Integrated Scheduling Method for Fresh Agricultural Product Harvesting and Distribution Based on Co-Evolutionary Algorithm

Abstract

The present invention provides an integrated scheduling method for the picking and distribution of fresh agricultural products based on a co-evolutionary algorithm, which relates to the technical field of agricultural product supply chain management and optimization. The method includes setting constraint conditions for the integrated scheduling problem of fresh agricultural product picking and distribution; defining a fresh degree loss function of agricultural products and a work efficiency function of the picking team; defining two objective functions, including a minimum operating cost function and a maximum customer satisfaction function; establishing a bi-objective optimization model for the integrated scheduling of fresh agricultural product picking and distribution based on the constraint conditions and objective functions; solving the bi-objective optimization model to obtain an optimal solution set, and performing agricultural product scheduling according to the picking and distribution information in the optimal solution set. The present invention can not only significantly reduce the operating costs of farmers and increase the profit margin, but also ensure that consumers obtain high-quality and high-freshness agricultural products during the demand period, thereby greatly improving the service satisfaction.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural product supply chain management and optimization, and particularly to an integrated scheduling method for picking and distribution of fresh agricultural products based on a co-evolution algorithm. Background Art

[0002] In recent years, the continuous improvement of the national living standard has prompted the public to increasingly focus on diet health, and the demand for fresh organic agricultural products has increased sharply. Against this background, "Community Supported Agriculture" (CSA), as a new agricultural product supply chain model, has received extensive attention in society. This model aims to provide fresh and healthy fresh agricultural products and establish an operation mechanism of sharing risks and interests through the direct connection between urban community residents and local farms. Under the "Community Supported Agriculture" model, farms provide the whole-process service from the picking to the distribution of fresh agricultural products to meet the needs of community residents, ensuring the same-day picking, distribution, and delivery of agricultural products.

[0003] However, although the "Community Supported Agriculture" model has developed rapidly, it also faces some practical problems that need to be solved urgently. Among them, the high freshness requirement of agricultural products and the high cost of picking logistics are two major challenges. Farms are usually located in the suburbs of cities, with remote geographical locations, while customers are scattered throughout the city, resulting in low distribution efficiency and high distribution costs. In addition, the freshness of fresh agricultural products is a key factor in their value, and once picked, their freshness will gradually decrease over time.

[0004] Therefore, how to effectively integrate and schedule the picking and distribution processes of farms to provide consumers with fresh agricultural products at low cost has become the key to the survival and profitability of farms. However, most current studies only focus on a single stage of picking or distribution, and few studies pay attention to the integrated optimization of these two stages. Summary of the Invention

[0005] For this reason, the embodiments of the present invention provide an integrated scheduling method for picking and distribution of fresh agricultural products based on a co-evolution algorithm, which is used to solve the problem that most existing technologies only focus on the single-stage optimization of picking or distribution of fresh agricultural products and lack an effective solution for the integrated optimization of the two stages of picking and distribution.

[0006] To solve the above problems, the embodiments of the present invention provide an integrated scheduling method for picking and distribution of fresh agricultural products based on a co-evolution algorithm, and the method includes:

[0007] For the integrated scheduling problem of picking and distribution of fresh agricultural products, set constraint conditions, including that farmers receive customer orders, arrange the picking team to pick agricultural products in the optimized picking order, and arrange vehicles to depart from the farm along the optimized route to deliver agricultural products to customers and finally return to the farm;

[0008] Define the fresh - degree loss function of agricultural products and the working - efficiency function of the picking team;

[0009] Define two objective functions, including the function of minimizing the operation cost and the function of maximizing the customer satisfaction. The operation cost includes the picking cost, transportation cost, and vehicle - usage cost. The customer satisfaction is related to the freshness of the delivered agricultural products and the delivery time;

[0010] Based on the above - mentioned constraints and objective functions, establish a bi - objective optimization model for the integrated scheduling of fresh agricultural product picking and distribution;

[0011] Solve the bi - objective optimization model to obtain the optimal solution set, and conduct agricultural product scheduling according to the picking and distribution information in the optimal solution set.

