Special-shaped multi-machine cooperative scheduling method and system based on genetic programming

Abstract

The invention provides a special-shaped multi-machine cooperative scheduling method and system based on genetic programming, and belongs to the field of agricultural machinery scheduling. Comprising the following steps: S1, initializing parameters of a special-shaped multi-machine system; s2, generating an initial population according to the farmland operation sequence and the agricultural machinery set; s3, calculating agricultural machinery scheduling efficiency as a first fitness value; s4, judging whether a population migration condition is met or not, if so, migrating the population according to the first fitness value and recalculating the fitness value, otherwise, continuing to execute the subsequent steps; s5, performing crossover variation and duplicate removal operation on the population according to the current fitness value to form a new population, and repeatedly executing the steps S3 to S5; s6, the maximum number of iterations is reached, and an optimal agricultural machinery scheduling rule is obtained; and S7, operating the optimal rule in the test data set to obtain a final scheduling scheme. According to the method, excellent adaptability and robustness can be shown in the face of problems of different types and scales, and more intelligent and efficient decision support is provided for the agricultural machinery scheduling problem.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural machinery scheduling, and in particular to a method and system for cooperative scheduling of heterogeneous multiple machines based on genetic programming. Background Art

[0002] The operation mode of cross-regional agricultural machinery dispatch is a modern agricultural management method, which aims to improve agricultural production efficiency and economic benefits through scientific and reasonable resource allocation. This mode usually relies on Internet of Things technology, big data analysis and advanced dispatching algorithms to achieve remote monitoring and efficient deployment of agricultural machinery. Specifically, during the busy farming season, the flow path of agricultural machinery is planned according to the crop planting area, farming progress and mechanical resource distribution in various places, the idle rate is reduced, and the machinery can arrive at the required location in time for operation. For the situation where the farmland of individual households in rural and urban areas that do not have cooperative conditions has continuous operation needs, scattered distribution and large number during the crop season change, different types of agricultural machinery are usually required for collaborative operation. The service mode of heterogeneous multi-machine collaborative dispatch can meet this demand. Cross-regional agricultural machinery dispatch not only helps to balance the contradiction between supply and demand of agricultural machinery between regions, but also promotes the overall improvement of the level of agricultural mechanization. It is an indispensable part of the development of modern agriculture.

[0003] In the prior art, there are still the following problems:

[0004] At present, some cross-regional scheduling methods provide farmers and agricultural machinery operators with communication channels and interaction conditions through the information service management platform, and the two directly communicate about job requests or the scheduling system provides solutions. The existing agricultural machinery scheduling methods, regardless of the scheme, have the problem of low efficiency in agricultural machinery scheduling. If the scheme is a direct communication between farmers and agricultural machinery operators, it is impossible to arrange suitable agricultural machinery for various demand operations in the farmland, or resources will be occupied, resulting in low scheduling efficiency; if a specific arrangement scheme is given to the scheduling system, there is a low efficiency in scheduling schemes, because the scheduling schemes currently used are mostly traditional heuristic methods. When dealing with complex optimization problems, heuristic algorithms often rely on repeated iterations and local improvements of the initial solution. However, this strategy has a significant limitation - it is overly dependent on the characteristics of specific scenarios and the knowledge and experience of domain experts. In addition, the agricultural machinery scheduling operation process needs to take into account both the field operation process and the agricultural machinery transfer process. Therefore, a method that can take into account both operation processes and break away from the constraints of traditional heuristics is needed to optimize agricultural machinery scheduling. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for coordinated scheduling of heterogeneous multi-machines based on genetic programming, which can solve the problem that the current agricultural machinery scheduling model in agricultural operations is overly dependent on fixed operation scenes, resulting in weak generalization and low scheduling efficiency, especially in the face of continuous farmland operations, such as the scheduling challenges when a piece of farmland needs to perform multiple continuous operation processes such as harvesting, plowing, fertilizing, and sowing. It is necessary to simultaneously deploy multiple types of multiple agricultural machinery and take into account both field operations and agricultural machinery transfer processes. A heterogeneous multi-machine coordinated scheduling model is established by collecting the parameters of agricultural machinery and operating farmland, with the shortest sum of operation time and transfer time as the optimization goal, and an adaptive multi-population genetic programming algorithm and system are used to allocate cross-regional operation tasks, thereby obtaining the deployment route of each agricultural machinery.

