A method for agricultural machinery scheduling based on order resource sharing and agricultural machinery resource sharing

By establishing an agricultural machinery scheduling model that minimizes scheduling costs and a genetic algorithm for co-evolution of multiple groups, the problem of lack of scientific management and emergency management of agricultural machinery scheduling in the existing technology is solved, and the utilization rate and scheduling efficiency of agricultural machinery is improved, and the socialization of agricultural machinery is achieved. The service provides theoretical support.

CN114626718BActive Publication Date: 2024-11-08NORTHEAST AGRICULTURAL UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210251605.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-15
Publication Date
2024-11-08
Estimated Expiration
2042-03-15

AI Technical Summary

Technical Problem

The existing agricultural machinery scheduling methods mainly rely on manual formulation, and have not formed a scientific and systematic management model, cannot effectively deal with emergencies and emergencies during the operation process, and lack a general scheduling model.

Method used

Adopt the agricultural machinery scheduling method based on order resource sharing to establish a mixed integer linear planning model for agricultural machinery scheduling with maximization of operation volume based on the premise of minimizing scheduling cost. Combining dynamic order processing strategies, operation model selection strategies and emergency order insertion strategies, use Multiple group co-evolution genetic algorithms with heuristic rules solve the problem of agricultural machinery scheduling.

Benefits of technology

It improves the efficiency of operation order decision-making and the utilization rate of agricultural machinery resources, provides scientific agricultural machinery scheduling plans and resource layout plans, and provides theoretical support for the scale and professional development of agricultural machinery social services.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114626718B_ABST
    Figure CN114626718B_ABST
Patent Text Reader

Abstract

The invention discloses an agricultural machinery scheduling method based on order resource sharing and agricultural machinery resource sharing, including: establishing an agricultural machinery scheduling mixed integer linear programming model with the premise of minimizing scheduling cost and maximizing operation volume; planning the agricultural machinery path with the shortest agricultural machinery driving path as the objective function for the received orders, and arranging the operation tasks of the agricultural machinery according to the time window in the order; when the agricultural machinery leaves the agricultural machinery point to perform farmland operations, the received new orders are used as optional tasks, and the operation tasks of the agricultural machinery are arranged according to the maximum operation area with the maximum number of farmers served as the objective function of the current stage; a multi-population co-evolution genetic algorithm based on heuristic rules with a dynamic order processing strategy, an operation machine type selection strategy and an emergency order insertion strategy is used to solve the agricultural machinery scheduling mixed integer linear programming model, and the operation requirements are allocated according to the operation priority of the agricultural machinery point. The invention improves the decision-making efficiency of operation orders and the utilization rate of agricultural machinery resources.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of agricultural machinery scheduling, and in particular to an agricultural machinery scheduling method based on order resource sharing and agricultural machinery resource sharing. Background Art

[0002] With the increasing number of agricultural machinery service organizations in rural areas of my country and the further advancement of agricultural modernization, the degree of agricultural mechanization in rural areas has been continuously improved. At the same time, inspired by the "Internet +" model and the sharing economy model and the advancement of communication technology, the emergence of "online agricultural machinery" has brought greater convenience to the marketization of agricultural machinery. Although this method has solved some problems in the utilization of agricultural machinery resources, the scheduling plan still relies mainly on manual formulation, and a scientific and systematic management model has not been formed.

[0003] Combine harvester cross-regional operation is a socialized agricultural operation service with Chinese characteristics that was created in 1996. It uses the time difference of wheat harvesting at different latitudes to carry out mechanized harvesting across regional boundaries. With the rapid progress of agricultural machinery modernization and the valuable suggestions put forward by many scholars through research and practice, agricultural machinery scheduling has been continuously developed and promoted. However, most of the existing research on agricultural machinery scheduling is focused on static problems, and the sudden and emergency situations in the operation process are not fully considered, and no universal scheduling model has been formed. Summary of the invention

[0004] The purpose of the present invention is to solve the deficiencies in the prior art and to provide an agricultural machinery scheduling method based on order resource sharing and agricultural machinery resource sharing.

[0005] To achieve the above object, the present invention is implemented according to the following technical solutions:

[0006] An agricultural machinery scheduling method based on order resource sharing and agricultural machinery resource sharing comprises the following steps:

[0007] Step 1: Establish a mixed integer linear programming model for agricultural machinery scheduling that maximizes the workload under the premise of minimizing the scheduling cost;

[0008] Step 2: Plan the agricultural machinery path based on the received orders with the shortest agricultural machinery driving path as the objective function, and arrange the operation tasks of the agricultural machinery according to the time window in the order;

[0009] Step 3: When the agricultural machinery leaves the agricultural machinery point to carry out farmland operations, the received new orders are taken as optional tasks, and the objective function of the current stage is to serve the most farmers, and the operation tasks of the agricultural machinery are arranged according to the maximum operation area;

[0010] Step 4: Use a heuristic rule-based multi-population co-evolutionary genetic algorithm with dynamic order processing strategy, operation machine selection strategy and emergency order insertion strategy to solve the mixed integer linear programming model of agricultural machinery scheduling and allocate operation requirements according to the operation priority of agricultural machinery points.

[0011] Furthermore, the mixed integer linear programming model for agricultural machinery scheduling in step 1 is defined as G=(V, A),

[0012] Where: V is the set of all nodes, V = {v1,v2,…,v n ,v n+1 ,…v n+m}, including each agricultural machinery cooperative and each farmland operation point, A is the set of all paths, A={(v i ,v j )|v i ,v j ∈V,i≠j}.

[0013] Furthermore, the objective function in step 2 is:

[0014]

[0015] Based on the actual agricultural machinery scheduling problem and the model assumptions, the constraints are:

[0016] (1) The number of agricultural machinery owned by each agricultural machinery station is limited, and the number of agricultural machinery dispatched cannot exceed the total amount of agricultural machinery owned by the agricultural machinery station:

[0017]

[0018] (2) One agricultural machine can only provide services for one operation point, that is:

[0019]

[0020] (3) The needs of each operation point can be met and can only be served once by one agricultural machine. Multiple services from multiple agricultural machines are not accepted, that is:

[0021]

[0022] (4) After each agricultural machine is set off, it cannot travel between agricultural machine points. The route between agricultural machine points is invalid. It can only set off to the operation point, that is:

[0023]

[0024] (5) The agricultural machinery must leave the current operation point after completing the operation task at the current operation point, that is:

[0025]

[0026] (6) After completing the tasks at each operation point, agricultural machinery cannot be parked at will, or parked at other agricultural machinery points. It must return to the starting agricultural machinery point, that is:

[0027]

[0028] (7) The path planning of agricultural machinery should meet the time window requirements of the farmers to whom the initial order belongs, and should not be violated unless in an emergency, that is:

[0029]

[0030] In the formula is the time when agricultural machine k starts working at operation point i;

[0031] (8) The time window from the current operation point to the next operation point of the agricultural machinery should meet the following conditions, namely:

[0032]

[0033] Where M is a large positive number;

[0034] (9) The total operating time of each agricultural machine on a given day shall meet the following conditions: the sum of the driving time and operating time of the agricultural machine on that day shall not exceed the prescribed working time, namely:

[0035]

[0036] (10) The workload of agricultural machinery at the operating point satisfies the following relationship, namely:

[0037]

[0038] Furthermore, the objective function in step three is:

[0039]

[0040] The path is divided into five time periods A→B→C→D→E according to the operable time. The operation points that have participated in the planning are divided into four insertable spaces. Constraints are set for these four parts and the entire path:

