematical model for optimizing the energy consumption of assembly lines includes three objective functions and multiple constraints.

[0095] Preferably, the three objective functions include: minimizing the maximum completion time F1, minimizing the total power consumption F2, and minimizing the shutdown switching loss F3:

[0096] F1=min{C max}

[0097]

[0098] Among them, C max represents the maximum completion time of all orders, E ij S is the electricity consumption per unit area of ​​order j on production line i; ij is the downtime switching loss of order j on production line i.

[0099] Preferably, the plurality of constraints include:

[0100] The first constraint is:

[0101]

[0102] Among them, C v,i,jrepresents the completion time of order j on production line i in voltage partition v; the first constraint is used to constrain the completion time of each order to never exceed the maximum completion time of the total order.

[0103] Second constraint:

[0104]

[0105] The third constraint is:

[0106]

[0107] Among them, X v,j,k Indicates whether order j is processed at position k in voltage partition v. If X v,j,k =1 means processing at this position, X v,j,k = 0 means not to process at this position. The second constraint and the third constraint together restrict that each order can only be processed at one position in one voltage zone at the same time.

[0108] The fourth constraint is:

[0109]

[0110] Among them, Y v,i,h,j Indicates whether order h is located before order j on production line i in voltage partition v. If Y v,i,h,j =1 means that order h comes before order j; otherwise, it means that order h does not come before order j.

[0111] The fifth constraint is:

[0112]

[0113] Among them, Y v,i,h+1,j Indicates whether order j is after order h+1 on production line i in voltage partition v. If Y v,i,h+1,j =1, it means that order j is after order h+1, otherwise order j is not after order h+1. The fourth constraint and the fifth constraint together ensure that each job can be a subsequent job or a predecessor job in the factory assigned to it.

[0114] The sixth constraint is:

[0115]

[0116] Among them, Y v,i,1,j It represents the preparation time caused by machine downtime for order j at the first position of production line i in voltage partition v. The sixth constraint is used to ensure that the orders at the first position of each voltage do not involve preparation time caused by machine downtime.

[0117] The seventh constraint:

[0118]

[0119] Among them, C v,i,k represents the processing completion time at the kth position on the production line i in the voltage partition v; C v,i,k-1 is the processing completion time at the k-1th position on production line i in voltage partition v; SP i is the processing speed of production line i; p i,j is the processing time of order j on production line i. The seventh constraint is used to constrain the completion time of the production line without idle time.

[0120] The eighth constraint is:

[0121]

[0122] Where PEC represents the total power consumption in the production process; E i It represents the power consumption per unit time when production line i produces an order. The eighth constraint is used to constrain the total power consumption PEC in the production process.

[0123] Ninth constraint:

[0124] S(J c ,J d )=k1×T switch (J c ,J d )+k2×C fixed

[0125] Among them, S(J c ,J d ) represents the loss when the production of order c and order d is switched; k1 and k2 represent the downtime loss coefficient and the downtime switching fixed loss coefficient respectively; T switch (J c ,J d ) represents the time when order c and order d are switched. fixed It represents the fixed loss when the order is shut down for switching. The ninth constraint is used to constrain the loss caused when the production of two orders is switched.

[0126] Tenth constraint:

[0127]

[0128] Among them, SEC represents the total loss caused by order switching during the production process. represents the switching loss when order c and order d are produced on production line i. The tenth constraint is used to constrain the total loss caused by order switching.

[0129] Eleventh constraint:

[0130]

[0131] Among them, C v,i,j represents the processing completion time of the jth order on production line i in voltage partition v. The eleventh constraint is used to constrain the completion time of each order to be a non-negative value.

[0132] The twelfth constraint is:

[0133]

[0134] Thirteenth constraint:

[0135]

[0136] The twelfth constraint and the thirteenth constraint jointly define the value range of the decision variable.

[0137] S1 obtains multiple orders of the target assembly line and obtains the corresponding first scheduling plan, second scheduling plan and third scheduling plan based on the improved LR rule, the principle of minimum total power consumption and the principle of random generation.

[0138] Preferably, set the IMOGA algorithm parameters: population size N, crossover probability p c , mutation probability p m As well as the taboo table and the length of the taboo table. Based on the set algorithm parameters, the population is initialized for multiple targets to obtain three initial populations with population sizes of N1, N2, and N3.