[0006] To achieve the above object, the present invention provides a heterogeneous multi-machine collaborative scheduling method based on genetic programming, comprising the following steps:

[0007] Step S1, initializing the parameters of the heterogeneous multi-machine system scheduling problem;

[0008] Step S2, generating an initial population according to the operation sequence of the farmland and the set of agricultural machinery;

[0009] Step S3, calculating the agricultural machinery dispatching efficiency as the first fitness value according to the sum of the agricultural machinery operation time and the transfer time and the sum of the areas of all processes of the farmland;

[0010] Step S4, determining whether the population migration conditions are met;

[0011] When the population migration condition is met, a migration operation is performed according to the first fitness value to obtain a population after migration, and the fitness value of the population after migration is recalculated to obtain a second fitness value;

[0012] When the population migration condition is not met, jump to step S5;

[0013] Step S5, performing crossover mutation and deduplication operations on the initial population or the migrated population according to the first fitness value or the second fitness value to form a new population;

[0014] Step S6, repeating steps S3 to S5 until the maximum number of iterations is reached, the evolution process ends, and the optimal agricultural machinery scheduling rule is obtained;

[0015] Step S7: Run the optimal agricultural machinery scheduling rule obtained in step S6 on the test data set to obtain the final agricultural machinery scheduling plan.

[0016] Preferably, in step S1, initializing the heterogeneous multi-machine system scheduling problem parameters includes the following steps:

[0017] Step S11, setting basic information of farmland and agricultural machinery;

[0018] Define the farmland operation set F = {f1, f2, ..., f n}, f i represents the i-th job, i∈[1,n], n represents the number of farmlands; the farmland attribute is represented by f i ={sf, of, lfx, lfy}, where sf represents the area of ​​the current farmland; of represents the operation sequence required for the current farmland; lfx represents the longitude coordinate of the current farmland; lfy represents the latitude coordinate of the current farmland;

[0019] Define the operation sequence of the farmland i ={o1, o2, ..., o x}, ok represents the kth process, k∈[1,x], x represents the number of processes in any farmland; the attributes of each process in the operation sequence are expressed as ok={jo,co}, where jo represents the number of the farmland to which the current process belongs; co represents the type of the current process;

[0020] Define the agricultural machinery set M = {m1, m2, ..., mp}, mj represents the jth agricultural machinery, j∈[1, p], p represents the total number of agricultural machinery; the agricultural machinery attributes are represented by mj = {tm, em, ms, lmx, lmy}, where tm represents the type of agricultural machinery; em represents the efficiency of agricultural machinery; ms represents the transfer speed of agricultural machinery; lmx represents the current longitude coordinate of agricultural machinery; lmy represents the current latitude coordinate of agricultural machinery;

[0021] Step S12, setting population number, population size, evolutionary generation, initial exchange generation and reproduction parameters;

[0022] Step S13, dividing the set into training set and test set.

[0023] Preferably, in step S2, the size of the initial population is 1024, including 1 main population and 2 auxiliary populations.

[0024] Preferably, in step S3, the calculation formula of the first fitness value is as follows:

[0025]

[0026] Among them, maxEfficiency represents the efficiency of agricultural machinery scheduling, that is, the first fitness value; represents the total area of ​​n farmlands; represents the total scheduling time of n farmlands; PT i Represents the processing time of farmland operations; RM i Indicates the agricultural machinery transfer time; S k It represents the total effective dispatch area of ​​a certain farmland;

[0027] S k The calculation formula is as follows:

[0028]

[0029] Among them, SF q Indicates the area required for the corresponding operation in the farmland;

[0030] PT i The calculation formula is as follows:

[0031]

[0032] Among them, SF k represents the area of ​​the operation corresponding to the kth process; EM k represents the working efficiency of the agricultural machinery corresponding to the kth process;

[0033] RM i The calculation formula is as follows:

[0034]

[0035] Among them, D represents the distance required to transfer from the previous farmland to the current farmland; V represents the transportation speed of the agricultural machinery in the current process.

