A production scheduling method for freeze-dried fruits based on genetic algorithm

Abstract

The present invention discloses a fruit freeze-drying production scheduling method based on a genetic algorithm, comprising the following steps: step S1: obtaining the fruit type, slice weight, slice thickness, arrival time and equipment information of n fruit slices to be freeze-dried; step S2: constructing a model using a genetic algorithm and determining relevant constraints; step S3: processing the above model using a genetic algorithm, through initial population and encoding, dynamic programming function, selection operation and fitness function, crossover and mutation operations, until an optimal solution, i.e., an optimal production scheduling method, is generated; step S4: displaying the results of the optimal production scheduling method generated by the algorithm. Compared with the traditional genetic algorithm using a greedy strategy, the scheduling scheme obtained by the genetic algorithm operation proposed by the present invention can obviously shorten the maximum completion time of fruit freeze-drying and improve processing efficiency. The more types of fruits there are or the more vacuum freeze dryers there are, the more obvious the advantages are.

Description

Technical Field

[0001] The invention belongs to the technical field of production scheduling methods, and in particular relates to a fruit freeze-dried production scheduling method based on a genetic algorithm. Background Art

[0002] Fruit freeze-drying refers to the process of quickly freezing fruit slices and then dehydrating them in a vacuum ice state. Fruits with different sugar contents have different freezing temperatures. Generally, fruit slices must be quickly frozen to minus 50°C to minus 70°C, and then kept warm and dehydrated in a vacuum for dozens of hours. Due to the long processing time and the high price of freeze dryers, a reasonable production scheduling method is the key to improving production efficiency. The production process of fruit freeze-drying has typical batch scheduling characteristics. A vacuum freeze dryer can process multiple fruit slices at the same time, and a fruit slice can be processed on at most one vacuum freeze dryer at any time. Batch scheduling problems are widely present in various industrial environments. For example, semiconductor manufacturing aging tests, steel continuous casting, and mold heat treatment all have batch scheduling characteristics.

[0003] Genetic algorithms have been widely used in scheduling problems. Although they have achieved good results in the classic single-piece scheduling problem, their application in batch scheduling is not ideal. For the classic single-piece scheduling problem, in the evolutionary iteration, as long as the order of the workpieces is optimized and the workpieces are arranged on the earliest idle machine, the ideal scheduling solution can be obtained. However, for the batch scheduling problem, after the evolutionary process obtains the workpiece order, it is also necessary to determine how to group the workpieces into batches, and then arrange each batch on the machine. The existing batch scheduling genetic algorithm is based on a greedy strategy, that is, each batch is filled as much as possible, and then the earliest start time of each batch is determined and arranged on the earliest idle machine. Designing a more advanced batch grouping method and embedding it into the genetic algorithm is the key to improving the efficiency of batch scheduling by the genetic algorithm. Summary of the invention

[0004] The purpose of the present invention is to minimize the maximum completion time as the optimization goal, provide a fruit freeze-dried production scheduling method based on a genetic algorithm, which is applied to the scheduling of the fruit freeze-dried production process, improves production efficiency, shortens the processing cycle, and solves the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a fruit freeze-dried production scheduling method based on genetic algorithm, the main structure of which is: comprising the following steps:

[0006] Step S1: obtaining the fruit type, slice weight, slice thickness, arrival time and equipment information of n fruit slices to be freeze-dried;

[0007] Step S2: construct a model using a genetic algorithm and determine relevant constraints;

[0008] Step S3: The above model is processed by genetic algorithm, through initial population and coding, dynamic programming function, selection operation and fitness function, crossover and mutation operation, until the optimal solution is generated, that is, the optimal production scheduling method;

[0009] Step S4: Display the optimal production scheduling method results generated by the algorithm.

[0010] Preferably, in step S1, multiple vacuum freeze dryers of the same specifications are used, and each vacuum freeze dryer can accommodate multiple fruit slices for simultaneous processing; the total weight of each batch of fruit slices loaded cannot exceed the machine capacity c; the processing time of each batch consists of quick freezing time and heat preservation and dehydration time. Different types of fruits need to be quick frozen to different temperatures. Only fruits of the same type can be loaded into the same batch, that is, there are incompatible workpiece families. The quick freezing time is inversely proportional to the total weight of the batch, and the heat preservation time mainly depends on the maximum fruit slice thickness in the batch. Batch B b The processing time can be expressed as:

[0011]

[0012] Among them, α i represents the quick freezing time coefficient of type i, while all varieties have the same insulation and dehydration time coefficient β, w j Indicates the slice weight, s j represents the slice thickness. The vacuum freeze dryer process is not allowed to pause and insert other batches. The scheduling goal is to minimize the maximum completion time C max .