[0012] Preferably, the constraints also include:

[0013] Each type of agricultural product can only be picked by one picking team;

[0014] Each picking team can only pick one type of agricultural product at the same time;

[0015] The working efficiency of each picking team is different and will become fatigued as the picking time prolongs;

[0016] Each type of agricultural product has a specific fresh - degree loss rate, and the freshness of the agricultural product decreases over time after being picked;

[0017] Each customer may need more than one type of agricultural product;

[0018] Each customer has a scheduled delivery time. If the vehicle arrival time is later than the delivery time, a delay will occur and the customer satisfaction will decrease;

[0019] Each customer is only served by one vehicle;

[0020] The departure time of the vehicle from the farm is equal to the completion time of picking the last type of agricultural product to be loaded;

[0021] The vehicle's carrying capacity is limited, that is, the total load of the vehicle at any point on the route is not greater than the vehicle's carrying capacity.

[0022] Preferably, the fresh - degree loss function of agricultural products is used to describe the change of freshness over time during the process from picking the agricultural products to delivering them to customers, and is expressed as:

[0023]

[0024] In the formula, F represents the initial freshness of the agricultural product, that is, the maximum freshness; θ j is the fresh - degree loss rate of agricultural product j; F jRepresents the freshness of agricultural product j after time t.

[0025] Preferably, the picking team work efficiency function is used to reflect the fatigue effect of the picking team as the working time increases, and is expressed as:

[0026] t′ jl =t jl +λ jl ·S j ;

[0027] In the formula, t′ jl represents the actual picking time of picking team l for agricultural product j; t jl represents the picking time of picking team l for agricultural product j; λ jl represents the fatigue rate of picking team l for agricultural product j; S j represents the start time of picking agricultural product j.

[0028] Preferably, the minimized operating cost function is:

[0029] minf 1 =b·∑ j∈P\{0} ∑ l∈M Y jl +e·∑ m∈Q ∑ n∈Q ∑ h∈K Z mnk ·v mn +r·∑ m∈J ∑ h∈K Z 0mh ;

[0030] In the formula, f 1 represents the operating cost, b represents the picking cost per unit time; e represents the driving cost per unit time; r represents the fixed usage cost per vehicle; Y jl represents the total picking time of picking team l for agricultural product j; Z mnk represents that if vehicle h visits node m and then visits node n, then Z mnh =1; otherwise, Z mnh =0; v mn represents the driving time from node m to node n; Z 0mn represents that if vehicle h departs from the farm to customer m, then Z 0mh =1; otherwise, Z 0mh= 0; m, n ∈ Q, representing node indices, where Q = {0} ∪ J, representing the set of all nodes, 0 representing the farm, J = {1, 2, …, q}, representing the set of customers, q representing the total number of customers; h ∈ K, representing vehicle indices, where K = {1, 2, …, k}, representing the set of vehicles, k representing the number of vehicles; l ∈ M, representing picking team indices, where M = {1, 2, …, g}, representing the set of picking teams, g representing the number of picking teams; j ∈ P, representing agricultural product indices, where P = {0, 1, 2, …, p}, representing the set of agricultural products, where 0 represents a virtual agricultural product, p representing the number of types of agricultural products; P\{0} represents the set of agricultural products excluding 0.

[0031] Preferably, the maximizing customer satisfaction function is expressed by minimizing freshness loss and delivery delay:

[0032] min f 2 = ∑ m∈J ψ m + ∑ m∈J ξ m ;

[0033] In the formula, f 2 represents freshness loss and delivery delay; ψ m represents the maximum freshness loss in the agricultural products obtained by customer m; ξ m represents the delivery delay generated by visiting customer m; m ∈ J, J = {1, 2, …, q}, representing the set of customers, q representing the total number of customers.

[0034] Preferably, the bi-objective optimization model is expressed as:

[0035]

[0036]

[0037] ∑ n∈Q ∑ m∈J ∑ j∈P\{0} Z nmh · d mj ≤ δ, h ∈ K;

[0038]

[0039]

[0040] Where \(i, j\in P\), representing the agricultural product index, where \(P = \{0, 1, 2, \ldots, p\}\), representing the set of agricultural products, where \(0\) represents the virtual agricultural product and \(p\) represents the number of types of agricultural products; \(P\setminus\{0\}\) represents the set of agricultural products excluding \(0\); \(m, n\in Q\), representing the node index, where \(Q=\{0\}\cup J\), representing the set of all nodes, \(0\) represents the farm, \(J = \{1, 2, \ldots, q\}\), representing the set of customers, and \(q\) represents the total number of customers; \(h\in