[0036] Preferably, in step S3, when calculating the first fitness value, the following constraints must be met:

[0037] StarTO k-1 <FinTO k-1 <StarTO k <FinTO k ,

[0038] Completion time (O ik ) = start time (O ik )+t ik ,

[0039] st k >max(ft μ μ∈P(k));

[0040]

[0041] Among them, StarTO k-1 Indicates the start time of the k-1th process; FinTO k-1 Indicates the end time of the k-1th process; StarTO k Indicates the start time of the kth process; FinTO k represents the end time of the kth process; Oik represents the kth process of the ith job; t ik represents the processing time of the kth process of the ith job; st k is the start time of process k, and process k has a set of processes that must be completed first, P(k); ft μ represents the end time of the process in P(k); x j represents the number of processes handled by the j-th agricultural machine at a certain moment; y i represents the completion status of the i-th farm operation; x ik Indicates the completion status of the kth process of the ith job.

[0042] Preferably, in step S4, a migration operation is performed according to the first fitness value to obtain a population after migration, and the specific operation is:

[0043] The first fitness value is ranked from small to large, and the first one-third of the individuals are selected as excellent individuals, and the remaining two-thirds of the individuals are ordinary individuals;

[0044] Given the initial migration rate mrate and migration interval interval and their respective changes g and h, migration is carried out according to the migration rate and migration interval, the common individuals of the main population are exchanged with the individuals of the auxiliary population, and the average fitness change is counted after the migration is completed to adaptively adjust the migration interval and migration rate;

[0045] When the migration operation is completed, if this migration evolves in a better direction than the previous migration, the migration interval will be increased and the migration rate will be reduced, that is:

[0046] mrate'=mrate-g;

[0047] interval'=interval+h;

[0048] Otherwise, reduce the migration interval and increase the migration rate, that is:

[0049] mrate'=mrate+g;

[0050] interval'=interval-h;

[0051] Wherein, mrate' is the new migration rate; interval' is the new migration interval.

[0052] Preferably, in step S4, the calculation formula of the second fitness value is as follows:

[0053] Fitness2=totalTime / totalProcessArea;

[0054] Among them, Fitness2 is the second fitness value; totalTime is the time consumed in the entire operation process; totalProcessArea is the sum of the operation areas of each farmland, and the calculation formula of totalProcessArea is as follows:

[0055]

[0056] Preferably, in step S5, the crossover mutation operation is as follows:

[0057] Reproduction selection is performed through pre-set crossover rate and mutation rate;

[0058] Eligible parents are selected for breeding operations using a tournament selection algorithm of size 6;

[0059] Call the corresponding reproduction operation through the ECJ framework to reproduce;

[0060] The calculation formula of the crossover rate is as follows:

[0061] crossRate=c0+((c1-c0) / (evolveSum×a))×nowGeneration;

[0062] Among them, c0 and c1 are the initial crossover rate and the final crossover rate respectively; a is the convergence point; evolveSum represents the total evolutionary generation; nowGeneration is the current generation;

[0063] The mutation rate is 1-crossRate.

[0064] Preferably, in step S5, the operation of forming a new population is as follows:

[0065] After each breeding operation, new individuals are placed into a new population;

[0066] When the individuals in the new population reach the predetermined size, the breeding operation is ended and the new population covers the original population.

[0067] The present invention also provides a system for implementing the above-mentioned heterogeneous multi-machine collaborative scheduling method based on genetic programming, comprising:

[0068] Initialization module, used to initialize the parameters of heterogeneous multi-machine system scheduling problem and generate the initial population;

[0069] The algorithm evolution module is used to evolve GP individuals according to the pre-set algorithm parameters to obtain the optimal agricultural machinery scheduling rules;

[0070] The result calculation module is used to calculate the final agricultural machinery scheduling plan based on the optimal agricultural machinery scheduling rules and test data sets.

[0071] Therefore, the present invention adopts the above-mentioned genetic programming-based heterogeneous multi-machine collaborative scheduling method and system, solves various types of cross-regional scheduling models of agricultural machinery by improving the genetic programming algorithm, solves the stereotyped problem of the current traditional heuristic method, and can obtain efficient and generalized agricultural machinery scheduling rules, avoiding the influence of human subjective factors. Significantly improve economic benefits and efficiently perform continuous operations on crops. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 This is a flow chart of a heterogeneous multi-machine collaborative scheduling method based on genetic programming of the present invention;

[0073] Figure 2 This is a schematic diagram of cross-regional heterogeneous multi-machine collaborative scheduling;

[0074] Figure 3 Diagram representing individuals for genetic planning. DETAILED DESCRIPTION

[0075] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0076] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field to which the present invention belongs.