[0013] Preferably, the genetic algorithm processing steps in step S3 are as follows:

[0014] Step SA: read in processing information, set population size, number of iterations, crossover and mutation probabilities, and start genetic algorithm;

[0015] Step SB: Initialize the population, generate chromosomes based on the artifact sequence encoding, and the number of chromosomes is the preset population size;

[0016] Step SC: Call the dynamic programming function, calculate the maximum completion time objective function value corresponding to each chromosome, convert it into a fitness value, and save the historical optimal chromosome and its objective function value;

[0017] Step SD: Select the remaining chromosomes by roulette wheel;

[0018] Step SE: Based on the preset crossover probability, the parent chromosomes undergo multi-point crossover to generate new individuals;

[0019] Step SF: Generate new individuals based on the preset mutation probability;

[0020] Step SG: whether the number of iterations has been reached;

[0021] Step SH: When the number of iterations has not been reached, return to step SC and recalculate; when the number of iterations has been reached, output the historical optimal chromosome and its objective function value.

[0022] Preferably, the quality of the initial population in step SB has a great influence on the calculation efficiency of the algorithm. In order to ensure that there are better chromosomes in the initial population, the first chromosome is generated in the following way: sorted by arrival time from small to large, and if the arrival times are the same, sorted by weight from large to small; the last chromosome is generated in the following way: sorted by weight from large to small, and if the weights are the same, sorted by arrival time from small to large.

[0023] Preferably, the dynamic programming function in step SC uses the chromosome encoding as the processing order of the fruit slices, and calculates the scheduling plan and the corresponding maximum completion time through the dynamic programming function.

[0024] Preferably, the process of the dynamic programming function in step SC is as follows:

[0025] Step SC1: read the first idle vacuum freeze dryer k, set the idle time as t, and set the latest start processing time of the vacuum freeze dryer k as t+T, where T is 80% of the longest processing time of a single batch, and then go to step SC2;

[0026] Step SC2: Group the fruit slices of each variety that arrive before time t according to sequence L. If the weight of a certain fruit to be processed reaches the minimum processing weight, the variety with the largest weight is processed first, and the next idle time of the vacuum freeze dryer is updated as the completion time of the batch, and then go to step SC4; otherwise, if the weight of each fruit to be processed does not reach the minimum processing weight, then go to step SC3;

[0027] Step SC3: If the time t is not less than the latest start time of the vacuum freeze dryer, and the fruit slices to be scheduled that arrive before the time t are not empty, arrange the processing of the variety with the largest total weight to be scheduled, and update the next idle time of the vacuum freeze dryer as the completion time of the batch, and go to step SC4; otherwise, take the workpiece arriving in the next time window into consideration, that is, update the time t, set t=t+T / 3, and go to step SC2;

[0028] Step SC4: If all workpieces have been scheduled, output the scheduling results; otherwise, go to step SC1.

[0029] Preferably, the formula for converting the objective function value in step SC into the fitness value is:

[0030]

[0031] where r j represents the arrival time, max(r j ) represents the maximum arrival time of the workpiece, C max (i) represents the maximum completion time calculated by decoding the i-th chromosome in the population, and 3600 represents the conversion of the time unit from seconds to hours;

[0032] The formula for converting the chromosome fitness value into a probability value is:

[0033] p(i)=fitness(i) / ∑ k∈pop fitness(k) (2)

[0034] Where pop represents the contemporary population.

[0035] Preferably, in step SD, a roulette wheel is used to select the dominant chromosome remaining after crossover and mutation. The probability range of each chromosome in the roulette wheel is obtained by the following method:

[0036] The probability interval of the first chromosome in the roulette wheel: [0, p(1)];

[0037] The probability interval of the nth chromosome in the roulette wheel:

[0038] Preferably, in step SE, the parent chromosome to be crossovered is first copied and recorded as parent chromosome 1 and parent chromosome 2, the crossover probability is pre-set, a floating point number sequence is randomly generated and compared with the crossover probability one by one; if the random floating point number is less than the crossover probability, the gene at the corresponding position of parent chromosome 1 is retained, and the corresponding gene code in parent chromosome 2 is deleted; conversely, if the random floating point number is not less than the crossover probability, the gene at the corresponding position of parent chromosome 1 is changed to -1, and then the remaining code of parent chromosome 2 after the change is filled in the position coded as -1 in parent chromosome 1 in sequence, so as to obtain a new daughter chromosome.