[0041] (1) In the A→B section, the agricultural machine starts from the operation point A but cannot return to the operation point A, and the agricultural machine must end at the operation point B and cannot leave the operation point B, that is:

[0042]

[0043] In the A→B section, the time when the agricultural machinery arrives at the operation point B must be earlier than the upper limit of the time window of the operation point B, that is:

[0044]

[0045] (2) Similarly, the constraint expression in the B→C segment is:

[0046]

[0047] (3) The constraint expression in the C→D segment is:

[0048]

[0049]

[0050] (4) The constraint expression in the D→E segment is:

[0051]

[0052] In the planning of the entire path, an inserted operation point can only be inserted into one time period to avoid repeated insertion;

[0053] (5) Avoid inserting the same task repeatedly at different times:

[0054]

[0055]

[0056] (6) Ensure that a job task is executed at most once in the entire path:

[0057]

[0058] (7)ρ i is the time to reach operation point i:

[0059]

[0060]

[0061] Furthermore, the specific steps of solving the mixed integer linear programming model for agricultural machinery scheduling in step 4 are:

[0062] Step 1: Integer coding is used, 1 to n are farmland operation points, (n+1) to (n+m) are agricultural machinery points, and individuals that do not meet the time window constraints and operation time constraints in the initial population are eliminated according to the decoding rules. At this time, all individuals in the initial population are feasible solutions;

[0063] Step 2: Calculate the fitness value of the individual. The fitness function uses the inverse of the objective function and arranges them in descending order of fitness value.

[0064] Step 3: Divide the problem into an elite sub-population and two ordinary populations. Each sub-population evolves into a sub-population according to the set crossover probability and mutation probability.

[0065] Step 4: Calculate the fitness value of the new generation of individuals to determine whether the evolution meets the termination condition. If it does, then terminate and output the optimal solution; otherwise, go to Step 4;

[0066] Step 5: Based on the initial scheduling plan and the constraints of the agricultural machinery's operating capacity and order time window, the nearest neighbor search algorithm is used to plan the dynamic orders according to the principle of the closest distance between the new order and the current operating point of the agricultural machinery, and a dynamic scheduling plan is generated;

[0067] Step 6: For the selection of agricultural machinery models, the improved saving algorithm is used to allocate appropriate models according to the principle of minimum agricultural machinery power consumption per unit operating area, and the time window constraint of the initial order cannot be violated. At this time, the agricultural machinery can select the path according to the feasible solution of the initial solution of the planning concentration;

[0068] Step 7: When an agricultural machinery failure causes the operation to stop, the agricultural machinery with the shortest distance and the smallest fluctuation in the overall plan caused by the path transfer is urgently dispatched to the rescue until the faulty agricultural machinery can return to participate in the dispatch operation.

[0069] Furthermore, the specific steps of the dynamic order processing strategy are:

[0070] Step 1: Calculate the transportation distance between the inserted node and each pair of nodes including the operation point and the agricultural machinery service point;

[0071] Step 2: Select a pair of nodes with the shortest transmission distance;

[0072] Step 3: Connect nodes from a start node to an end node by visiting the sequence in ascending order of the earliest time of the node;

[0073] Step 4: Connect the next node to the previously connected end node as much as possible to form a route until all nodes are assigned to different routes;

[0074] Step 5: Arrange the route so that the working vehicle starts the access sequence of this period after finishing the access sequence of the previous period;

[0075] Step 6: Assign vehicles to routes in sequence until the demand and time windows of all orders are met.

[0076] Furthermore, in step 4, the operation machine type selection strategy is to use an improved conservation algorithm to determine the agricultural machinery driving route and agricultural machinery model, specifically including:

[0077] First, by planning the orders that have been received, a specific path has been determined for each farmland operation point to perform the operation task. These routes will be combined and re-optimized according to the constraints; at this time, a travel time can be calculated for each path relative to different types of agricultural machinery, and the minimum travel time is recorded as T d :

[0078] S jw : The time saved by inserting the work point w into the path where the work point j is located;

[0079] t jw : The travel time between work point w and work point j (assuming t jw =t wj );

[0080] t dw : The travel time between the agricultural machinery point d and the operation point w;

[0081] t dj : The travel time between the agricultural machinery point d and the operation point j.

[0082] Therefore,

[0083] S jw =t dw +t dj -t jw ;

[0084] S jw The routes are sorted in descending order of numerical value, with the routes with higher ranking having greater probability of being selected, and then the largest S jw The relevant paths begin to be optimized;

[0085] If the new path satisfies the problem constraints and does not form a sub-route, and can satisfy the time window constraints as follows, the insertion of the new order can be accepted, otherwise the insertion of the new operation point will be rejected;

[0086]

[0087] Among them, T w represents the time when agricultural machine k visits work point w, represents the operation time of agricultural machine k at operation point w, t wj represents the transfer time of agricultural machine k between operation point w and operation point j, l j It represents the latest time allowed for the start of operation at operation point j;

[0088] Compare the total workload of path r with the operating power of each type of agricultural machinery, and finally select the vehicle. If more than one type of agricultural machinery can be selected based on the demand quantity, the most economical choice The calculation is as follows, and the minimum Assigned to route r;

[0089] The power consumed by the optimized agricultural machinery to complete a unit of work;

[0090] Q r : Path Q r The total amount of all work;

[0091] P k : The amount of work done by this model of agricultural machinery per unit time, that is, the power of the agricultural machinery.

[0092] Therefore

[0093]

[0094] Agricultural machinery of different power can undertake different workloads, but the smaller the power consumed to complete the unit workload, the more matching and economical the scheduling scheme is. Therefore, different agricultural machinery-path matching combinations are calculated, and the path is taken to the minimum The more solutions you get, the better ones you get.

[0095] Compared with the prior art, the present invention proposes an agricultural machinery scheduling mode based on order resource sharing and agricultural machinery resource sharing in view of the current supply and demand situation of agricultural machinery scheduling, comprehensively considers the influencing factors and characteristics of the agricultural machinery scheduling process, and establishes a two-stage agricultural machinery scheduling model with dynamic demand. The model adopts a scheduling strategy that combines soft time windows and hard time windows as well as real-time insertion of emergency orders and batch processing of dynamic demand. The optimization solution algorithm and the solution program are improved and designed according to the characteristics of the model. The initial scheduling plan is generated by a multi-population co-evolution genetic algorithm with heuristic rules. The dynamic demand batch insertion of the nearest neighbor search strategy is used to process new orders in each time period according to the emergency situation and priority rules of the order. The agricultural machinery model selection method is formulated using a saving algorithm. Finally, an agricultural machinery operation status index is established to specify rescue agricultural machinery to handle emergency situations at the minimum cost. The decision-making efficiency of operation orders and the utilization rate of agricultural machinery resources are improved.

[0096] The present invention can provide effective and feasible scheduling solutions for managers and operators of agricultural machinery service organizations, and also provide a strong basis for the selection of cross-regional agricultural machinery operations and the layout of agricultural machinery resources, and provide theoretical support for the large-scale and professional development of agricultural machinery social services. BRIEF DESCRIPTION OF THE DRAWINGS

[0097] Figure 1 This is the flow chart of the co-evolutionary genetic algorithm.

[0098] Figure 2 Plan a timing diagram for the order.

[0099] Figure 3 Insert schematic for dynamic order.

[0100] Figure 4 For path r i Complete solution set for 5 periods in the current state.

[0101] Figure 5 This is the cost trend chart of the initial scheduling solution.

[0102] Figure 6 Schedule initial paths for agricultural machinery.