[0077] Embodiment 1

[0078] like Figure 1 FIG. 1 is a flowchart of a method for coordinating heterogeneous multi-machines based on genetic programming according to the present invention, which specifically includes the following steps:

[0079] The present invention provides a method for heterogeneous multi-machine collaborative scheduling based on genetic programming, comprising the following steps:

[0080] Step S1, initializing the parameters of the heterogeneous multi-machine system scheduling problem, includes the following steps:

[0081] Step S11, setting basic information of farmland and agricultural machinery;

[0082] Define the farmland operation set F = {f1, f2, ..., f n}, f i represents the i-th job, i∈[1,n], n represents the number of farmlands; the farmland attribute is represented by f i ={sf, of, lfx, lfy}, where sf represents the area of ​​the current farmland; of represents the operation sequence required for the current farmland; lfx represents the longitude coordinate of the current farmland; lfy represents the latitude coordinate of the current farmland;

[0083] Define the operation sequence of the farmland i={o1, o2, ..., o x}, ok represents the kth process, k∈[1,x], x represents the number of processes in any farmland; the attributes of each process in the operation sequence are expressed as ok={jo,co}, where jo represents the number of the farmland to which the current process belongs; co represents the type of the current process;

[0084] Define the agricultural machinery set M = {m1, m2, ..., mp}, mj represents the jth agricultural machinery, j∈[1, p], p represents the total number of agricultural machinery; the agricultural machinery attributes are represented by mj = {tm, em, ms, lmx, lmy}, where tm represents the type of agricultural machinery; em represents the efficiency of agricultural machinery; ms represents the transfer speed of agricultural machinery; lmx represents the current longitude coordinate of agricultural machinery; lmy represents the current latitude coordinate of agricultural machinery;

[0085] Step S12, setting population number, population size, evolutionary generation, initial exchange generation and reproduction parameters;

[0086] Step S13, dividing the set into training set and test set.

[0087] Step S2, generating an initial population according to the operation sequence of the farmland and the set of agricultural machinery;

[0088] The individuals in genetic programming are tree structures, consisting of terminal nodes and non-terminal nodes. Terminal nodes represent program attributes and are selected from the terminal set, and non-terminal nodes represent functions and are selected from the function set. The specific combination process is implemented by the ECJ framework.

[0089] The initial population size is 1024, including 1 main population and 2 auxiliary populations.

[0090] Step S3, calculating the agricultural machinery dispatching efficiency as the first fitness value according to the sum of the agricultural machinery operation time and the transfer time and the sum of the areas of all processes of the farmland;

[0091] The calculation formula of the first fitness value is as follows:

[0092]

[0093] Among them, maxEfficiency represents the efficiency of agricultural machinery scheduling, that is, the first fitness value; represents the total area of ​​n farmlands; represents the total scheduling time of n farmlands; PT i Represents the processing time of farmland operations; RM i Indicates the agricultural machinery transfer time; S k It represents the total effective dispatch area of ​​a certain farmland;

[0094] S k The calculation formula is as follows:

[0095]

[0096] Among them, SF q Indicates the area required for the corresponding operation in the farmland;

[0097] PT i The calculation formula is as follows:

[0098]

[0099] Among them, SF k represents the area of ​​the operation corresponding to the kth process; EM k represents the working efficiency of the agricultural machinery corresponding to the kth process;

[0100] RM i The calculation formula is as follows:

[0101]

[0102] Among them, D represents the distance required to transfer from the previous farmland to the current farmland; V represents the transportation speed of the agricultural machinery in the current process.

[0103] When calculating the first fitness value, the following constraints must be met:

[0104] StarTO k-1 <FinTO k-1 <StarTO k <FinTO k ,

[0105] Completion time (O ik ) = start time (O ik )+t ik ,

[0106] st k >max(ft μ μ∈P(k));

[0107]

[0108] Among them, StarTO k-1 Indicates the start time of the k-1th process; FinTO k-1 Indicates the end time of the k-1th process; StarTO k Indicates the start time of the kth process; FinTO k represents the end time of the kth process; O ik represents the kth process of the ith job; t ikrepresents the processing time of the kth process of the ith job; st k is the start time of process k, and process k has a set of processes that must be completed first, P(k); ft μ represents the end time of the process in P(k); x j represents the number of processes handled by the j-th agricultural machine at a certain moment; y i represents the completion status of the i-th farm operation; x ik Indicates the completion status of the kth process of the ith job.