[0041] K, representing the vehicle index, where \(K = \{1, 2, \ldots, k\}\), representing the set of vehicles, and \(k\) represents the number of vehicles; \(l\in M\), representing the picking team index, where \(M = \{1, 2, \ldots, g\}\), representing the set of picking teams, and \(g\) represents the number of picking teams; \(d mj The demand of customer \(m\) for agricultural product \(j\); \(R mj Represents that if customer \(m\) needs agricultural product \(j\), then \(R mj = 1; otherwise, \(R mj = 0; \(\theta j Represents the freshness loss rate of agricultural product \(j\); \(t lj Represents the picking time of picking team \(l\) for agricultural product \(j\); \(\lambda jl Represents the fatigue rate of picking team \(l\) for picking agricultural product \(j\); \(L m Represents the delivery time required by customer \(m\); \(v mn Represents the driving time from node \(m\) to node \(n\); \(v 0m Represents the driving time from node \(m\) to the farm; \(\delta\) represents the maximum load of each vehicle; \(B\) represents infinity; \(U jl Represents that if picking team \(l\) picks agricultural product \(j\), then \(U jl = 1; otherwise, \(U jl = 0; \(X ijl Represents that if picking team \(l\) picks agricultural product \(i\) and then picks agricultural product \(j\), then \(X ijl = 1; otherwise, \(X ijl = 0; \(W mh Represents that if vehicle \(h\) serves customer \(m\), then \(W mh = 1; otherwise, \(W mh = 0; \(Z mnh Represents that if vehicle \(h\) visits node \(m\) and then visits node \(n\), then \(Z mnh = 1; otherwise, \(Z mnh = 0; \(Z 0mh Represents that if vehicle \(h\) departs from the farm to customer \(m\), then \(Z 0mh = 1; otherwise, \(Z 0mh = 0; \(D h Represents the departure time of vehicle \(h\) from the farm; \(A m Represents the arrival time of the vehicle at customer \(m\); \(S j Represents the start time of picking agricultural product \(j\); \(Y jlrepresents the total picking time of agricultural product j by picking team l; C j represents the completion time of picking agricultural product j; represents the freshness loss incurred when agricultural product j is delivered to customer m; ψ m represents the maximum freshness loss of agricultural product j obtained by customer m; t′ jl represents the actual picking time of agricultural product j by picking team l; ξ m The delivery delay caused by visiting customer m.

[0042] Preferably, the method for solving the bi-objective optimization model to obtain the optimal solution set is as follows:

[0043] Use a co-evolution algorithm to solve the bi-objective optimization model to obtain the optimal solution set, where the co-evolution algorithm includes:

[0044] S51: Initialize the algorithm parameters, including the population size N, the local search probability α, the local search iteration times β, the number of sub-problems T in the decomposition method, the Q table, the initial state, and the maximum iteration times;

[0045] S52: Randomly generate the initial population, including solutions represented by N integer encodings;

[0046] S53: Evaluate the objective values of the N solutions in the population and store the non-dominated solution set in the external archive;

[0047] S54: Select the evolutionary framework of the population according to the current state;

[0048] S55: The population performs evolution and updates the external archive;

[0049] S56: The external archive performs evolution and updates the population;

[0050] S57: Calculate the reward value based on the change of state and update the Q table;

[0051] S58: If the maximum iteration number is reached, terminate; otherwise, execute S54;

[0052] S59: Finally, output the optimal solution set obtained from the external archive.

[0053] An embodiment of the present invention also provides an integrated scheduling system for picking and distribution of fresh agricultural products based on a co-evolution algorithm. This system is used to implement the integrated scheduling method for picking and distribution of fresh agricultural products based on the co-evolution algorithm as described above, and specifically includes:

[0054] Constraint setting module, which is used to set constraints for the integrated scheduling problem of fresh agricultural product picking and distribution, including that farmers receive customer orders, arrange picking teams to pick agricultural products according to the optimized picking sequence, and arrange vehicles to depart from the farm according to the optimized route to deliver agricultural products to customers and finally return to the farm;

[0055] Function definition module, which is used to define the fresh - degree loss function of agricultural products and the working - efficiency function of picking teams;

[0056] Objective - function definition module, which is used to define two objective functions, including the minimum - operation - cost function and the maximum - customer - satisfaction function, where the operation cost includes picking cost, transportation cost and vehicle - usage cost, and the customer satisfaction is related to the freshness of the delivered agricultural products and the delivery time;

[0057] Dual - objective optimization - model establishment module, which is used to establish a dual - objective optimization model for the integrated scheduling of fresh agricultural product picking and distribution based on the above - mentioned constraints and objective functions;

[0058] Solution and scheduling module, which is used to solve the dual - objective optimization model to obtain the optimal solution set and conduct agricultural product scheduling according to the picking and distribution information in the optimal solution set.