[0103] Figure 7 Provide dynamic paths for agricultural machinery for each agricultural machinery cooperative. DETAILED DESCRIPTION

[0104] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments. The specific embodiments described herein are only used to explain the present invention and are not used to limit the invention.

[0105] Based on the current status of rural operations and agricultural machinery development, the following explanation is given to the agricultural machinery scheduling problem:

[0106] (1) Multiple agricultural machinery service organizations participate in the same collaborative scheduling process. The agricultural machinery service organizations are used as agricultural machinery points for agricultural machinery to depart and return. Multiple farmland operation points are dispersed in different areas of the scheduling area. Therefore, the problem to be solved in this embodiment is to generate a path optimization solution for driving from multiple agricultural machinery points to multiple agricultural machinery operation points for farmland operations.

[0107] (2) Each piece of agricultural machinery in operation is equipped with necessary communication equipment to facilitate the transmission of information between the dispatch center, agricultural machinery operators and farmers.

[0108] (3) Each operation point will only be served by one agricultural machine once. The geographical location and area of ​​the farmland and other related information are all known, and the coordinates of all operation locations and parking lots are also known. If the distance of the operation route is known, then the cost of the route is also known. Each agricultural machine must return to the parking lot where it departed for the day.

[0109] (4) During normal driving, agricultural machinery is not restricted by factors such as traffic congestion. The driving speed is uniform, and the agricultural machinery can obtain dynamic information related to roads and weather during driving. In other words, the system can estimate the most likely driving time of the agricultural machinery between two work points.

[0110] (5) Agricultural machinery service organizations should estimate the number of agricultural machinery to be used based on the local farmland planting area. They can also open the order receiving system a certain time in advance and expand the scope of agricultural machinery dispatch when necessary.

[0111] (6) In principle, the agricultural machinery must operate on the farmland within the time window specified by the farmer. It cannot be earlier than the earliest start time allowed by the farmer, nor later than the latest end time allowed by the farmer. The start time of each planned operation point must strictly comply with the time window [e i ,l i ] requires that the time window requirement can be violated if and only if an urgent job is inserted.

[0112] A mixed integer linear programming model for agricultural machinery scheduling with the premise of minimizing scheduling cost and maximizing workload is established. The assumptions of the model are as follows:

[0113] (1) The geographical locations of agricultural machinery points at different locations within the dispatching area are known and fixed;

[0114] (2) Each agricultural machinery site has different types of agricultural machinery. The number of each type of agricultural machinery is known. The operating power and operating cost of the same type of agricultural machinery are also known. The impact of the age of agricultural machinery on operating power is not considered for the time being.

[0115] (3) The geographical location, operation area, and operation time window requirements of the farmland in the order submitted by the farmer are all known, and the distances between all agricultural machinery points and farmland operation points are known;

[0116] (4) The needs of each operation point must be met and can only be visited by one agricultural machine at a time. In order to standardize the management and maintenance of agricultural machinery, as well as the calculation of subsequent operation revenue, each agricultural machine must return to the agricultural machinery point after completing the day's operation task;

[0117] (5) In the process of generating the initial plan, it is assumed that the agricultural machinery involved in the scheduling is sufficient to allocate an operating agricultural machinery for each operation path;

[0118] (6) The agricultural machinery in the model has no capacity limit. The research object is only the operating agricultural machinery. The operating time of the agricultural machinery consists of the transfer time between operating points and the operating time of the operating point. The total working time of the agricultural machinery is set to 10 hours per day, and the inflow of materials and the output of agricultural products are not considered for the time being.

[0119] In actual research, it is found that the current operating power of agricultural machinery can already meet the operating needs of farmers. If the span of the time window is too long, the operating time of the agricultural machinery will be included in the time window. If the span of the time window is shortened, multiple changes in the scheduling plan will cause confusion in the agricultural machinery path. Combined with the working hours of agricultural machinery operations, this embodiment divides the operating time of agricultural machinery into 5 discrete time windows of 2 hours, thereby achieving an ideal order response state. The division of each time window is shown in Table 1:

[0120] Table 1

[0121]

[0122] A mixed integer programming model is established to describe the agricultural machinery scheduling problem. G = (V, A) is defined as a complete graph that can fully represent the system composed of the problem. The symbols involved in the model and their meanings are as follows:

[0123] (1) Collection

[0124] V is the set of all nodes, V = {v1,v2,…,v n ,v n+1 ,…v n+m}, including various agricultural machinery cooperatives and farmland operation sites;

[0125] D is the set of parking lots, D={v n+1 ,v n+2 ,…,v n+m} represents m parking lots (agricultural machinery cooperatives);

[0126] C is the set of all customers, C = {v1, v2, ..., v n} represents n customers (farm operation points);

[0127] A is the set of all paths, A={(v i ,v j )|v i ,v j ∈V,i≠j};

[0128] K is the set of all agricultural machinery models k owned by all agricultural machinery cooperatives in the region, K = {1, 2, …, h};

[0129] (2) Parameters

[0130] K d : Total amount of agricultural machinery at the agricultural machinery point;

[0131] T: working time of each agricultural machine;

[0132] t ij: The travel time of agricultural machinery from operation point i to operation point j;

[0133] The operation time of agricultural machine k at operation point i;

[0134] c ij : The cost of agricultural machinery traveling per unit distance;

[0135] v ij : Transfer speed of agricultural machinery;

[0136] The number of agricultural machinery k models in use;

[0137] p k : operating efficiency of agricultural machinery k model;

[0138] The accumulated operation volume of agricultural machine k at the current operation point i.

[0139] (3) Variables

[0140] is a 0-1 decision variable, indicating that operation point i is served by agricultural machine k;

[0141] is a 0-1 decision variable, indicating that if agricultural machine k drives directly from operation point i to operation point j and starts to provide services, then otherwise

[0142] The first stage: The received orders are routed with the shortest agricultural machinery route as the objective function. In this process, the operation tasks of agricultural machinery are arranged strictly according to the time window in the order. The objective function is:

[0143]

[0144] Based on the actual agricultural machinery scheduling problem and the model assumptions, the following constraints are established:

[0145] (1) The number of agricultural machinery owned by each agricultural machinery station is limited, and the number of agricultural machinery dispatched cannot exceed the total number of agricultural machinery owned by the agricultural machinery station.

[0146]

[0147] (2) One agricultural machine can only provide services for one operation point, that is:

[0148]

[0149] (3) The needs of each operation point can be met and can only be served once by one agricultural machine. Multiple services from multiple agricultural machines are not accepted, that is:

[0150]

[0151] (4) After each agricultural machine is set off, it cannot travel between agricultural machine points. The route between agricultural machine points is invalid. It can only set off to the operation point, that is:

[0152]

[0153] (5) The agricultural machinery must leave the current operation point after completing the operation task at the current operation point, that is:

[0154]

[0155] (6) After completing the tasks at each operation point, agricultural machinery cannot be parked at will, or parked at other agricultural machinery points. It must return to the starting agricultural machinery point, that is:

[0156]

[0157] (7) The path planning of agricultural machinery should meet the time window requirements of the farmers to whom the initial order belongs, and should not be violated unless in an emergency, that is:

[0158]

[0159] In the formula is the time when agricultural machine k starts working at operation point i.

[0160] (8) The time window from the current operation point to the next operation point of the agricultural machinery should meet the following conditions, namely:

[0161]

[0162] Where M is a large positive number.