[0109] Step S4, determining whether the population migration conditions are met;

[0110] When the population migration condition is met, a migration operation is performed according to the first fitness value to obtain a population after migration, and the fitness value of the population after migration is recalculated to obtain a second fitness value;

[0111] When the population migration condition is not met, jump to step S5.

[0112] Perform migration operation according to the first fitness value to obtain the population after migration. The specific operation is:

[0113] The first fitness value is ranked from small to large, and the first one-third of the individuals are selected as excellent individuals, and the remaining two-thirds of the individuals are ordinary individuals;

[0114] Given the initial migration rate mrate and migration interval interval and their respective changes g and h, migration is carried out according to the migration rate and migration interval, the common individuals of the main population are exchanged with the individuals of the auxiliary population, and the average fitness change is counted after the migration is completed to adaptively adjust the migration interval and migration rate;

[0115] When the migration operation is completed, if this migration evolves in a better direction than the previous migration, the migration interval will be increased and the migration rate will be reduced, that is:

[0116] mrate'=mrate-g;

[0117] interval'=interval+h;

[0118] Otherwise, reduce the migration interval and increase the migration rate, that is:

[0119] mrate'=mrate+g;

[0120] interval'=interval-h;

[0121] Wherein, mrate' is the new migration rate; interval' is the new migration interval.

[0122] The calculation formula of the second fitness value is as follows:

[0123] Fitness2=totalTime / totalProcessArea;

[0124] Among them, Fitness2 is the second fitness value; totalTime is the time consumed in the entire operation process; totalProcessArea is the sum of the operation areas of each farmland, and the calculation formula of totalProcessArea is as follows:

[0125]

[0126] Step S5, performing crossover mutation and deduplication operations on the initial population or the migrated population according to the first fitness value or the second fitness value to form a new population;

[0127] The crossover mutation operation is as follows:

[0128] Reproduction selection is performed through pre-set crossover rate and mutation rate;

[0129] Eligible parents are selected for breeding operations using a tournament selection algorithm of size 6;

[0130] Call the corresponding reproduction operation through the ECJ framework to reproduce;

[0131] The calculation formula of the crossover rate is as follows:

[0132] crossRate=c0+((c1-c0) / (evolveSum×a))×nowGeneration;

[0133] Among them, c0 and c1 are the initial crossover rate and the final crossover rate respectively; a is the convergence point; evolveSum represents the total evolutionary generation; nowGeneration is the current generation;

[0134] The mutation rate is 1-crossRate.

[0135] Deduplication strategy: After the crossover and mutation operation is completed, check whether there are duplicate individuals. If so, perform deduplication operation to regenerate new individuals, reduce the proportion of duplicate individuals, and achieve the purpose of increasing individual diversity.

[0136] Generate a new population. The new population generated after the breeding operation is completed will cover the original population.

[0137] a. After each breeding operation, place the new individuals into a new population.

[0138] b. When the number of individuals in the new population reaches the predetermined size, the breeding operation is terminated and the new population covers the original population.

[0139] Step S6, repeating steps S3 to S5 until the maximum number of iterations is reached, the evolution process ends, and the optimal agricultural machinery scheduling rule is obtained;

[0140] Step S7: Run the optimal agricultural machinery scheduling rule obtained in step S6 on the test data set to obtain the final agricultural machinery scheduling plan.

[0141] like Figure 2 As shown, it is a simplified schematic diagram of the execution steps of the present invention, and the specific meanings are:

[0142] This diagram is a simplified diagram of the dispatching process. The triangle symbol is the location of the dispatching center, where the starting and ending points of all agricultural machinery are located. The circular symbol is the location of the farmland plot, and the line segments of different shapes represent the operation routes of the corresponding types of agricultural machinery. The required operation types of the farmland are shown in the table in the figure. The farmland cannot be cultivated before it is harvested. For example, the operation process of farmland 2 needs to be harvested by harvester 2 before the land can be cultivated. The farmland is considered to have completed the task after completing all the required steps.