[0059] An embodiment of the present invention also provides a computer storage medium, which stores computer software for agricultural products. The computer software for agricultural products includes several instructions for causing a computer device to execute the above - mentioned integrated scheduling method for fresh agricultural product picking and distribution based on the co - evolutionary algorithm.

[0060] It can be seen from the above technical solutions that the present invention application has the following beneficial effects:

[0061] (1) Compared with previous studies that only focused on a certain link in the supply of fresh agricultural products, the present invention innovatively integrates the two stages of picking and distribution. By implementing an integrated scheduling strategy, it effectively reduces the intermediate links and realizes the direct connection between consumers and farmers. This not only enables consumers to obtain more affordable and high - quality fresh agricultural products such as vegetables and fruits, but also actively responds to the policies advocated by the state, such as community - supported agriculture and direct procurement and direct supply, which has important practical significance.

[0062] (2) For the integrated scheduling problem of fresh agricultural product picking and distribution, the present invention constructs a bi-objective optimization model. This model not only comprehensively considers the operating costs incurred during the supply service process, but also deeply analyzes the customer satisfaction with product freshness and delivery time. This model can not only help farmers reduce operating costs and increase profit margins, but also ensure that consumers can obtain high-quality and highly fresh agricultural products during the demand period. Although these two objectives conflict to a certain extent, a balance between them can be achieved by seeking the Pareto equilibrium solution.

[0063] (3) To solve the bi-objective optimization model, the present invention proposes a co-evolutionary algorithm and designs corresponding coding methods, crossover methods, and local search strategies. At the same time, combined with reinforcement learning technology, the effectiveness of the algorithm for solving problems is further improved. On a series of test cases, this algorithm is compared with three classic multi-objective algorithms in a comparative experiment, fully verifying the effectiveness and superiority of the model and the algorithm. The finally obtained Pareto equilibrium solution performs excellently in reducing costs, reducing the loss of agricultural product freshness, and delivery delays, thus significantly improving service satisfaction. The present invention provides a useful reference for farmers and fresh food e-commerce. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly describe the drawings required in the embodiments. By referring to the drawings, the features and advantages of the present invention will be more clearly understood. The drawings are schematic and should not be construed as limiting the present invention in any way. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts. Among them:

[0065] Figure 1 It is a flowchart of an integrated scheduling method for fresh agricultural product picking and distribution based on a co-evolutionary algorithm provided in the embodiment;

[0066] Figure 2 It is a schematic diagram of the integrated scheduling problem of fresh agricultural product picking and distribution in two stages in the embodiment;

[0067] Figure 3 It is a flowchart of the co-evolutionary algorithm in the embodiment;

[0068] Figure 4 It is a schematic diagram of the coding and decoding of the solution in the embodiment;

[0069] Figure 5 It is a schematic diagram of the crossover method in the embodiment;

[0070] Figure 6Box plots of the IGD metrics obtained by the four algorithms in the embodiments for different examples, where (a), (b), (c), (d), (e), (f), (g), (h), and (i) are the box plots of the IGD metrics obtained by the four algorithms for examples of 4 types of agricultural products, 6 types of agricultural products, 8 types of agricultural products, 10 types of agricultural products, 25 customers, 50 customers, 2 picking teams, 3 picking teams, and 4 picking teams, respectively;

[0071] Figure 7 Box plots of the Hypervolume metrics obtained by the four algorithms in the embodiments for different examples, where (a), (b), (c), (d), (e), (f), (g), (h), and (i) are the box plots of the Hypervolume metrics obtained by the four algorithms for examples of 4 types of agricultural products, 6 types of agricultural products, 8 types of agricultural products, 10 types of agricultural products, 25 customers, 50 customers, 2 picking teams, 3 picking teams, and 4 picking teams, respectively;

[0072] Figure 8 Block diagram of an integrated scheduling system for fresh agricultural product picking and distribution based on a co-evolution algorithm provided in the embodiments. Detailed implementation manners