[0139] Preferably, Figure 2 As shown, the first scheduling scheme is obtained based on the improved LR rule, which specifically includes:

[0140] Calculate the completion time of each order on each production line;

[0141] Get a comprehensive score based on completion time;

[0142] All orders are sorted in descending order based on the comprehensive scores, and the first R orders are selected and allocated to the production lines of R different voltage partitions in turn;

[0143] The remaining orders calculate their index values ​​with the production line;

[0144] Select the order with the smallest index value and assign it to the production line with the shortest average completion time.

[0145] Repeat the above process until all remaining orders are assigned to the corresponding production lines, and obtain the first scheduling solution based on the improved LR rule.

[0146] Preferably, the comprehensive score based on the completion time specifically includes:

[0147] The time mean, time standard deviation and time skewness are calculated based on the completion time, and the time mean, time standard deviation and time skewness are converted to the same dimension and summed up as the comprehensive score of the order. A high score represents a more complex or time-consuming order.

[0148] Preferably, the index value calculation formula is:

[0149]

[0150] in, Indicates the index value of order j processed in voltage partition g*, represents the average completion time of order j in voltage partition g*, c represents a constant, c = n / r-2, n represents the number of orders, r represents the number of voltage partitions; n g* Indicates the number of orders in the current voltage partition; [n g * ] indicates position n g* The corresponding order; C m,j represents the completion time of order j on production line m; t i,j represents the completion time of order j in voltage partition g*; C i-1,j represents the completion time of order j on production line i-1, C i ,[n g* ] indicates order [n g* ] is the completion time on production line i, i represents the production line number, and m represents a production line in production line i[1,l].

[0151] Preferably, the second scheduling scheme based on the principle of minimum total power consumption specifically includes:

[0152] Calculate the corresponding production time based on all orders and their corresponding maximum speed;

[0153] Sort all orders in descending order based on production time;

[0154] Based on the sorting results, the first two orders are selected and preferentially assigned to the production line with the lowest power consumption;

[0155] The remaining orders are allocated to the corresponding production lines so that the remaining orders can be processed on the corresponding production lines with the minimum additional power consumption, thereby obtaining a second scheduling plan based on the principle of minimum total power consumption.

[0156] Preferably, the random generation principle is: all orders are randomly assigned to different production lines.

[0157] S2 obtains the corresponding chromosome based on the corresponding scheduling scheme and generates the corresponding first population, second population and third population.

[0158] Preferably, S2 specifically includes:

[0159] Arrange the corresponding production line identifiers based on the corresponding production line sequences in the first scheduling scheme, the second scheduling scheme, and the third scheduling scheme, and obtain a first arrangement result, a second arrangement result, and a third arrangement result accordingly;

[0160] Based on the first arrangement result, the second arrangement result and the third arrangement result, the first chromosome, the second chromosome and the third chromosome are used accordingly;

[0161] A first population, a second population and a third population are obtained based on the first chromosome, the second chromosome and the third chromosome respectively.

[0162] Preferably, a small portion of the orders are adjusted and changed based on the first chromosome and the second chromosome to generate two dominant populations: a first population N1 and a second population N2 respectively; and a plurality of different third chromosomes are generated based on the random generation principle to obtain a third population N3.

[0163] Preferably, Figure 3 As shown, each chromosome individual represents a scheduling plan. The chromosome individual is decoded, the production line sequence is matched with the order sequence, and each order is processed by the corresponding production line.

[0164] Preferably, the present invention adopts a variety of initialization strategies to improve the diversity of the initial population, which helps to cover the solution space more comprehensively and reduce the risk of the algorithm falling into a local optimum.

[0165] S3 performs genetic operations on each population to generate corresponding new chromosomes and obtain corresponding updated populations.

[0166] Preferably, performing genetic operations on each population to generate corresponding new chromosomes specifically includes:

[0167] Genetic operations are performed simultaneously based on the first population, the second population, and the third population;

[0168] Genetic operations include: selection operation, crossover operation and mutation operation;

[0169] The selection operation is: randomly selecting two chromosomes from the first population, the second population or the third population as selected chromosomes; selecting a selected chromosome with a low non-dominated level and a large crowding distance as a selected chromosome based on the selected chromosome;

[0170] like Figure 4 As shown, the crossover operation is as follows: based on the selected chromosome, a plurality of orders for production line crossover are randomly selected as crossover orders; based on the first position of each crossover order on the parent generation P1, the second position of the production line on the selected chromosome is found on the parent generation P2, and the first position and the second position are exchanged until all crossover orders are completed and the corresponding crossover chromosome is generated;

[0171] like Figure 5 As shown, the mutation operation is: randomly select two orders from the orders with position exchange in the cross chromosome as mutation orders; randomly select a production line that is not originally processed from the optional production lines based on the mutation order to replace it, and another mutation order selects a production line with a power consumption less than the set value to replace the original production line, and obtains the corresponding mutation chromosome;

[0172] The selected chromosome and the mutated chromosome are used together as new chromosomes.