[0143] like Figure 3 As shown, it is a schematic diagram of the priority rule representation scheme of the present invention, and the specific meaning is:

[0144] The present invention uses a tree structure to represent the rule scheme, which is used to make decisions at decision points. For example, when the agricultural machine completes the current work and the waiting queue is not empty, a decision needs to be made to select the next operation to be performed. The rule scheme consists of the terminal part T β and function part F ε Composition (β and ε are both positive integers). During the initialization process, corresponding elements are extracted from the feature set and function set to form an individual tree, and a predetermined number of elements are repeatedly generated to form the initial population for continuous evolution in subsequent algorithms.

[0145] The present invention screens and proposes terminal set contents suitable for the agricultural machinery scheduling model. The detailed contents of the feature set are shown in Table 1, where the elements represent the common characteristics in the agricultural machinery scheduling problem model. In addition, the contents of the function set are shown in Table 2.

[0146] The present invention will be further described below through specific examples.

[0147] In order to make the algorithm reasonable and explainable, this embodiment designs experiments in four different scenarios.

[0148] The experimental setting has 3 candidate machines for each process, a total of 15 machines are tested, and the failure factors of the machine during work or driving are not considered. The number of farmlands is 50, and the dispatch center is located in the regional center. In order to verify whether the proposed AMGP can improve the experimental results, a variety of algorithms are compared, including standard genetic programming GP, genetic algorithm GA, multi-population genetic programming MSGP and particle swarm algorithm PSO.

[0149] In order to verify the stability and effectiveness of the AMGP algorithm and its versatility in different scenarios, four different scenarios were designed and tested with the number of farmlands and the location of the dispatch center as variants:

[0150] Scenario 1: 80 farmlands and the dispatch center is located in the center of the region;

[0151] Scenario 2: 50 farmlands and the dispatch center is located in the center of the region;

[0152] Scenario 3: 80 farmlands and the dispatch center is located at the edge of the area;

[0153] Scenario 4: 50 farmlands and the dispatch center is located at the edge of the area.

[0154] The four groups of variable experiments all take the shortest scheduling time as the optimization goal. Ten independent repeated experiments were conducted on AMGP, GP, GA, MSGP and PSO, and the experimental results are shown in Tables 3, 4, 5 and 6. The parameters used in the experiment are shown in Table 7.

[0155] Table 1 Common features of agricultural machinery scheduling models

[0156] property Notes T System current time NIQ The number of processes in the queue WIQ The total processing time of the operations in the queue PT Processing time of the process NPT Processing time of the next process WKR Treatment time for remaining farmland NOR The number of remaining operations WINQ Next queue processing time NINQ The number of processes in the next queue MWT Waiting time for agricultural machinery OWT Waiting time of process NWT Waiting time for the next machine to be ready rF Relative process end time rD Relative completion time SL time left DIST The distance between farmland and corresponding agricultural machinery FDD Estimated completion time of the process MRT Machine ready time NRT Next machine readiness time DD Farmland estimated completion time EFF Agricultural machinery operation efficiency

[0157] Table 2 Function set used in the agricultural machinery scheduling problem model

[0158]

[0159]

[0160] Table 3 Comparison of completion time of ten independent repeated experiments between the algorithm of the present invention (the first one) and the other four algorithms

[0161]

[0162]

[0163] Table 4 Comparison of the average completion time of the algorithm of the present invention (the first one) and the other four algorithms after ten independent repeated experiments

[0164] Table 5 Comparison of standard deviations of completion time of ten independent repeated experiments between the algorithm of the present invention (the first one) and the other four algorithms

[0165]

[0166] The experimental results are summarized and compared in Table 6.

[0167] Table 6 Average comparison of algorithms

[0168]

[0169] Table 7 AMGP parameters

[0170]

[0171] According to the information shown in Table 6, it can be verified by multiple independent repeated experiments:

[0172] 1) The dispatching time of AMGP with the dispatching center located in the central area is reduced by 6.775%, 16.01%, 5.165% and 11.14% compared with GP, GA, MSGP and PSO respectively;

[0173] 2) The dispatching time of AMGP with the dispatching center located in the edge area is reduced by 7.925%, 17.995%, 4.43% and 11.88% compared with GP, GA, MSGP and PSO respectively;

[0174] 3) When the number of farmland plots is 80, the scheduling time of AMGP is reduced by 6.865%, 21.225%, 4.275% and 16.24% compared with GP, GA, MSGP and PSO, respectively;

[0175] 4) When the number of farmland plots is 50, the scheduling time of AMGP is reduced by 7.835%, 12.78%, 5.31% and 6.78% compared with GP, GA, MSGP and PSO, respectively.