[0163] (9) The total operating time of each agricultural machine on a given day shall meet the following conditions: the sum of the driving time and operating time of the agricultural machine on that day shall not exceed the prescribed working time, namely:

[0164]

[0165] (10) The workload of agricultural machinery at the operating point satisfies the following relationship, namely:

[0166]

[0167] Phase 2: When the agricultural machinery leaves the agricultural machinery point to carry out farmland operations, the received new orders are taken as optional tasks, and the objective function of the current phase is to serve the most farmers and maximize the operating area. When the initial time window and operation constraints are met, the new orders are inserted into the current route of the agricultural machinery. In this process, the time window of the new order is used as a soft time window. If the operating area requirements of the current route can be met, the new order will be inserted. If not, the insertion will be rejected and the order will be retained as an order for the next working day waiting for planning.

[0168] Assume that A→B→C→D→E is the operation sequence of the first operation point in each period of the day, as shown in Table 2.

[0169] Table 2

[0170]

[0171] The variables and parameters defined are as follows:

[0172] (1) Collection

[0173] I is the operation point to be inserted into the path in all new orders;

[0174] (2) Parameters

[0175] The travel time of agricultural machinery from operation point i to operation point j;

[0176] The operation time of agricultural machine k at operation point i;

[0177] t i : The time when the agricultural machinery starts working at the operation point i;

[0178] l i : The time when the agricultural machinery arrives at the operation point i;

[0179] ρ i : The time when the agricultural machinery arrives at the operation point i.

[0180] (3) Variables

[0181] x i ,y i , z i , μ i Both are binary variables, indicating whether the job point is inserted into the job task queue of the scheduling scheme. If the insertion operation is accepted, the variable value is 1, and if the insertion operation is rejected, the variable value is 0.

[0182]

[0183] The number of operating points inserted in the initial plan is the largest, and the objective function is:

[0184]

[0185] Based on the initial scheduling scheme of this embodiment, the path is divided into 5 time periods according to the operable time, and the operation points that have participated in the planning are divided into 4 insertable spaces. Therefore, constraints are set for these 4 parts and the entire path as follows.

[0186] Constraints: (1) In the A→B section, the agricultural machine starts from the operation point A but cannot return to the operation point A, and the agricultural machine must end at the operation point B and cannot leave the operation point B, that is:

[0187]

[0188] In the A→B section, the time when the agricultural machinery arrives at the operation point B must be earlier than the upper limit of the time window of the operation point B, that is:

[0189]

[0190] (2) Similarly, the constraint expression in the B→C segment is:

[0191]

[0192] (3) The constraint expression in the C→D segment is:

[0193]

[0194]

[0195] (4) The constraint expression in the D→E segment is:

[0196]

[0197]

[0198] In the planning of the entire path, an inserted operation point can only be inserted into one time period to avoid repeated insertion.

[0199] (5) Avoid repeatedly inserting the same task at different time periods.

[0200]

[0201]

[0202] (6) Ensure that a job task is executed at most once in the entire path.

[0203]

[0204] (7)ρi is the time to reach operation point i.

[0205]

[0206]

[0207] In the process of processing new orders, since the time window requirement of new orders in this embodiment is a soft time window constraint, the processing of new orders first takes the minimum cost as the goal, and comprehensively considers the time window constraint under the premise that the agricultural machinery operation capacity meets the constraint. Once the farmer's order is scheduled for a certain period of operation on the same day, it cannot be changed again unless an emergency order appears. The system needs to send the operation time to the farmer through the communication system after making a new plan.

[0208] If a job order is [t i-1 ,t i ] time period, and can be served in the next optimized time period [t i ,t i+1 ] is successfully scheduled, then there is no need to immediately insert the order into the current path, as long as the order is scheduled in the period [t i ,t i+1 Therefore, this embodiment adopts the urgent order insertion and batch driving strategy, and can divide the order demand into two categories: urgent orders and non-urgent orders according to the urgency.

[0209] Assume that the farmer is at time t(t∈[t i-1 ,t i ]) sent a job request, [t i-1 ,t i ] and [t i ,t i+1 ] are two adjacent optimization time periods. If the order can be placed in [t i ,t i+1 ] period to the current path plan, which neither exceeds the operation time of the agricultural machinery nor meets the time window constraint of the current operation task, is called a non-urgent customer; on the contrary, if it cannot wait [t i ,t i+1 ] time period is inserted into the current route plan, it is called an emergency customer.

[0210] The overall agricultural machinery dispatch of the day is divided into several interconnected time intervals. Decisions are made on the acceptance and rejection of non-urgent orders and the priority insertion of urgent orders in each stage in the order of each time interval, and the decision results are fed back to farmers and agricultural machinery operators respectively. Farmers make preparations before the operation or arrange other agricultural activities to wait for the arrival of agricultural machinery according to the decision results. Agricultural machinery operators continue to move to the next operation point according to the optimized path of the decision results, or change the operation route urgently to ensure the effective implementation of the decision plan.

[0211] When a job order refuses to be inserted into all the paths of the current scheduling plan, the job order will be retained in the initial scheduling plan of the next operation day for optimization, and the operation time window of the order will be used as the hard time window constraint scheduling plan to form a framework for dynamic order insertion for scheduling path optimization. When the number of rejected orders is large, the scheduling area should expand the scope of agricultural machinery scheduling or publish cross-regional operation demand information before starting the operation the next day to attract external agricultural machinery to participate in the operation.

[0212] The agricultural machinery scheduling problem in this embodiment is a large-scale global optimization problem. The algorithm has a large search space and many variable dimensions. The use of multiple subpopulations for co-evolution information can exchange information with each other to avoid the algorithm from falling into a local optimal solution. According to the model objective function expression, the operation paths and sequences of different types of agricultural machinery at different agricultural machinery sites need to be decided at the same time. The solution process of this embodiment is as follows:

[0213] (1) In the initial scheduling stage, the operation points are divided into different operation priorities relative to the agricultural machinery points according to the distance between the agricultural machinery points and the farmland operation points and between the farmland operation points. A threshold of the operation volume and operation duration is given to the priority allocation process to generate multiple operation paths and determine which agricultural machinery point should dispatch vehicles to perform the operation tasks at all locations;

[0214] (2) In the dynamic scheduling stage, the unfinished tasks in each time interval and the tasks received in that time period are counted as the initial state of the next time period planning. Routes with large workloads are allocated to high-power agricultural machinery with high probability, forming a complete scheduling plan.

[0215] (3) The operating status of the agricultural machinery is determined according to its driving speed and operating speed, thereby estimating the transfer time and operating time of the agricultural machinery. When an emergency order is inserted or the agricultural machinery fails, an alternative plan is called to replace the current operation plan to perform the operation task.

[0216] When evaluating individuals, individuals with higher fitness values ​​are called elite individuals. To prevent the current best individuals from disappearing during the next generation of evolution, this embodiment divides elite individuals and ordinary individuals into different populations for separate evolution. Individuals with higher fitness values ​​that appear during the evolution process replace the next generation of individuals with poor performance. The algorithm flow chart is shown in the figure below. Figure 1 shown.

[0217] The update operation of the subpopulation during the evolution process is as follows:

[0218] P μ and P ω Represents the two selected parent individuals, C μ and C ω represents the two new individuals evolved from these two parent individuals, P ν represents the individual before mutation, P' ν Represents the individual after mutation.

[0219] Using individual P μ and P ω The new individual C generated after the crossover operation μ and C ω To update the subpopulation, the update rules are as follows:

[0220]

[0221] Using individual P ν The new individual P' generated after the mutation operation ν To update the subpopulation, the update rules are as follows:

[0222]

[0223] The solution steps of the co-evolutionary genetic algorithm are:

[0224] Step 1: Generate an initial population of N individuals P(gen)|gen=0 by random method, where gen is the population evolution generation and the maximum evolution generation is maxgen. Call the decoding program to decode and eliminate infeasible solutions, calculate the individual fitness values ​​and arrange them in descending order.