[0039] Preferably, a mutation probability is preset in step SF. If the random number is less than the preset mutation probability, the current chromosome undergoes a mutation operation. When a chromosome with n genes is mutated, two integers are first generated using the random function randint(0,n), and then the genes at the corresponding positions of the two integers are exchanged to obtain a new chromosome through mutation.

[0040] Compared with the traditional genetic algorithm using greedy strategy, the scheduling scheme calculated by the genetic algorithm proposed in the present invention can obviously shorten the maximum completion time of fruit freeze-drying and improve processing efficiency. The more types of fruits or the more vacuum freeze dryers there are, the more obvious the advantage. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is the flow chart of the genetic algorithm of the present invention. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the invention claimed for protection, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0043] See also Figure 1 The present invention provides a technical solution, a fruit freeze-dried production scheduling method based on genetic algorithm, comprising the following steps:

[0044] Step S1: obtaining the fruit type, slice weight, slice thickness, arrival time and equipment information of n fruit slices to be freeze-dried;

[0045] Step S2: construct a model using a genetic algorithm and determine relevant constraints;

[0046] Step S3: The above model is processed by genetic algorithm, through initial population and coding, dynamic programming function, selection operation and fitness function, crossover and mutation operation, until the optimal solution is generated, that is, the optimal production scheduling method;

[0047] Step S4: Display the optimal production scheduling method results generated by the algorithm.

[0048] In this embodiment, multiple vacuum freeze dryers of the same specifications are used in step S1, and each vacuum freeze dryer can accommodate multiple fruit slices for simultaneous processing; the total weight of each batch of fruit slices loaded cannot exceed the machine capacity c; the processing time of each batch consists of quick freezing time and heat preservation and dehydration time. Different types of fruits need to be quick frozen to different temperatures. Only fruits of the same type can be loaded into the same batch, that is, there are incompatible workpiece families. The quick freezing time is inversely proportional to the total weight of the batch, and the heat preservation time mainly depends on the maximum fruit slice thickness in the batch. Batch B b The processing time can be expressed as:

[0049]

[0050] Among them, α i represents the quick freezing time coefficient of type i, while all varieties have the same insulation and dehydration time coefficient β, w j Indicates the slice weight, s j represents the slice thickness. The vacuum freeze dryer process is not allowed to pause and insert other batches. The scheduling goal is to minimize the maximum completion time C max .

[0051] In this embodiment, the genetic algorithm processing steps in step S3 are as follows:

[0052] Step SA: read in processing information, set population size, number of iterations, crossover and mutation probabilities, and start genetic algorithm;

[0053] Step SB: Initialize the population, generate chromosomes based on the artifact sequence encoding, and the number of chromosomes is the preset population size;

[0054] Step SC: Call the dynamic programming function, calculate the maximum completion time objective function value corresponding to each chromosome, convert it into a fitness value, and save the historical optimal chromosome and its objective function value;

[0055] Step SD: Select the remaining chromosomes by roulette wheel;

[0056] Step SE: Based on the preset crossover probability, the parent chromosomes undergo multi-point crossover to generate new individuals;

[0057] Step SF: Generate new individuals based on the preset mutation probability;

[0058] Step SG: whether the number of iterations has been reached;

[0059] Step SH: When the number of iterations has not been reached, return to step SC and recalculate; when the number of iterations has been reached, output the historical optimal chromosome and its objective function value.

[0060] Among them, in this embodiment, there are two encoding methods for batch scheduling problems: batch-based and workpiece sequence-based. The batch-based encoding method is only applicable to special batch scheduling scenarios, and the workpiece-based encoding method has strong versatility. Therefore, the present invention adopts a workpiece-based encoding method, each workpiece is represented by a unique natural number, and the encoding sequence represents the workpiece sorting sequence. The quality of the initial population in the step SB has a great influence on the calculation efficiency of the algorithm. In order to make the initial population have better chromosomes, the first chromosome is generated in the following way: sort by arrival time from small to large, and if the arrival time is the same, sort by weight from large to small; the last chromosome is generated in the following way: sort by weight from large to small, and if the weight is the same, sort by arrival time from small to large.