[0176] A heterogeneous multi-machine collaborative scheduling system based on genetic programming, comprising:

[0177] Initialization module, used to initialize the parameters of heterogeneous multi-machine system scheduling problem and generate the initial population;

[0178] The algorithm evolution module is used to evolve GP individuals according to the pre-set algorithm parameters to obtain the optimal agricultural machinery scheduling rules;

[0179] The result calculation module is used to calculate the final agricultural machinery scheduling plan based on the optimal agricultural machinery scheduling rules and test data sets.

[0180] Therefore, the present invention adopts the above-mentioned genetic programming-based heterogeneous multi-machine collaborative scheduling method and system, solves various types of cross-regional scheduling models of agricultural machinery by improving the genetic programming algorithm, solves the stereotyped problem of the existing traditional heuristic method, and can obtain efficient and generalized agricultural machinery scheduling rules, avoiding the influence of human subjective factors. Significantly improve economic benefits and efficiently perform continuous operations on crops.

[0181] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for heterogeneous multi-machine collaborative scheduling based on genetic programming, characterized in that: The following steps are involved: Step S1, initializing the parameters of the heterogeneous multi-machine system scheduling problem; Step S2, generating an initial population according to the operation sequence of the farmland and the set of agricultural machinery; Step S3, calculating the agricultural machinery dispatching efficiency according to the total area of ​​the farmland and the total dispatching time of the farmland as the first fitness value; Step S4, determining whether the population migration conditions are met; When the population migration condition is met, a migration operation is performed according to the first fitness value to obtain a population after migration, and the fitness value of the population after migration is recalculated to obtain a second fitness value; When the population migration condition is not met, jump to step S5; Step S5, performing crossover mutation and deduplication operations on the initial population or the migrated population according to the first fitness value or the second fitness value to form a new population; Step S6, repeating steps S3 to S5 until the maximum number of iterations is reached, the evolution process ends, and the optimal agricultural machinery scheduling rule is obtained; Step S7: Run the optimal agricultural machinery scheduling rule obtained in step S6 on the test data set to obtain the final agricultural machinery scheduling plan.

2. According to the genetic programming-based heterogeneous multi-machine collaborative scheduling method of claim 1, it is characterized in that: In step S1, the parameters of the heterogeneous multi-machine system scheduling problem are initialized, including the following steps: Step S11, setting basic information of farmland and agricultural machinery; Define the farmland operation set F = {f1, f2, ..., f n }, f i represents the i-th job, i∈[1,n], n represents the number of farmlands; the farmland attribute is represented by f i ={sf, of, lfx, lfy}, where sf represents the area of ​​the current farmland; of represents the operation sequence required for the current farmland; lfx represents the longitude coordinate of the current farmland; lfy represents the latitude coordinate of the current farmland; Define the operation sequence of the farmland i ={o1, o2, ..., o x }, ok represents the kth process, k∈[1,x], x represents the number of processes in any farmland; the attributes of each process in the operation sequence are expressed as ok={jo,co}, where jo represents the number of the farmland to which the current process belongs; co represents the type of the current process; Define the agricultural machinery set M = {m1, m2, ..., mp}, mj represents the jth agricultural machinery, j∈[1, p], p represents the total number of agricultural machinery; the agricultural machinery attributes are represented by mj = {tm, em, ms, lmx, lmy}, where tm represents the type of agricultural machinery; em represents the efficiency of agricultural machinery; ms represents the transfer speed of agricultural machinery; lmx represents the current longitude coordinate of agricultural machinery; lmy represents the current latitude coordinate of agricultural machinery; Step S12, setting population number, population size, evolutionary generation, initial exchange generation and reproduction parameters; Step S13, dividing the set into training set and test set.

3. The method for heterogeneous multi-machine collaborative scheduling based on genetic programming according to claim 2 is characterized in that: In step S2, the size of the initial population is 1024, including 1 main population and 2 auxiliary populations.