[0173] Preferably, in this embodiment, three different positions are selected on the two parents respectively.

[0174] Preferably, the first population, the second population and the third population are all subjected to genetic operations simultaneously, and the first new chromosome, the second new chromosome and the third new chromosome are generated correspondingly, thereby obtaining the first updated population, the second updated population and the third updated population.

[0175] S4 selects two new chromosomes as the best solution according to the rules based on the new chromosomes generated by all populations.

[0176] Preferably, the best solution acquisition method is:

[0177] Based on all the new chromosomes in the updated population, two new chromosomes X1 and X2 are selected as the best solutions based on the comprehensive objective function;

[0178] The comprehensive objective function includes: minimizing the maximum completion time F1, minimizing the total power consumption F2 and minimizing the shutdown switching loss F3.

[0179] Preferably, a comprehensive evaluation is performed based on the first new chromosome, the second new chromosome and the third new chromosome according to the objective functions F1, F2 and F3, and they are sorted in descending order based on the evaluation values, and the first two new chromosomes X1 and X2 are selected as the best solutions.

[0180] S5 obtains the optimal solution based on the best solution search and adds it to all update populations to obtain the corresponding final population.

[0181] Preferably, Figure 6 As shown, the optimal solution is obtained as follows:

[0182] S41 initializes the taboo table and sets the best solution as the current solution and the initial global optimal solution;

[0183] S42 selects multiple candidate solutions based on all domain solutions of the current solution;

[0184] S43 determines whether the candidate solution is better than the initial global optimal solution. If so, the candidate solution is used as a selected candidate solution and added to the taboo table; otherwise, the optimal solution is selected from all the candidate solutions that have not been tabooed as a selected candidate solution and added to the taboo table;

[0185] S44 determines whether all candidate solutions are in the taboo table. If so, the best selected candidate solution is selected in the taboo table as the optimal solution; otherwise, returns to S42.

[0186] Preferably, the present invention adopts a taboo search method and searches for a better solution based on a neighborhood structure and a contempt criterion.

[0187] Preferably, the three objective function values ​​are comprehensively considered to determine whether the candidate solution is better than the initial global optimal solution.

[0188] Preferably, even if the current solution is equal to the global optimal solution, the search cannot be stopped. On the one hand, the non-tabooed neighborhood should be continuously explored, and on the other hand, falling into the local optimal solution should be avoided.

[0189] S6 performs chromosome exchange based on all final populations according to the corresponding objective function to obtain the corresponding non-dominated solution.

[0190] Preferably, S6 specifically includes:

[0191] Based on the chromosomes in the first final population, the second final population, and the third final population, calculations are performed based on minimizing completion time, minimizing total power consumption, and minimizing shutdown switching losses, respectively, to obtain corresponding first optimization target values, second optimization target values, and third optimization target values;

[0192] Sort the chromosomes in the corresponding final population in ascending order based on the corresponding optimization target value;

[0193] Based on the sorting results, the first quarter of chromosomes in each final population are transferred as transfer individuals to the other two final populations, and the corresponding fusion populations are obtained;

[0194] The first hierarchical result is obtained by fast non-dominated sorting based on the fused population;

[0195] The top solution based on the first stratification results is defined as the upper elite solution;

[0196] Chromosomes that have not entered the upper elite solution in the fusion population are used as discarded individuals and the corresponding discarded population is generated;

[0197] Based on the discarded population, non-dominated sorting is performed to obtain the second stratification result;

[0198] Based on the second layer result, the top chromosome is selected as the priority solution;

[0199] Based on the priority solution and the upper elite solution, non-dominated operations are performed and inferior solutions are replaced to obtain the corresponding non-dominated solutions;

[0200] The above process is repeated to obtain a first non-dominated solution, a second non-dominated solution and a third non-dominated solution based on the first final population, the second final population and the third final population respectively.

[0201] Preferably, the inferior solution is the chromosome that has not entered the upper elite solution.