[0073] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0074] Embodiment 1

[0075] To solve the problem that most of the existing technologies only focus on the single-stage optimization of fresh agricultural product picking or distribution and lack an effective solution for the integrated optimization of the two stages of picking and distribution. As Figure 1 shown, an integrated scheduling method for fresh agricultural product picking and distribution based on a co-evolution algorithm is proposed in the embodiments of the present invention, and the method includes:

[0076] S1: For the integrated scheduling problem of fresh agricultural product picking and distribution, set constraint conditions, including that farmers receive customer orders, arrange picking teams to pick agricultural products in an optimized picking order, and arrange vehicles to depart from the farm along an optimized route to deliver the agricultural products to customers and finally return to the farm;

[0077] S2: Define an agricultural product freshness loss function and a picking team work efficiency function;

[0078] S3: Define two objective functions, including the operation cost minimization function and the customer satisfaction maximization function. The operation cost includes the picking cost, transportation cost, and vehicle usage cost. The customer satisfaction is related to the freshness of the delivered agricultural products and the delivery time.

[0079] S4: Based on the above constraints and objective functions, establish a bi-objective optimization model for the integrated scheduling of fresh agricultural product picking and distribution.

[0080] S5: Solve the bi-objective optimization model to obtain the optimal solution set, and perform agricultural product scheduling according to the picking and distribution information in the optimal solution set.

[0081] As can be seen from the above technical solutions, an integrated scheduling method for fresh agricultural product picking and distribution based on the co-evolution algorithm of the present invention realizes the reduction of the intermediate link in the fresh agricultural product supply chain and the direct connection between consumers and farmers by integrating the two stages of picking and distribution. The core of its technology lies in constructing a bi-objective optimization model, which not only considers the minimization of operation costs but also takes into account the maximization of customer satisfaction, especially for the sensitivity of the freshness of agricultural products and the delivery time. To solve this complex model, the present invention designs a co-evolution algorithm, combined with reinforcement learning technology, effectively improving the solving efficiency. The advantage of this technical solution is that it can not only significantly reduce the operation costs of farmers and increase the profit margin but also ensure that consumers obtain high-quality and highly fresh agricultural products during the demand period, thus greatly improving the service satisfaction. In addition, this method actively responds to the policies advocated by the state, such as community-supported agriculture and direct procurement and direct supply, and has important practical significance and application value, providing useful reference and inspiration for farmers and fresh e-commerce.

[0082] The integrated scheduling problem of the two stages of fresh agricultural product picking and distribution studied by the present invention is as Figure 2 shown, including 4 decisions: (1) The task assignment of the picking team, that is, allocating the picked agricultural products to each picking team; (2) The picking order of agricultural products for each picking team; (3) The allocation of transportation vehicles for customer points, that is, allocating customer points to each transportation vehicle; (4) The delivery routes of each vehicle.

[0083] In step S1, for the integrated scheduling problem of fresh agricultural product picking and distribution, constraint conditions are set, including that farmers (farmers) receive customer orders, arrange picking teams to pick agricultural products (various fruits and vegetables) in the optimized picking order, and arrange vehicles to depart from the farm according to the optimized route to deliver the agricultural products to customers and finally return to the farm. The constraint conditions also include that each type of agricultural product can only be picked by one picking team; each picking team can only pick one type of agricultural product at the same time; the working efficiency of each picking team is different and will cause fatigue as the picking time extends; each type of agricultural product has a specific freshness loss rate, and the freshness of the agricultural product decreases over time after picking; each customer may need more than one type of agricultural product; each customer has a predetermined delivery time, and if the vehicle arrival time is later than the delivery time, a delay will occur and the customer satisfaction will decrease; each customer is only delivered by one vehicle; the departure time of the vehicle from the farm is equal to the picking completion time of the last type of agricultural product to be loaded; the vehicle carrying capacity is limited, that is, the total load of the vehicle at any point on the route is not greater than the vehicle carrying capacity.

[0084] Define model symbols, including sets, indices, parameters, variables and other symbols to be used, as shown in Table 1 below.