[0061] For example, for the five fruit slices in Table 1, according to the above encoding rules, the encoding of the first chromosome is (3, 4, 2, 5, 1), and the encoding of the last chromosome is (3, 4, 5, 2, 1).

[0062] Table 1 Fruit slice information table

[0063]

[0064] The remaining chromosome genes of the initial population are generated by random natural numbers. For example, for a scheduling problem with n fruit slices, a random natural number sequence from 1 to n is generated as the genes of one chromosome.

[0065] Among them, in this embodiment, the code of one chromosome is passed as a parameter to the dynamic programming function. The dynamic programming function in step SC uses the chromosome code as the processing order of the fruit slices, and calculates the scheduling plan and the corresponding maximum completion time through the dynamic programming function.

[0066] The basic idea of ​​the dynamic programming strategy is: based on the idle time of the vacuum freeze dryer, combined with a variety of batch scheduling rules, determine the fruit slices included in the current batch and the start processing time, and dynamically update the next idle time of the vacuum freeze dryer. Comparing the scheduling performance of various batch scheduling rules, the present invention adopts a batch grouping method that combines the minimum processing weight, the longest idle waiting time of the equipment and the maximum weight priority. The minimum processing weight refers to the lower weight limit that the total weight of the fruit slices in each batch must reach, which is set to 60% of the capacity of the vacuum freeze dryer in this patent. The longest idle waiting time of the vacuum freeze dryer is set to 80% of the longest processing time of a single batch, and the longest idle waiting time of the vacuum freeze dryer is divided into 3 time windows to dynamically evaluate the weight of the fruit slices that have been reached. Maximum weight priority means that the fruit slices are sorted by weight, and fruit slices with large weight are prioritized for batching.

[0067] In this embodiment, the process of the dynamic programming function in step SC is as follows:

[0068] Step SC1: read the first idle vacuum freeze dryer k, set the idle time as t, and set the latest start processing time of the vacuum freeze dryer k as t+T, where T is 80% of the longest processing time of a single batch, and then go to step SC2;

[0069] Step SC2: Group the fruit slices of each variety that arrive before time t according to sequence L. If the weight of a certain fruit to be processed reaches the minimum processing weight, the variety with the largest weight is processed first, and the next idle time of the vacuum freeze dryer is updated as the completion time of the batch, and then go to step SC4; otherwise, if the weight of each fruit to be processed does not reach the minimum processing weight, then go to step SC3;

[0070] Step SC3: If the time t is not less than the latest start time of the vacuum freeze dryer, and the fruit slices to be scheduled that arrive before the time t are not empty, arrange the processing of the variety with the largest total weight to be scheduled, and update the next idle time of the vacuum freeze dryer as the completion time of the batch, and go to step SC4; otherwise, take the workpiece arriving in the next time window into consideration, that is, update the time t, set t=t+T / 3, and go to step SC2;

[0071] Step SC4: If all workpieces have been scheduled, output the scheduling results; otherwise, go to step SC1.

[0072] In this embodiment, the formula for converting the objective function value in step SC into the fitness value is:

[0073]

[0074] where r j represents the arrival time, max(r j ) represents the maximum arrival time of the workpiece, C max (i) represents the maximum completion time calculated by decoding the i-th chromosome in the population, and 3600 represents the conversion of the time unit from seconds to hours;

[0075] The formula for converting the chromosome fitness value into a probability value is:

[0076] p(i)=fitness(i) / ∑ k∈pop fitness(k) (2)

[0077] Where pop represents the contemporary population. From formula (1) and formula (2), we can see that the larger the completion time corresponding to a chromosome, the smaller its fitness value, and the smaller the probability of being selected.

[0078] In this embodiment, the roulette wheel method is used in step SD to select the dominant chromosomes remaining after crossover and mutation. The probability range of each chromosome in the roulette wheel is obtained by the following method:

[0079] The probability interval of the first chromosome in the roulette wheel: [0, p(1)];

[0080] The probability interval of the nth chromosome in the roulette wheel:

[0081] Generate a random [0,1] floating-point array according to the preset population size of the genetic algorithm, and select the corresponding chromosome according to the position of the random number in the roulette wheel. After the selection operation is completed, the number of chromosomes retained is the initially set population size.

[0082] Among them, in this embodiment, crossover is the most direct method for genetic algorithm to produce offspring, and commonly used crossover operations include single-point crossover and double-point crossover. According to the coding characteristics of batch scheduling problems, the present invention adopts a multi-point crossover method to generate a random [0,1] floating point number sequence.