[0225] Step 2: Divide the population P(t) into three populations, one elite population and two ordinary populations. The individuals occupying the first M positions are placed in the first population, called the elite population. The higher the individual fitness value, the closer to the optimal solution. The remaining (NM) individuals are randomly divided into two ordinary populations, each with a size of

[0226] Step 3: Select individuals, reorganize through crossover operation, and enhance population diversity through mutation operation. The common population adopts a higher mutation probability than the elite population.

[0227] Step 4: Evaluate the fitness of the offspring individuals. If the offspring individual has a higher fitness value, it will replace the parent individual with a lower fitness value. If the fitness values ​​of the offspring individuals are all lower than the fitness values ​​of the parent individuals, no individual replacement will be performed. Update the three sub-populations according to the individual update rule, gen = gen + 1;

[0228] Step 5: Merge the offspring elite population and the offspring common population. The merged population is recorded as P(gen+1), and is also sorted according to the fitness value;

[0229] Step 6: Determine whether the evolution process has reached the loop end condition. If so, the loop ends and the current population P(gen+1) is output. Otherwise, go to Step 2.

[0230] Since the agricultural machinery scheduling problem involves multiple agricultural machinery service organizations involved in the scheduling, and each agricultural machinery service organization has different models and numbers of operating agricultural machinery, the complexity of the scheduling problem and the difficulty of solving the algorithm are greatly increased. In addition, how to allocate farmland operation points in different locations not only determines the overall framework of the agricultural machinery operation path, but also affects the scheduling cost of the entire plan. Therefore, how to allocate the operation tasks in the initial order to different agricultural machinery points to achieve the goal of optimizing the scheduling plan is the most important part of the algorithm design. The operation steps of the scheduling algorithm are as follows:

[0231] Step 1: Input the location coordinate information of the agricultural machinery service organization, the parameter information such as the agricultural machinery model, operating power and operating cost, the location of the operating point and the operating area. In addition, there are also control parameters such as the elite population size, crossover probability, and mutation probability of the co-evolutionary genetic algorithm;

[0232] Step 2: Integer coding is used, 1 to n are farmland operation points, (n+1) to (n+m) are agricultural machinery points, and individuals that do not meet the time window constraints and operation time constraints in the initial population are eliminated according to the decoding rules. At this time, all individuals in the initial population are feasible solutions;

[0233] Step 3: Calculate the fitness value of the individual. The fitness function uses the inverse of the objective function and arranges them in descending order of fitness value.

[0234] Step 4: Divide the problem into an elite sub-population and two ordinary populations. Each sub-population evolves into a sub-population according to the set crossover probability and mutation probability.

[0235] Step 5: Calculate the fitness value of the new generation of individuals to determine whether the evolution meets the termination condition. If so, terminate and output the optimal solution; otherwise, go to Step 4;

[0236] Step 6: Based on the initial scheduling plan and the constraints of the agricultural machinery's operating capacity and order time window, the nearest neighbor search algorithm is used to plan the dynamic orders according to the principle of the closest distance between the new order and the current operating point of the agricultural machinery, and a dynamic scheduling plan is generated;

[0237] Step 7: For the selection of agricultural machinery models, the improved saving algorithm is used to allocate appropriate models according to the principle of minimum agricultural machinery power consumption per unit operating area, and the time window constraint of the initial order cannot be violated. At this time, the agricultural machinery can select the path according to the feasible solution of the initial solution of the planning concentration;

[0238] Step 8: When an agricultural machinery failure causes the operation to stop, the agricultural machinery with the closest distance and the smallest fluctuation in the overall plan caused by the path transfer is urgently dispatched to the rescue until the faulty agricultural machinery can return to participate in the dispatch operation.

[0239] Whether a new order is accepted depends not only on the remaining operating capacity of the agricultural machinery, but also on the operating costs and various events that may occur during the driving and operation of the agricultural machinery, such as weather factors that extend the operation time or poor road conditions that reduce the driving speed and the arrival of new orders. How to quickly respond to real-time dynamic information and implement response strategies is a very critical issue to improve operating efficiency. At the end of the time period, the order that needs to be executed in the time period is received. The order reception and decision-making planning sequence is as follows: Figure 2 shown.

[0240] By managing a planning set to respond to dynamic events in real time, the feasible solutions obtained by the algorithm are retained with some high-quality solutions. The structure of the planning set is shown in Table 3.

[0241] Assume that p * is the optimal solution in the current plan set. Before the agricultural machine starts working, all the path plans in the plan set are feasible, and these plans are arranged in descending order of the total work volume. After completing each work task at a work point and before leaving for the next work point, the agricultural machine must check whether the remaining driving route of the path is feasible. In this way, the plan set will be updated after the agricultural machine completes each work task.

[0242] Table 3

[0243]

[0244] The computational time of dynamic optimization depends on the frequency of re-optimization and the expected time for decision makers to respond to new information. The nearest neighbor search algorithm is used to construct a work path for each new work point. Different from the distance clustering method, this embodiment is implemented by selecting a pair of continuous work tasks for an agricultural machinery service point in the next planning period. First, determine the two work points closest to a new work order, and then select the next work point closest to the previous work point until all work points have established paths with a certain agricultural machinery point. Repeat the following heuristic process until all new orders in each group are assigned to a route. The order insertion diagram is shown in the figure. Figure 3 As shown, the specific steps are as follows:

[0245] Step 1: Calculate the transportation distance between the inserted node and each pair of nodes including the operation point and the agricultural machinery service point.

[0246] Step 2: Select a pair of nodes with the shortest transmission distance.

[0247] Step 3: Connect nodes from a start node to an end node by visiting the sequence in ascending order of the earliest time of the node.

[0248] Step 4: Connect the next node to the previously connected end node as much as possible to form a route, until all nodes are assigned to different routes.

[0249] Step 5: Arrange the route so that the working vehicle starts the access sequence of this period after completing the access sequence of the previous period.

[0250] Step 6: Assign vehicles to routes in sequence until the demand and time windows of all orders are met.

[0251] The current scheduling status is described as:

[0252] L k It represents the current position i of the agricultural machinery, that is, the operation task of operation point i has just been completed by the agricultural machinery.

[0253] S k Indicates that except L k Other work sites that require agricultural machinery to operate, S k ={S1,S2,…}.

[0254] T k is the current time.

[0255] Suppose there is a matrix representing the set of feasible solutions obtained by the algorithm.

[0256]

[0257] where b ij ∈I,i∈{1,2,…,n},j∈{n∈N|n≤N max}. In particular, when n = N max When the matrix is ​​full rank, it represents the optimal solution in the plan set. Each row contains the optional operation points in a time period. If the value is 0, the agricultural machine will go directly to the next operation point according to the original operation task order without waiting for new operation tasks. For example, given a plan matrix

[0258]

[0259] Then the complete solution in the current state is Figure 4 shown.

[0260] The purpose of managing the plan set is to find the most suitable feasible solution. Inserting a new order will not affect the initial operation task, and the agricultural machinery will arrive at the operation site and start the operation task according to the time window requirements. If the order insertion fails, it will not affect other agricultural machinery and completed operations. This part of the order will participate in the initial scheduling plan for the next operation scheduling, and the time window in the order is a hard time window and is not allowed to be violated.