[0085] Table 1

[0086]

[0087]

[0088] In step S2, define the freshness loss function of agricultural products and the working efficiency function of picking teams. Specifically, the freshness loss function of agricultural products is used to describe the change of freshness over time during the process from picking agricultural products to delivering them to customers. The freshness of agricultural products is one of the important factors affecting customer satisfaction. Therefore, this function is crucial for evaluating customer satisfaction. Generally speaking, the freshness loss function of agricultural products can be expressed as a function of time. As time goes by, the freshness of agricultural products gradually decreases. Specifically, the freshness loss function of agricultural products is expressed as:

[0089]

[0090] In the formula, F represents the initial freshness of the agricultural product, that is, the maximum freshness; F j represents the freshness of agricultural product j after time t.

[0091] The picking team work efficiency function is used to reflect the fatigue effect of the picking team as the working time increases. The work efficiency of the picking team directly affects the picking cost and time, and thus affects the operating cost and customer satisfaction. The work efficiency function can be expressed as a function of working time. As the working time increases, the work efficiency of the picking team gradually decreases, that is, the fatigue effect gradually appears. Specifically, the picking team work efficiency function is expressed as:

[0092] t′ jl =t jl +λ jl ·S j 。

[0093] In step S3, two objective functions are defined, including minimizing the operating cost function and maximizing the customer satisfaction function. Among them, the operating cost is an important objective of model optimization, including picking cost, transportation cost and vehicle usage cost. The picking cost is affected by the work efficiency of the picking team, while the transportation cost and vehicle usage cost are related to the distribution route and vehicle arrangement. Therefore, by optimizing the work efficiency of the picking team and the distribution route, the operating cost can be reduced. Specifically, the minimizing operating cost function can be expressed as:

[0094] minf 1 =b·∑ j∈P\{0} ∑ l∈M Y jl +e·∑ m∈Q ∑ n∈Q ∑ h∈K Z mnk ·v mn +r·∑ m∈J ∑ h∈K Z 0mh ;

[0095] In the formula, f 1 represents the operating cost.

[0096] Customer satisfaction is another important optimization objective, which is related to the freshness of the delivered agricultural products and the delivery time. The higher the freshness of the agricultural products and the more punctual the delivery time, the higher the customer satisfaction. The freshness of agricultural products is affected by the freshness loss function of agricultural products, and the delivery time is affected by the overall scheduling arrangement of the picking and distribution processes. Specifically, the maximizing customer satisfaction function is expressed by minimizing the freshness loss and delivery delay:

[0097] minf 2 =∑ m∈J ψ m +∑ m∈J ξ m ;

[0098] In the formula, f 2Represent freshness loss and delivery delay.

[0099] In step S4, based on the above constraints and objective function, a bi-objective optimization model for the integrated scheduling of fresh agricultural product picking and distribution is established, where the bi-objective optimization model is expressed as:

[0100]

[0101]

[0102] ∑ n∈Q ∑ m∈J ∑ j∈P\{0} Z nmh ·d mj ≤δ, h ∈ K;

[0103]

[0104]

[0105] In step S5, the above bi-objective optimization model is solved to obtain the optimal solution set (Pareto equilibrium solution), and agricultural product scheduling is performed according to the picking and distribution information in the optimal solution set, including determining task allocation, picking order, vehicle allocation, and delivery route, etc.

[0106] Specifically, in this embodiment, a co-evolution algorithm is used to solve the bi-objective optimization model to obtain the optimal solution set, where the process of the co-evolution algorithm is as Figure 3 shown, specifically including:

[0107] S51: Initialize the algorithm parameters, including population size N, local search probability α, local search iteration times β, the number of sub-problems T in the decomposition method, Q table, initial state, and maximum iteration times;

[0108] S52: Randomly generate the initial population, including solutions represented by N integer encodings;

[0109] S53: Evaluate the objective values of the N solutions in the population and store the non-dominated solution set in the external archive;

[0110] S54: Select the evolutionary framework of the population according to the current state, including the domination-based framework and the decomposition-based framework;

[0111] S55: The population performs evolution and updates the external archive;

[0112] S56: The external archive performs evolution and updates the population;

[0113] S57: Calculate the reward value based on the change of state and update the Q table;

[0114] S58: If the maximum number of iterations is reached, terminate; otherwise, execute S54.

[0115] S59: Finally, output the optimal solution set obtained from the external file.