[0202] Preferably, based on the sorting results, the first quarter of the chromosomes in each final population are transferred as transfer individuals to the other two final populations respectively: that is, the first final population, the second final population and the third final population obtain the first transfer individual, the second transfer individual and the third transfer individual respectively, the first final population receives the second transfer individual and the third transfer individual, the second final population receives the first transfer individual and the third transfer individual, and the third final population receives the first transfer individual and the second transfer individual.

[0203] Preferably, based on the sorting results, the last quarter of chromosomes in each final population are used as discarded individuals to maintain the stability of the population size, which correspondingly include: the first discarded individual, the second discarded individual and the third discarded individual, and crossover and mutation operations are performed based on the corresponding discarded individuals to generate the corresponding discarded population.

[0204] Preferably, non-dominated sorting is performed based on discarding the population, and the solutions in the population are divided into multiple levels according to the dominance relationship between the solutions: if solution 1 is not inferior to solution 2 in all objectives and is better than solution 2 in at least one objective, then solution 1 is said to dominate solution 2, and the solutions in the population are stratified, each layer contains a group of non-dominated solutions, and a stratified result is obtained.

[0205] Preferably, the first layer of the hierarchical results contains all solutions that are not dominated by any solution, the second layer contains solutions that are dominated only by the first layer of solutions, the third layer contains solutions that are dominated only by the first two layers of solutions, and so on. The higher the level, the more obvious the advantage of the solution.

[0206] S7 obtains the final non-dominated solution set based on all non-dominated solutions.

[0207] Preferably, S7 specifically includes:

[0208] Search and replace based on the first non-dominated solution, the second non-dominated solution and the third non-dominated solution to update the non-dominated solution set;

[0209] Calculate the fitness value based on the non-dominated solution set and the comprehensive objective function;

[0210] Determine whether the fitness value is greater than the population fitness value;

[0211] If so, output the non-dominated solution set as the final non-dominated solution set;

[0212] Otherwise return to S3.

[0213] Preferably, the final non-dominated solution set considers three objective functions at the same time. Each solution balances the other two objectives while optimizing the current objective, thereby improving the production efficiency and resource utilization of the entire assembly line.

[0214] Example 3

[0215] The method of the present invention verifies:

[0216] In order to verify the effectiveness of the present invention, the hypervolume metric (HV) and inverted generational distance (IGD) are selected as evaluation indicators, and the influence of population initialization and taboo search method on the performance of the algorithm are tested respectively. Finally, IMOGA is compared with multiple optimization algorithms, including MOGWO, INSGA-II and PSO-GA, which further verifies the effectiveness of the improved genetic algorithm proposed in this paper for solving the energy consumption optimization scheduling problem of the assembly line.

[0217] The calculation formula of HV evaluation index is as follows:

[0218]

[0219] Where: p represents a set of non-inferior solutions found by the algorithm, U represents a reference point. After normalizing the non-inferior solution set p, the coordinates of U are set to (1.01, 1.01, 1.01) to assist in the calculation; HV represents the hypervolume union formed by all points surrounded by the non-dominated solution set p and the reference point U in the target space; Volume represents the hypervolume, which refers to an indicator that measures the volume of the hypercube space formed from the non-inferior solution set p to the specified reference point U. This area is composed of the target vectors of all solutions and the boundary determined by the reference point U.

[0220] The calculation formula of IGD evaluation index is as follows:

[0221]

[0222] Where: IGD(s,S * ) represents the average distance between the ideal point and the actual solution set; s represents a specific solution; distance(s,S *) represents a specific solution s and the set S of all optimal non-dominated solutions found by a certain algorithm * The minimum gap between * It represents the final solution set formed by summarizing the non-dominated solution sets obtained by each algorithm and removing those solutions dominated by other solutions. Its number of members is |S * |.

[0223] This embodiment uses IMOGA+Random, a simple random initialization strategy, and IMOGA+HI, which combines two advanced initialization strategies, to verify the impact of the population initialization strategy on the present invention. The average and optimal values ​​of the HV index of these two algorithm versions running 20 times independently on multiple different test data sets are shown in Table 1:

[0224] Table 1 Comparison results of HV values ​​between IMOGA+Random and IMOGA+HI algorithms

[0225]

[0226] It can be seen from Table 1 that under specific test data, the HV average value obtained by IMOGA+Random is slightly higher than that of IMOGA+HI; however, in the remaining test instances, the HV average values ​​obtained by IMOGA+HI are significantly better than those of IMOGA+Random, and in the test instances of this embodiment, IMOGA+HI can achieve higher optimal HV values.