[0083] The following is an analysis of the crossover operation process of parent chromosome 1 and chromosome 2 through a simple example.

[0084] Parental chromosome 1: (1, 3, 5, 2, 4, 6)

[0085] Parental chromosome 2: (2, 4, 6, 1, 3, 5)

[0086] Random floating point number sequence: (0.9, 0.7, 0.5, 0.3, 0.5, 0.7)

[0087] In the step SE, the parent chromosome to be crossovered is first copied, recorded as parent chromosome 1 and parent chromosome 2, the crossover probability is pre-set to 0.6, a floating point sequence is randomly generated and compared with the crossover probability one by one; if the random floating point number is less than the crossover probability, the gene at the corresponding position of parent chromosome 1 is retained, and the corresponding gene code in parent chromosome 2 is deleted; on the contrary, if the random floating point number is not less than the crossover probability, the gene at the corresponding position of parent chromosome 1 is changed to -1. After the cyclic comparison operation, the coding of the parent chromosome changes as follows:

[0088] Parental chromosome 1: (-1, -1, 5, 2, 4, -1)

[0089] Parental chromosome 2: (6, 1, 3)

[0090] Then, the remaining codes of the parent chromosome 2 after the change are filled into the positions coded as -1 in the parent chromosome 1 in sequence, so that a new daughter chromosome is obtained, and the code of the daughter chromosome is: (6, 1, 5, 2, 4, 3).

[0091] Among them, in this embodiment, the preset mutation probability in the step SF is 0.1, and a [0,1] floating point number is randomly generated. If the random number is less than the preset mutation probability, the current chromosome undergoes a mutation operation; when a chromosome with n genes is mutated, two integers are first generated using the random function randint(0,n), and then the genes at the corresponding positions of the two integers are exchanged to mutate a new chromosome.

[0092] Compared with the traditional genetic algorithm using greedy strategy, the scheduling scheme calculated by the genetic algorithm proposed in the present invention can obviously shorten the maximum completion time of fruit freeze-drying and improve processing efficiency. The more types of fruits or the more vacuum freeze dryers there are, the more obvious the advantage.

[0093] It is obvious to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential features of the present invention. Therefore, the embodiments should be regarded as exemplary and non-limiting from any point of view, and the scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention, and any reference numerals in the claims should not be regarded as limiting the claims involved.

[0094] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A fruit freeze-dried production scheduling method based on genetic algorithm, characterized in that: The following steps are involved: Step S1: obtaining the fruit type, slice weight, slice thickness, arrival time and equipment information of n fruit slices to be freeze-dried; Step S2: construct a model using a genetic algorithm and determine relevant constraints; Step S3: The above model is processed by genetic algorithm, through initial population and coding, dynamic programming function, selection operation and fitness function, crossover and mutation operation, until the optimal solution, i.e. the optimal production scheduling method, is generated; the genetic algorithm processing steps are as follows: Step SA: read in processing information, set population size, number of iterations, crossover and mutation probabilities, and start genetic algorithm; Step SB: Initialize the population, generate chromosomes based on the artifact sequence encoding, and the number of chromosomes is the preset population size; Step SC: Call the dynamic programming function, calculate the maximum completion time objective function value corresponding to each chromosome, convert it into a fitness value, and save the historical optimal chromosome and its objective function value; the formula for converting the objective function value into the fitness value is: where r j represents the arrival time, max(r j ) represents the maximum arrival time of the workpiece, C max (i) represents the maximum completion time calculated by decoding the i-th chromosome in the population, and 3600 represents the conversion of the time unit from seconds to hours; The formula for converting the chromosome fitness value into a probability value is: p(i)=fitness(i) / ∑ k∈pop fitness(k) Where pop represents the contemporary population; Step SD: Select the remaining chromosomes by roulette wheel; Step SE: Based on the preset crossover probability, the parent chromosomes undergo multi-point crossover to generate new individuals; Step SF: Generate new individuals based on the preset mutation probability; Step SG: whether the number of iterations has been reached; Step SH: When the number of iterations is not reached, return to step SC to recalculate; when the number of iterations is reached, output the historical optimal chromosome and its objective function value; Step S4: Display the optimal production scheduling method results generated by the algorithm.