[0116] The present invention uses integer coding to represent the solution to the problem, including three integer strings. Among them, integer string 1 represents the number of types of agricultural products picked by each picking team; integer string 2 represents the picking order of agricultural products; integer string 3 represents the delivery order of customers. Two decisions in the picking stage are decoded through integer string 1 and integer string 2. As Figure 4 shown, picking team 1 is responsible for picking two types of agricultural products, picking agricultural product 2 and then agricultural product 1, and picking team 2 is responsible for picking two types of agricultural products, picking agricultural product 4 and then agricultural product 3. According to the customer demand, the loading limit of the vehicle, and integer string 3, two decisions in the delivery stage are decoded. As Figure 4 shown, the customer demand has been given. Assuming that the maximum loading limit of the vehicle is 6, then according to integer string 3, the customers are assigned to the same vehicle in order until the loading capacity of a vehicle reaches the upper limit, and then a new vehicle is rearranged. Finally, 4 delivery routes are obtained.

[0117] In addition, as Figure 5 shown, in the process of using the co-evolution algorithm to solve the optimal solution set, integer string 1 adopts the uniform crossover method, and both integer string 2 and integer string 3 adopt the order crossover method. The uniform crossover is designed as follows: randomly generate the number 0 or 1. If the number 1 is generated, copy the gene from parent 1; if the number 0 is generated, copy the gene from parent 2. Then perform the repair operation, that is, calculate whether the sum of the types of agricultural products responsible for picking by each picking team is equal to the sum of the types of products on the farm. If it is greater, randomly select a picking team and subtract 1 from the number of types of agricultural products it picks, and continue to calculate and check until the sum is just equal to the sum of the types of products on the farm; if it is less, randomly select a picking team and add 1 to the number of types of agricultural products it picks, and continue to calculate and check until the sum is exactly equal to the sum of the types of products on the farm. As Figure 5 shown, copy the first gene (2) of parent individual 1 and the second gene (3) of parent individual 2 to the offspring individual, and then calculate and find that 2 + 3 > 4. Randomly select the first picking team and subtract 1 from its picking type number. Therefore, the final integer string 1 of the offspring is (1, 3). The order crossover is designed as follows: randomly select two different cutting points, and select the gene segment between the cutting points as the crossover segment; copy the crossover segment in parent 1 to the corresponding gene position of the offspring individual, delete the genes of the crossover segment in parent 2 in parent 2, and assign the remaining genes to the offspring individual in turn.

[0118] To verify the effectiveness and superiority of the co-evolutionary algorithm (denoted as Q-CEA-K) proposed in the present invention, three algorithms are selected as comparison algorithms, namely the Non-dominated Sorting Genetic Algorithm II (NSGA-II), the Multi-objective Evolutionary Algorithm Based on Decomposition (MOEA / D), and the Multi-objective Artificial Bee Colony Algorithm (MOABC), which are run 20 times on different test cases. The IGD metric and the Hypervolume metric are used to measure the performance of the algorithms. It should be noted that the smaller the value of the IGD metric, the better the performance of the algorithm, while the larger the value of the Hypervolume metric, the better the performance of the algorithm. Figure 6 and Figure 7 The box plots obtained on the two metrics are given. It can be seen that the performance of the co-evolutionary algorithm is better than that of the other three algorithms.

[0119] Embodiment 2

[0120] As Figure 8 shown, the present invention provides an integrated scheduling system for fresh agricultural product picking and distribution based on a co-evolutionary algorithm, which is used to implement the integrated scheduling method for fresh agricultural product picking and distribution based on the co-evolutionary algorithm in the above Embodiment 1, and specifically includes:

[0121] The constraint condition setting module 100 is used to set constraint conditions for the integrated scheduling problem of fresh agricultural product picking and distribution, including that farmers receive customer orders, arrange the picking team to pick agricultural products in the optimized picking order, and arrange vehicles to depart from the farm according to the optimized route to deliver the agricultural products to customers and finally return to the farm;

[0122] The function definition module 200 is used to define the fresh agricultural product freshness loss function and the picking team work efficiency function;

[0123] The objective function definition module 300 is used to define two objective functions, including the minimum operating cost function and the maximum customer satisfaction function, where the operating cost includes the picking cost, the transportation cost, and the vehicle usage cost, and the customer satisfaction is related to the freshness of the delivered agricultural products and the delivery time;

[0124] The double-objective optimization model establishment module 400 is used to establish a double-objective optimization model for the integrated scheduling of fresh agricultural product picking and distribution based on the constraint conditions and the objective functions;

[0125] The solution and scheduling module 500 is used to solve the double-objective optimization model, obtain the optimal solution set, and perform agricultural product scheduling according to the picking and distribution information in the optimal solution set.