[0227] This embodiment uses IMOGA / DNTS without the taboo search strategy and the improved version IMOGA+TS with the taboo search strategy, and ensures that the other parameter configurations of the two are consistent. The HV indicators of the two algorithms run 20 times independently on a series of preset test data sets are shown in Table 2:

[0228] Table 2 Comparison of HV values ​​between IMOGA / DNTS and IMOGA+TS algorithms

[0229]

[0230] It can be seen from Table 2 that when solving the production scheduling problem of the flow shop, compared with IMOGA / DNTS that does not adopt this strategy, IMOGA+TS that integrates the taboo search strategy shows more significant advantages in overall performance, which is manifested in higher HV average values ​​and optimal values, thus confirming the effectiveness of the taboo search strategy in improving the solution effect.

[0231] In order to further verify the effectiveness of the IMOGA algorithm proposed in this invention in solving the workshop production scheduling problem, IMOGA is compared with a variety of optimization algorithms, including MOGWO, INSGA-II and PSO-GA. In order to ensure the fairness of the evaluation of various algorithms, Tables 3 and 4 show the average and optimal values ​​of HV and IGD when all algorithms are independently run 20 times using the same test data:

[0232] Table 3 Comparison results of HV values ​​of various algorithms

[0233]

[0234]

[0235] Table 4 Comparison results of IGD values ​​of various algorithms

[0236]

[0237] It can be seen from the results in Table 3 and Table 4 that the HV value and IGD value of the present invention show obvious advantages in the test data of most months, indicating that the improved initialization strategy, taboo search strategy, and population collaboration strategy proposed in the algorithm improve the quality and diversity of the solution, enhance the global search and local development capabilities, and make the solution set obtained by the algorithm more diverse and more convergent.

[0238] Example 4

[0239] like Figure 7 As shown, the flow shop energy consumption optimization scheduling system based on the improved genetic algorithm includes: a scheme generation module, a population generation module, a population update module, a best solution output module, a final population generation module and a result output module;

[0240] The scheme generation module is used to obtain multiple orders of the target flow shop and obtain the corresponding first scheduling scheme, second scheduling scheme and third scheduling scheme based on the improved LR rule, the total power consumption minimum principle and the random generation principle;

[0241] The population generation module is used to obtain the corresponding chromosome based on the corresponding scheduling scheme and generate the corresponding first population, second population and third population;

[0242] The population updating module is used to perform genetic operations on each population to generate corresponding new chromosomes and obtain corresponding updated populations;

[0243] The best solution output module is used to select two of the new chromosomes as the best solutions according to the rules based on the new chromosomes generated by all populations;

[0244] The final population generation module is used to obtain the optimal solution based on the best solution search and add it to all the updated populations respectively to obtain the corresponding final population;

[0245] The result output module is used to perform chromosome exchange based on all the final populations according to the corresponding objective function to obtain corresponding non-dominated solutions; and obtain a final non-dominated solution set based on all the non-dominated solutions.

[0246] Preferably, the functional implementation of each functional module in this embodiment corresponds to the above method one by one, and will not be repeated here.

[0247] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a method and system for optimizing the energy consumption of a flow shop based on an improved genetic algorithm, which has the following beneficial effects:

[0248] 1. This method adopts a diversified initialization strategy, using the improved LR rule and the minimum total power consumption to generate two relatively good individuals and use the two individuals to make the population explore the direction of the optimal solution. The remaining individuals are all randomly generated. This helps to cover the solution space more comprehensively, reduce the risk of the algorithm falling into the local optimum, and ensure the diversity of the population, thereby improving the quality of the initial solution and the quality of the population initialization, making the final solution set more diverse and more convergent.

[0249] 2. This method adopts the taboo search strategy to jump out of the local optimum and explore more solution spaces, thereby improving the global optimization effect of the algorithm; in addition, the algorithm incorporates the taboo search strategy as a mechanism to strengthen local exploration, allowing in-depth exploration of each member of the group; this strategy relies on its unique taboo principle to screen and adopt new solutions, which helps to break through the limitations of local optimality and enhance the global exploration ability of the algorithm. It improves the global search ability of the algorithm in complex solution space, enhances the depth of local exploration, and at the same time enhances the quality and efficiency of the algorithm's solution.