[0261] The scheduling of dynamic orders can be uniformly processed according to the divided time periods to reduce the frequency of route updates and avoid the confusion caused by multiple instructions to agricultural machinery. The disadvantage is that these orders are not strictly started according to the time window required by farmers, which will reduce farmers' satisfaction.

[0262] The dynamic path optimization strategy uses the new order batch processing strategy for scheduling calculations. According to the divided scheduling cycle, the size of the optimization time interval and the time window sorting are set to determine whether the batch processing conditions are met. If so, the current state of the scheduling execution is accessed to extract the order operation demand information, the current operation information of the agricultural machinery, and the current path plan. The path optimization algorithm is used to replan the path, and then the currently executed scheduling plan is updated. After the planned path changes, the operation changes must be sent to the agricultural machinery operator immediately, and the order processing results must be sent to the farmers, so that agricultural machinery resources can be dispatched in time to carry out operations.

[0263] The order information is input into the database and waits for the optimization solution to be solved. If it is an urgent customer, the current state of the scheduling execution is also accessed, and the order operation demand information, the current operation information of the agricultural machinery and the current path plan are extracted. The customer is processed immediately, and then the scheduling plan is updated and the agricultural machinery is required to execute immediately. If the urgent customer cannot be served within the specified time window, it is necessary to consider sending a new agricultural machinery and adding it to the dynamic order processing process of the next period, because the activation cost of the agricultural machinery is also a part that needs to be considered in the scheduling plan.

[0264] At the same time, inserted non-urgent orders can only wait until the next time period to be processed. If there are too many failed insertion orders, customer satisfaction will be too poor. The dispatch center must consider plans to increase the number of agricultural machinery. It can be done by renting agricultural machinery from local farmers or by publishing agricultural machinery demand information to allocate agricultural machinery from other places to perform operations locally.

[0265] The agricultural machinery models and related information of all farmland operations are determined. Therefore, the purpose of this section is to determine the service routes and agricultural machinery models used for each route for each farmland of each farmer. In other words, through this step, it can be determined that the operation route should be assigned to a certain model of agricultural machinery for execution.

[0266] The operating costs and operating powers of different types of agricultural machinery are different, and the amount of work tasks they can undertake is also different. If there is no constraint on the type of agricultural machinery, many rejected orders will be generated in the subsequent dynamic solution process. The working time of each agricultural machinery is composed of the driving time between two operating points and the operating time at all operating points. In order to maximize the amount of work within the same working time, the existing agricultural machinery resources should be used to the maximum extent. The calculation process of the operating time is as follows:

[0267]

[0268] The power consumed by agricultural machinery to complete a unit of work;

[0269] Q k : The total amount of agricultural machinery operations on the current path;

[0270] P k : The amount of work done by this model of agricultural machinery per unit time, that is, the power of the agricultural machinery.

[0271]

[0272] An improved economy algorithm is used to determine the routes and models of agricultural machinery. First, by planning the orders that have been received, a specific path has been determined for each farmland operation point to perform the operation task. These routes will be combined and re-optimized based on the constraints. At this point, a travel time can be calculated for each path relative to different types of agricultural machinery, and the minimum travel time is recorded as T d .

[0273] S jw : The time saved by inserting the work point w into the path where the work point j is located;

[0274] t jw : The travel time between work point w and work point j (assuming t jw =twj );

[0275] t dw : The travel time between the agricultural machinery point d and the operation point w;

[0276] t dj : The travel time between the agricultural machinery point d and the operation point j.

[0277] Therefore,

[0278] S jw =t dw +t dj -t jw (4-11)

[0279] S jw The routes are sorted in descending order of numerical value, with the routes with higher ranking having greater probability of being selected, and then the largest S jw The relevant paths begin to be optimized.

[0280] If the new path satisfies the constraints of the problem and does not form a sub-route, and can satisfy the time window constraints of the following equation (4-12), the insertion of the new order can be accepted, otherwise the insertion of the new operation point will be rejected.

[0281]

[0282] Among them, T w represents the time when agricultural machine k visits work point w, represents the operation time of agricultural machine k at operation point w, t wj represents the transfer time of agricultural machine k between operation point w and operation point j, l j It represents the latest time allowed for the operation point j to start the operation.

[0283] Compare the total workload of path r with the operating power of each type of agricultural machinery, and finally select the vehicle. If more than one type of agricultural machinery can be selected based on the demand quantity, the most economical choice The calculation is as follows, and the minimum Assigned to route r.

[0284] The power consumed by the optimized agricultural machinery to complete a unit of work;

[0285] Q r : Path Q r The total amount of all work;

[0286] P k : The amount of work done by this model of agricultural machinery per unit time, that is, the power of the agricultural machinery.

[0287] Therefore

[0288]

[0289] Agricultural machinery of different power can undertake different workloads, but the less power is consumed to complete a unit of workload, the more matching and economical the scheduling scheme is. Therefore, different agricultural machinery-path matching combinations are calculated, and the path is taken to the minimum The more solutions you get, the better ones you get.

[0290] In the process of agricultural machinery scheduling, when the agricultural machinery has been dispatched, if there is a failure in the agricultural machinery that needs to be repaired, if the failure is serious, it has to wait for the rescue vehicle to tow it to the agricultural machinery repair center for repair, and the impact of weather changes may cause the operation efficiency to decrease or even be interrupted. Whenever a new job is submitted, the algorithm will check the urgency of the order. If it is an urgent order, it will be inserted into the current scheduling plan immediately. If it is a non-urgent order, it will be added to the planning area and wait to be inserted into the solution of the next time period. At this time, the dynamic order processing strategy will be called to improve the current scheduling plan. If the order cannot find an agricultural machinery to receive the order, it will be rejected for insertion. This dynamic process is repeated until there are no new service orders.

[0291] The operation points that appear as urgent orders are inserted into the current path first, and the remaining orders are determined by the remaining operation volume of the agricultural machinery and the time window operation constraints of the planned operation tasks to determine whether non-urgent orders are accepted. ξ When an agricultural machine encounters a fault and cannot be repaired and re-enters the operation process in the current scheduling period, the initial operation order undertaken by the agricultural machine in the current period will be immediately assigned to other agricultural machines that meet the constraints as an emergency order, and the dynamic order of the current period assigned to the agricultural machine, if there is no other agricultural machine to receive in the next period, the order will be postponed to the next scheduling plan or retained as the next working day as the initial order with a hard time window constraint. In order to reduce the rejection of orders, non-urgent orders are inserted into the planned path according to the soft time window constraint. If too many orders are rejected during the scheduling process, it is necessary to consider increasing the number of scheduled agricultural machines by leasing or publishing agricultural machinery needs.

[0292] In the dynamic programming algorithm, each visit to a node requires a corresponding state quantity to represent its state, which is defined as (i, p, R i , C, T p , S), the meanings of each part are as follows:

[0293] (1) i represents the farm operation point being visited;

[0294] (2) p represents the state quantity index of the visited node, through which the state quantity of the farmland operation point where the faulty agricultural machinery is located can be obtained;

[0295] (3)R i Indicates the remaining workload of the agricultural machinery accessing node i;

[0296] (4) C represents the distance cost of the agricultural machine after visiting node i;

[0297] (5)T p Indicates the time when the agricultural machine starts to access node i;

[0298] (6) S is used to record the set of nodes that the path has visited after visiting node i, S = {v j |j∈V}. Where v j =0 means node j has not been visited; v j =1 means that the agricultural machine has completed the visit to node j on the current path and does not need to visit again.