4. The method for heterogeneous multi-machine collaborative scheduling based on genetic programming according to claim 3 is characterized in that: In step S3, the calculation formula of the first fitness value is as follows: Among them, maxEfficiency represents the efficiency of agricultural machinery scheduling, that is, the first fitness value; represents the total area of ​​n farmlands; represents the total scheduling time of n farmlands; PT i Represents the processing time of farmland operations; RM i Indicates the agricultural machinery transfer time; S k It represents the total effective dispatch area of ​​a certain farmland; S k The calculation formula is as follows: Among them, SF q Indicates the area required for the corresponding operation in the farmland; PT i The calculation formula is as follows: Among them, SF k represents the area of ​​the operation corresponding to the kth process; EM k represents the working efficiency of the agricultural machinery corresponding to the kth process; RM i The calculation formula is as follows: Among them, D represents the distance required to transfer from the previous farmland to the current farmland; V represents the transportation speed of the agricultural machinery in the current process.

5. The method for heterogeneous multi-machine collaborative scheduling based on genetic programming according to claim 4 is characterized in that: In step S3, when calculating the first fitness value, the following constraints must be met: Completion time (O ik ) = start time (O ik )+t ik , st k >max(ft μ |μ∈P(k)); Among them, StarTO k-1 Indicates the start time of the k-1th process; FinTO k-1 Indicates the end time of the k-1th process; StarTO k Indicates the start time of the kth process; FinTO k represents the end time of the kth process; O ik represents the kth process of the ith job; t ik represents the processing time of the kth process of the ith job; st k is the start time of process k, and process k has a set of processes that must be completed first, P(k); ft μ represents the end time of the process in P(k); x j represents the number of processes handled by the j-th agricultural machine at a certain moment; y i represents the completion status of the i-th farm operation; x ik Indicates the completion status of the kth process of the ith job.

6. The method for heterogeneous multi-machine collaborative scheduling based on genetic programming according to claim 5, characterized in that: In step S4, a migration operation is performed according to the first fitness value to obtain a population after migration. The specific operation is: The first fitness value is ranked from small to large, and the first one-third of the individuals are selected as excellent individuals, and the remaining two-thirds of the individuals are ordinary individuals; Given the initial migration rate mrate and migration interval interval as well as their respective changes g and h, migration is carried out according to the migration rate and migration interval, exchanging the common individuals of the main population with the excellent individuals of the auxiliary population; When the migration operation is completed, if this migration evolves in a better direction than the previous migration, the migration interval will be increased and the migration rate will be reduced, that is: mrate'=mrate-g; interval'=interval+h; Otherwise, reduce the migration interval and increase the migration rate, that is: mrate'=mrate+g; interval'=interval-h; Wherein, mrate' is the new migration rate; interval' is the new migration interval.

7. The method for heterogeneous multi-machine collaborative scheduling based on genetic programming according to claim 6 is characterized in that: In step S4, the calculation formula of the second fitness value is as follows: Fitness2=totalTime / totalProcessArea; Among them, Fitness2 is the second fitness value; totalTime is the time consumed in the entire operation process; totalProcessArea is the sum of the operation areas of each farmland, and the calculation formula of totalProcessArea is as follows:

8. The method for heterogeneous multi-machine collaborative scheduling based on genetic programming according to claim 7, characterized in that: In step S5, the crossover mutation operation is as follows: Reproduction selection is performed through pre-set crossover rate and mutation rate; Eligible parents are selected for breeding operations using a tournament selection algorithm of size 6; Call the corresponding reproduction operation through the ECJ framework to reproduce; The calculation formula of the crossover rate is as follows: crossRate=c0+((c1-c0) / (evolveSum×a))×nowGeneration; Among them, c0 and c1 are the initial crossover rate and the final crossover rate respectively; a is the convergence point; evolveSum represents the total evolutionary generation; nowGeneration is the current generation; The mutation rate is 1-crossRate.

9. The method for heterogeneous multi-machine collaborative scheduling based on genetic programming according to claim 8, characterized in that: In step S5, the operation of forming a new population is as follows: After each breeding operation, new individuals are placed into a new population; When the individuals in the new population reach the predetermined size, the breeding operation is ended and the new population covers the original population.

10. A system for implementing the heterogeneous multi-machine collaborative scheduling method based on genetic programming as described in any one of claims 1 to 9.

Images