2. A method for scheduling freeze-dried fruit production based on genetic algorithm according to claim 1, characterized in that: In the step S1, multiple vacuum freeze dryers of the same specifications are used, and each vacuum freeze dryer can accommodate multiple fruit slices for simultaneous processing; the total weight of each batch of fruit slices loaded cannot exceed the machine capacity c; the processing time of each batch consists of quick freezing time and heat preservation and dehydration time. Different types of fruits need to be quick frozen to different temperatures. Only fruits of the same type can be loaded into the same batch, that is, there are incompatible workpiece families. The quick freezing time is inversely proportional to the total weight of the batch, and the heat preservation time mainly depends on the maximum fruit slice thickness in the batch. Batch B b The processing time is expressed as: Among them, α i represents the quick freezing time coefficient of type i, while all varieties have the same insulation and dehydration time coefficient β, w j Indicates the slice weight, s j represents the slice thickness. The vacuum freeze dryer process is not allowed to pause and insert other batches. The scheduling goal is to minimize the maximum completion time C max .

3. A fruit freeze-dried production scheduling method based on genetic algorithm according to claim 1, characterized in that: The first chromosome in step SB is generated by: sorting by arrival time from small to large, and if the arrival time is the same, sorting by weight from large to small; the last chromosome is generated by: sorting by weight from large to small, and if the weight is the same, sorting by arrival time from small to large.

4. A fruit freeze-dried production scheduling method based on genetic algorithm according to claim 1, characterized in that: The dynamic programming function in step SC uses the chromosome encoding as the processing order of the fruit slices, and calculates the scheduling plan and the corresponding maximum completion time through the dynamic programming function.

5. A fruit freeze-dried production scheduling method based on genetic algorithm according to claim 4, characterized in that: The process of the dynamic programming function in step SC is as follows: Step SC1: read the first idle vacuum freeze dryer k, set the idle time as t, and set the latest start processing time of the vacuum freeze dryer k as t+T, where T is 80% of the longest processing time of a single batch, and then go to step SC2; Step SC2: group the fruit slices of each variety arriving before time t according to sequence L. If the weight of a certain fruit to be processed reaches the minimum processing weight, the variety with the largest weight is processed first, and the next idle time of the vacuum freeze dryer is updated as the completion time of the batch, and then go to step SC4; otherwise, if the weight of the fruits to be processed does not reach the minimum processing weight, then go to step SC3; Step SC3: If the time t is not less than the latest start time of the vacuum freeze dryer, and the fruit slices to be scheduled that arrive before the time t are not empty, arrange the processing of the variety with the largest total weight to be scheduled, and update the next idle time of the vacuum freeze dryer as the completion time of the batch, and go to step SC4; Otherwise, the workpiece arriving in the next time window is taken into account, that is, the time t is updated, and t=t+T / 3 is set, and then the process goes to step SC2; Step SC4: If all workpieces are scheduled, output the scheduling results; Otherwise, go to step SC1.

6. A fruit freeze-dried production scheduling method based on genetic algorithm according to claim 1, characterized in that: In the step SD, a roulette wheel method is used to select the dominant chromosomes remaining after crossover and mutation, and the probability range of each chromosome in the roulette wheel is obtained by the following method: The probability interval of the first chromosome in the roulette wheel: [0, p(1)]; The probability interval of the nth chromosome in the roulette wheel:

7. A fruit freeze-dried production scheduling method based on genetic algorithm according to claim 1, characterized in that: In the step SE, the parent chromosome to be crossovered is first copied, recorded as parent chromosome 1 and parent chromosome 2, the crossover probability is pre-set, a floating point number sequence is randomly generated and compared with the crossover probability one by one; if the random floating point number is less than the crossover probability, the gene at the corresponding position of parent chromosome 1 is retained, and the corresponding gene code in parent chromosome 2 is deleted; on the contrary, if the random floating point number is not less than the crossover probability, the gene at the corresponding position of parent chromosome 1 is changed to -1, and then the remaining code of parent chromosome 2 after the change is sequentially filled into the position coded as -1 in parent chromosome 1, so as to obtain a new daughter chromosome.

8. A fruit freeze-dried production scheduling method based on genetic algorithm according to claim 1, characterized in that: In the step SF, a mutation probability is preset. If the random number is less than the preset mutation probability, the current chromosome is mutated. When a chromosome with n genes is mutated, two integers are first generated using the random function randint(0,n), and then the genes at the corresponding positions of the two integers are exchanged to obtain a new chromosome.

Images