[0299] Assume that at time T p When agricultural machinery k stops working due to a fault at operation point i, the state index p of operation point i will be accessed to obtain the remaining operation volume and order priority distribution of agricultural machinery k in the current period. The scheduling of rescue agricultural machinery needs to be determined based on the minimum cost and minimization of the volatility of the plan, and it is necessary to visit other farmland operation points based on the access record S.

[0300] When the operation point i is inserted into the operation path r of the agricultural machine k′ k′ The time when the agricultural machine k′ completes the operation point i and arrives at the next operation point j cannot exceed the latest time of the order required time window, that is:

[0301]

[0302] in,

[0303]

[0304] Where, t i represents the travel time of the agricultural machine k′ that meets the rescue conditions to reach the operation point i, represents the time required for the agricultural machine k′ to complete the remaining work at the operation point i and satisfies the relationship of formula (4-14), t ij represents the travel time of agricultural machine k′ to the next operation point on the operation path after completing the rescue operation, l j It means that the time spent by the agricultural machine k′ in completing the rescue mission cannot be later than the latest arrival time of the original initial order in its operation path. Among the agricultural machines that meet the above constraints, select the plan with the least cost increase C(i,k,j)=argminr(i,k′,j) and insert it into the current path of the agricultural machine, change the scheduling plan and send it to the rescue agricultural machine for execution.

[0305] The newly generated state will inherit the access record of the previous state. If the agricultural machinery can complete the task within the required operation time, the scheduling plan will not be changed. Otherwise, the nodes that have not been visited by the agricultural machinery in the current period will be released, and these nodes will be reallocated as new orders to other agricultural machinery with unsaturated workload to undertake the task. The priority rules of the order remain unchanged until the agricultural machinery can re-participate in the arrangement of the task in the next period and re-plan the operation path.

[0306] In order to verify and evaluate the scheduling method of the present invention, this embodiment compiled a MATLAB solution program based on the co-evolutionary genetic algorithm, and combined it with the rice harvesting operation in Wuchang City, Heilongjiang Province for example verification. Finally, the scheduling algorithm of this embodiment was compared with the standard genetic algorithm, proving the superiority of the improved algorithm of this embodiment.

[0307] The Wuchang rice project was the first batch of pilot projects approved by the Ministry of Agriculture in March 2017. In 2017, the rice planting area in Wuchang reached 2.355 million mu, accounting for 51% of the cultivated land in Wuchang and 4.12% of the rice planting area in the province, and the rice output continued to increase. With the rapid development of rice planting and related industries, the construction of agricultural infrastructure in Wuchang in 2018 is shown in Table 4. The source of Table 4: According to the data of the 2019 Wuchang Municipal Government Work Report.

[0308] Table 4

[0309]

[0310]

[0311] As of 2017, Wuchang City has actively promoted the construction of agricultural bases and built 15 "Internet + Agriculture" demonstration bases. In 2019, it plans to invest 80 million yuan to build 16,000 mu of green organic high-standard bases, 3,000 mu of fine seed breeding bases, 10,000 square meters of intelligent storage warehouses, and complete facilities such as rice fine seed research and development centers and social service centers.

[0312] Three agricultural machinery cooperatives in Dujia Town, Wuchang City were selected as research objects, and the actual situation of the survey and the relevant experimental data collected were used as examples for analysis. The relevant data of each agricultural machinery cooperative are shown in Table 5, and the cooperative numbers are M1 to M3. The number and model of rice harvesters owned by each cooperative are shown in Table 6. There are 10 rice harvesters of all models. The performance parameters and use cost parameters of harvesters of different models are shown in Table 7. During the model simulation process, the working hours of the harvester are from 6 am to 5 pm, with a one-hour rest interval in the middle, and the total working hours are 10 hours / day. The driving speed during the transfer of agricultural machinery is 35 km / h, and the transfer cost per unit distance is 2 yuan / km.

[0313] Table 5

[0314]

[0315] Table 6

[0316]

[0317] Table 7

[0318]

[0319] The entire scheduling problem will provide operation services for 50 operation points around the agricultural machinery cooperative, of which the initial number of operation points is 36, and their coordinates and demand information are shown in Table 8.

[0320] Table 8

[0321]

[0322] The scheduling model obtained by bringing the relevant data and some variable values ​​of the experimental process into the first-stage model is:

[0323]

[0324]

[0325] Experimental results and analysis

[0326] During the rice harvest season, the three agricultural machinery cooperatives collected the initial job orders received. In the initial stage of the algorithm, the crossover probability of the elite population was set to 0.9, the mutation probability was set to 0.05, and the crossover probability of the common population was set to 0.9, and the mutation probability was set to 0.1. In addition, after the 50th generation of evolution, the mutation probability of the common population increased by 0.001 per generation. The algorithm-related parameter settings are shown in Table 9.

[0327] Table 9

[0328]

[0329] Based on the above-established general model of resource-sharing agricultural machinery scheduling, the operating power of agricultural machinery model H3 is used as the threshold for operating point division in the heuristic decoding process. The designed multi-population co-evolutionary algorithm is used to solve the trend of the total cost change of agricultural machinery scheduling in the first stage through MATLAB2016a software. Figure 5 shown.

[0330] By solving the initial scheduling plan, the scheduling distance is 98.7 kilometers, the transfer cost is 197.4 yuan, and the operation cost is 9012.62 yuan. From the cost trend chart, it can be seen that the objective function has stabilized after 140 generations of evolution.

[0331] The operation point path of the initial stage scheduling plan is as follows: Figure 6 As shown in Table 10, in the initial scheduling plan, 18 operation points were assigned to the M1 agricultural machinery cooperative, 12 operation points were assigned to the M2 agricultural machinery cooperative, and 6 operation points were assigned to the M3 agricultural machinery cooperative. The walking path and operation access sequence of each agricultural machinery are shown in Table 10.

[0332] Table 10

[0333]

[0334] Summarize the 14 new orders that appear in the five time periods and plan a deployment plan that can meet the time window requirements and the constraints of agricultural machinery operation capacity. Figure 7 shown.

[0335] During the operation of agricultural machinery, agricultural machinery cooperative M1 received 6 new orders, agricultural machinery cooperative M2 received 4 new orders, and agricultural machinery cooperative M3 received 4 new orders. The walking paths of each agricultural machinery after receiving the new order are shown in Table 11. The scheduling scheme obtained in the dynamic scheduling stage obtained an operating area of ​​486.9 acres and an operating cost of 12,462.1 yuan. The amount of work allocated to each agricultural machinery is basically proportional to the model of the agricultural machinery, which shows that the scheduling method of the present invention can achieve the effect of uniform distribution of work.

[0336] Table 11

[0337]

[0338] The applicability of the present invention and the feasibility of the algorithm are verified through the above simulation.

[0339] The technical solution of the present invention is not limited to the above-mentioned specific embodiments. All technical variations made according to the technical solution of the present invention fall within the protection scope of the present invention.