[0250] 3. This method uses the population coordination mechanism to balance the three optimization objectives, and combines it with the improved elite strategy. The poor solutions discarded in the fused population are cross-mutated to obtain a new population for fast non-dominated sorting. The solutions in the upper layer of the new population are non-dominated with the selected elite solutions to replace the poor solutions and retain the better individuals. Each solution balances the other two objectives while optimizing the current goal, thereby improving the production efficiency and resource utilization of the entire assembly line, making the solved scheduling plan more in line with the actual situation of the enterprise.

[0251] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0252] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. The energy consumption optimization scheduling method of the assembly line based on the improved genetic algorithm is characterized by: include: S1 obtains multiple orders of the target flow shop and obtains the corresponding first scheduling plan, second scheduling plan and third scheduling plan based on the improved LR rule, the principle of minimum total power consumption and the principle of random generation; S2 obtains the corresponding chromosome based on the corresponding scheduling scheme and generates the corresponding first population, second population and third population; S3 performs genetic operations on each population to generate corresponding new chromosomes and obtain corresponding updated populations; S4 selects two of the new chromosomes as the best solutions according to the rules based on the new chromosomes generated by all populations; S5 obtains the optimal solution based on the best solution search and adds it to all the update populations respectively to obtain the corresponding final population; S6 performs chromosome exchange based on all the final populations according to the corresponding objective function to obtain corresponding non-dominated solutions; S7 obtains a final non-dominated solution set based on all the non-dominated solutions.

2. The method for optimizing energy consumption of a flow shop based on an improved genetic algorithm according to claim 1 is characterized in that: The first scheduling solution is obtained based on the improved LR rule, which specifically includes: Calculate the completion time of each order on each production line; Get a comprehensive score based on the stated completion time; Sort all the orders in descending order based on the comprehensive scores, select the first R orders and distribute them to the production lines of R different voltage partitions in sequence; The remaining orders calculate their index values ​​with the production line; Select the order with the smallest index value and assign it to the production line with the shortest average completion time. The above process is repeated until all remaining orders are assigned to corresponding production lines, thereby obtaining the first scheduling solution based on the improved LR rule.

3. The method for optimizing energy consumption of a flow shop based on an improved genetic algorithm according to claim 2 is characterized in that: The second scheduling scheme based on the principle of minimum total power consumption specifically includes: Calculate the corresponding production time based on all the orders and their corresponding maximum vehicle speeds; Sort all the orders in descending order based on the production time; Based on the sorting results, the first two orders are selected and preferentially assigned to the production line with the lowest power consumption; The remaining orders are allocated to the corresponding production lines so that the remaining orders are processed on the corresponding production lines with the minimum additional power consumption, thereby obtaining the second scheduling plan based on the principle of minimum total power consumption.

4. The method for optimizing energy consumption of a flow shop based on an improved genetic algorithm according to claim 3 is characterized in that: S2 specifically includes: Arrange the corresponding production line identifiers based on the order of the corresponding production lines in the first scheduling scheme, the second scheduling scheme, and the third scheduling scheme, and obtain a first arrangement result, a second arrangement result, and a third arrangement result accordingly; Based on the first arrangement result, the second arrangement result and the third arrangement result, the first chromosome, the second chromosome and the third chromosome are correspondingly used; The first population, the second population and the third population are correspondingly obtained based on the first chromosome, the second chromosome and the third chromosome.

5. The method for optimizing energy consumption of a flow shop based on an improved genetic algorithm according to claim 4 is characterized in that: Genetic operations are performed on each population to generate corresponding new chromosomes, including: performing the genetic operation simultaneously based on the first population, the second population and the third population respectively; The genetic operation includes: selection operation, crossover operation and mutation operation; The selection operation is: randomly selecting two chromosomes from the corresponding population as selected chromosomes; selecting a selected chromosome with a low non-dominated level and a large crowding distance as a selected chromosome based on the selected chromosomes; The crossover operation is as follows: based on the selected chromosome, a plurality of orders for production line crossover are randomly selected as crossover orders; based on the first position of each crossover order on the parent generation P1, the second position of the production line on the selected chromosome is found on the parent generation P2, and the first position and the second position are exchanged until all the crossover orders are crossed to generate corresponding crossover chromosomes; The mutation operation is: randomly selecting two orders from the orders with position exchange in the cross chromosome as mutation orders; randomly selecting a production line other than the original processing line from the optional production lines of the mutation order to replace it, and another production line with electricity consumption less than the set value selected from the other mutation order to replace the original production line, to obtain the corresponding mutation chromosome; The selected chromosome and the variant chromosome are used together as the new chromosome.

6. The method for optimizing energy consumption of a flow shop based on an improved genetic algorithm according to claim 5 is characterized in that: The best solution acquisition method is: Based on the new chromosomes in all updated populations, two new chromosomes are selected as the best solutions based on a comprehensive objective function; The comprehensive objective function includes: minimizing the maximum completion time, minimizing the total power consumption and minimizing the shutdown switching loss.

7. The method for optimizing energy consumption of a flow shop based on an improved genetic algorithm according to claim 5 is characterized in that: The optimal solution acquisition method is: S41 initializes the taboo table and sets the best solution as the current solution and the initial global optimal solution; S42 selects a plurality of candidate solutions based on all domain solutions of the current solution; S43 determines whether the candidate solution is better than the initial global optimal solution, and if so, takes the candidate solution as a selected candidate solution and adds it to the taboo table; Otherwise, selecting the best solution from all the candidate solutions that are not tabooed as the selected candidate solution and adding it to the taboo table; S44 determines whether all the candidate solutions are in the taboo table. If so, the best selected candidate solution is selected in the taboo table as the optimal solution; otherwise, returns to S42.

8. The method for optimizing energy consumption of a flow shop based on an improved genetic algorithm according to claim 6 is characterized in that: S6 specifically includes: Based on the chromosomes in the first final population, the second final population, and the third final population, calculations are performed based on minimizing completion time, minimizing total power consumption, and minimizing shutdown switching losses, respectively, to obtain corresponding first optimization target values, second optimization target values, and third optimization target values; Sort the chromosomes in the corresponding final population in ascending order based on the corresponding optimization target value; Based on the sorting results, the first quarter of chromosomes in each final population are transferred as transfer individuals to the other two final populations, and the corresponding fusion populations are obtained; Performing a fast non-dominated sorting based on the fused population to obtain a first hierarchical result; The top solution based on the first stratification result is defined as the upper elite solution; Based on the chromosomes in the fusion population that have not entered the upper elite solution as discarded individuals, a corresponding discarded population is generated; Perform non-dominated sorting based on the discarded population to obtain a second stratification result; Based on the second stratification result, the top chromosome is selected as the priority solution; Performing non-dominated operations based on the priority solution and the upper elite solution and replacing inferior solutions, obtaining the corresponding non-dominated solution; The above process is repeated to obtain a first non-dominated solution, a second non-dominated solution and a third non-dominated solution based on the first final population, the second final population and the third final population respectively.

9. The method for optimizing energy consumption of a flow shop based on an improved genetic algorithm according to claim 8, characterized in that: S7 specifically includes: Perform search and replacement based on the first non-dominated solution, the second non-dominated solution, and the third non-dominated solution to update a non-dominated solution set; Calculating a fitness value based on the non-dominated solution set and the comprehensive objective function; Determining whether the fitness value is greater than the population fitness value; If yes, output the non-dominated solution set as the final non-dominated solution set; Otherwise return to S3.

10. The flow shop energy consumption optimization scheduling system based on improved genetic algorithm is applied to the flow shop energy consumption optimization scheduling method based on improved genetic algorithm as claimed in any one of claims 1 to 9, characterized in that: include: Solution generation module, population generation module, population update module, best solution output module, final population generation module and result output module; The scheme generation module is used to obtain multiple orders of the target flow shop and obtain the corresponding first scheduling scheme, second scheduling scheme and third scheduling scheme based on the improved LR rule, the total power consumption minimum principle and the random generation principle; The population generation module is used to obtain the corresponding chromosome based on the corresponding scheduling scheme and generate the corresponding first population, second population and third population; The population updating module is used to perform genetic operations on each population to generate corresponding new chromosomes and obtain corresponding updated populations; The best solution output module is used to select two of the new chromosomes as the best solutions according to the rules based on the new chromosomes generated by all populations; The final population generation module is used to obtain the optimal solution based on the best solution search and add it to all the updated populations respectively to obtain the corresponding final population; A result output module is used to perform chromosome exchange based on all the final populations according to the corresponding objective function to obtain the corresponding non-dominated solution; A final non-dominated solution set is obtained based on all the non-dominated solutions.

Images