Claims

1. An agricultural machinery scheduling method based on order resource sharing and agricultural machinery resource sharing, characterized in that: The following steps are involved: Step 1: Establish a mixed integer linear programming model for agricultural machinery scheduling that maximizes the workload under the premise of minimizing the scheduling cost; Step 2: Plan the agricultural machinery path based on the received orders with the shortest agricultural machinery driving path as the objective function, and arrange the operation tasks of the agricultural machinery according to the time window in the order; Step 3: When the agricultural machinery leaves the agricultural machinery point to carry out farmland operations, the received new orders are taken as optional tasks, and the objective function of the current stage is to serve the most farmers, and the operation tasks of the agricultural machinery are arranged according to the maximum operation area; Step 4: Use a multi-population co-evolution genetic algorithm based on heuristic rules with dynamic order processing strategy, operation machine type selection strategy and emergency order insertion strategy to solve the mixed integer linear programming model of agricultural machinery scheduling, and allocate operation requirements according to the operation priority of agricultural machinery points; The mixed integer linear programming model for agricultural machinery scheduling in step 1 is defined as G = (V, A), Where: V is the set of all nodes, V = {v1,v2,…,v n ,v n+1 ,…v n+m }, including each agricultural machinery cooperative and each farmland operation point, A is the set of all paths, A={(v i ,v j )|v i ,v j ∈V,i≠j}; The specific steps of the dynamic order processing strategy are: Step 1: Calculate the transportation distance between the inserted node and each pair of nodes including the operation point and the agricultural machinery service point; Step 2: Select a pair of nodes with the shortest transmission distance; Step 3: Connect nodes from a start node to an end node by visiting them in ascending order of their earliest time; Step 4: Connect the next node to the previously connected end node to form a route until all nodes are assigned to different routes; Step 5: Arrange the route so that the working vehicle starts the access sequence of this period after finishing the access sequence of the previous period; Step 6: Assign vehicles to routes in sequence until the requirements and time windows of all orders are met; In step 4, the operation machine type selection strategy is to use an improved conservation algorithm to determine the agricultural machine driving route and agricultural machine model, specifically including: First, by planning the orders that have been received, a specific path has been determined for each farmland operation point to perform the operation task. These routes will be combined and re-optimized according to the constraints; at this time, a travel time can be calculated for each path relative to different types of agricultural machinery, and the minimum travel time is recorded as T d : S jw : The time saved by inserting the work point w into the path where the work point j is located; t jw : The travel time between work point w and work point j (assuming t jw =t wj ); t dw : The travel time between the agricultural machinery point d and the operation point w; t dj : The travel time between the agricultural machinery point d and the operation point j; Therefore, S jw =t dw +t dj -t jw ; S jw The routes are sorted in descending order of numerical value, with the routes with higher ranking having greater probability of being selected, and then the largest S jw The relevant paths begin to be optimized; If the new path satisfies the problem constraints and does not form a sub-route, and can satisfy the time window constraints as follows, the insertion of the new order will be accepted, otherwise the insertion of the new operation point will be rejected; Among them, T w represents the time when agricultural machine k visits work point w, represents the operation time of agricultural machine k at operation point w, t wj represents the transfer time of agricultural machine k between operation point w and operation point j, l j It represents the latest time allowed for the start of operation at operation point j; Compare the total workload of path r with the operating power of each type of agricultural machinery, and finally select the vehicle; according to the demand quantity, if more than one type of agricultural machinery can be selected, the most economical choice The calculation is as follows, and the minimum Assigned to route r; The power consumed by the optimized agricultural machinery to complete a unit of work; Q r : Path Q r The total amount of all work; P k : The amount of work per unit time of this type of agricultural machinery, that is, the power of the agricultural machinery; Therefore Agricultural machinery of different power can undertake different workloads, but the smaller the power consumed to complete the unit workload, the more matching and economical the scheduling scheme is. Therefore, different agricultural machinery-path matching combinations are calculated, and the path is taken to the minimum The more solutions you get, the better ones you get.

2. The agricultural machinery scheduling method based on order resource sharing and agricultural machinery resource sharing according to claim 1 is characterized in that: The objective function in step 2 is: Based on the actual agricultural machinery scheduling problem and the model assumptions, the constraints are: (1) The number of agricultural machinery owned by each agricultural machinery station is limited, and the number of agricultural machinery dispatched cannot exceed the total amount of agricultural machinery owned by the agricultural machinery station: (2) One agricultural machine can only provide services for one operation point, that is: (3) The needs of each operation point can be met and can only be served once by one agricultural machine. Multiple services from multiple agricultural machines are not accepted, that is: (4) After each agricultural machine is set off, it cannot travel between agricultural machine points. The route between agricultural machine points is invalid. It can only set off to the operation point, that is: (5) The agricultural machinery must leave the current operation point after completing the operation task at the current operation point, that is: (6) After completing the tasks at each operation point, agricultural machinery cannot be parked at will, or parked at other agricultural machinery points. It must return to the starting agricultural machinery point, that is: (7) The path planning of agricultural machinery should meet the time window requirements of the farmers to whom the initial order belongs, and should not be violated unless in an emergency, that is: In the formula is the time when agricultural machine k starts working at operation point i; (8) The time window from the current operation point to the next operation point of the agricultural machinery should meet the following conditions, namely: Where M is a large positive number; (9) The total operating time of each agricultural machine on a given day shall meet the following conditions: the sum of the driving time and operating time of the agricultural machine on that day shall not exceed the prescribed working time, namely: (10) The workload of agricultural machinery at the operating point satisfies the following relationship, namely:

3. The agricultural machinery scheduling method based on order resource sharing and agricultural machinery resource sharing according to claim 1 is characterized in that: The objective function in step 3 is: The path is divided into five time periods A→B→C→D→E according to the operable time. The operation points that have participated in the planning are divided into four insertable spaces. Constraints are set for these four parts and the entire path: (1) In the A→B section, the agricultural machine starts from the operation point A but cannot return to the operation point A, and the agricultural machine must end at the operation point B and cannot leave the operation point B, that is: In the A→B section, the time when the agricultural machinery arrives at the operation point B must be earlier than the upper limit of the time window of the operation point B, that is: (2) Similarly, the constraint expression in the B→C segment is: (3) The constraint expression in the C→D segment is: (4) The constraint expression in the D→E segment is: In the planning of the entire path, an inserted operation point can only be inserted into one time period to avoid repeated insertion; (5) Avoid inserting the same task repeatedly at different times: (6) Ensure that a job task is executed at most once in the entire path: (7)ρ i is the time to reach operation point i:

4. The agricultural machinery scheduling method based on order resource sharing and agricultural machinery resource sharing according to claim 1 is characterized in that: The specific steps of solving the mixed integer linear programming model for agricultural machinery scheduling in step 4 are: Step 1: Integer coding is used, 1 to n are farmland operation points, (n+1) to (n+m) are agricultural machinery points, and individuals that do not meet the time window constraints and operation time constraints in the initial population are eliminated according to the decoding rules. At this time, all individuals in the initial population are feasible solutions; Step 2: Calculate the fitness value of the individual. The fitness function uses the inverse of the objective function and arranges them in descending order of fitness value. Step 3: Divide the problem into an elite sub-population and two ordinary populations. Each sub-population evolves into a sub-population according to the set crossover probability and mutation probability. Step 4: Calculate the fitness value of the new generation of individuals to determine whether the evolution meets the termination condition; if so, terminate and output the optimal solution; otherwise, go to Step 4; Step 5: Based on the initial scheduling plan and the constraints of the agricultural machinery's operating capacity and order time window, the nearest neighbor search algorithm is used to plan the dynamic orders according to the principle of the closest distance between the new order and the current operating point of the agricultural machinery, and a dynamic scheduling plan is generated; Step 6: For the selection of agricultural machinery models, an improved saving algorithm is used to allocate appropriate models according to the principle of minimum agricultural machinery power consumption per unit operating area, and the time window constraint of the initial order cannot be violated. At this time, the agricultural machinery selects the path according to the feasible solution of the initial solution of the planning concentration; Step 7: When an agricultural machinery failure causes the operation to stop, the agricultural machinery with the shortest distance and the smallest fluctuation in the overall plan caused by the path transfer is urgently dispatched to the rescue until the faulty agricultural machinery can return to participate in the dispatch operation.