[0126] An integrated scheduling system for picking and delivering fresh agricultural products based on a co-evolutionary algorithm in this embodiment is used to implement the aforementioned integrated scheduling method for picking and delivering fresh agricultural products based on a co-evolutionary algorithm. Therefore, the specific implementation manners in the integrated scheduling system for picking and delivering fresh agricultural products based on a co-evolutionary algorithm can be seen in the embodiment part of the aforementioned integrated scheduling method for picking and delivering fresh agricultural products based on a co-evolutionary algorithm. For example, the constraint condition setting module 100, the function definition module 200, the objective function definition module 300, the dual-objective optimization model establishment module 400, and the solution and scheduling module 500 are respectively used to implement steps S1, S2, S3, S4, and S5 in the aforementioned integrated scheduling method for picking and delivering fresh agricultural products based on a co-evolutionary algorithm. Therefore, the specific implementation manners can refer to the descriptions of the corresponding various part embodiments. To avoid redundancy, they will not be elaborated here.

[0127] Embodiment III

[0128] An embodiment of the present invention provides a computer storage medium. The computer storage medium stores computer software for agricultural products. The computer software for agricultural products includes several instructions for causing a computer device to execute the aforementioned integrated scheduling method for picking and delivering fresh agricultural products based on a co-evolutionary algorithm.

[0129] Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, or computer programs for agricultural products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program for agricultural products implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0130] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer programs for agricultural products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the flows and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the specified functions in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0131] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to operate in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means that implement the functions specified in one or more of the processes and / or blocks Figure 1 one or more of the processes and / or blocks Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one or more of the processes and / or blocks Figure 1 one or more of the processes and / or blocks Figure 1 one or more of the blocks.

[0132] Obviously, the above embodiments are only examples for clear illustration and are not limitations on the implementation manners. For those of ordinary skill in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to enumerate all the implementation manners here. And the obvious changes or modifications derived therefrom are still within the protection scope of the present invention.

Claims

1. A method for integrated dispatching of fresh agricultural products picking and distribution based on a co-evolutionary algorithm, characterized in that: include: For the integrated scheduling problem of picking and delivering fresh agricultural products, constraints are set, including that farmers receive customer orders, arrange picking teams to pick agricultural products in an optimized picking order, and arrange vehicles to deliver agricultural products to customers from the farm along the optimized route, and finally return to the farm. The constraints also include: each kind of agricultural product can only be picked by one picking team; each picking team can only pick one kind of agricultural product at the same time; each picking team has different work efficiency, and fatigue will occur as the picking time increases; each agricultural product has a specific freshness loss rate, and the freshness of agricultural products decreases over time after picking; each customer may need more than one kind of agricultural product; each customer has a scheduled delivery time, and if the vehicle arrives later than the delivery time, a delay will occur, which will reduce customer satisfaction; each customer is delivered by only one vehicle; the time when the vehicle departs from the farm is equal to the time when the last agricultural product to be loaded is completed; the vehicle carrying capacity is limited, that is, the total vehicle load at any point on the route is not greater than the vehicle carrying capacity; The agricultural product freshness loss function and the picking team work efficiency function are defined. The agricultural product freshness loss function is used to describe the change of freshness over time in the process from picking to delivering agricultural products to customers, which is expressed as: Where F represents the initial freshness of agricultural products, that is, the maximum freshness; θ j is the freshness loss rate of agricultural product j; F j represents the freshness of agricultural product j after time t; The picking team work efficiency function is used to reflect the fatigue effect of the picking team as the working time increases, and is expressed as: t′ jk =t jl +λ jl ·S j ; Where t′ jl represents the actual picking time of picking team l picking agricultural product j; t jl represents the picking time of agricultural product j by picking team l; jl represents the fatigue rate of picking team l picking agricultural product j; S j represents the picking start time of agricultural product j; Two objective functions are defined, including minimizing the operating cost function and maximizing the customer satisfaction function, where the operating cost includes the picking cost, transportation cost and vehicle use cost, and customer satisfaction is related to the freshness and delivery time of the delivered agricultural products. The minimizing operating cost function is: minf1=b·∑ j∈P\{0} ∑ l∈M Y jl +e·∑ m∈Q ∑ n∈Q ∑ h∈ KZ mnk ·v mn +r·∑ m∈J ∑ h∈K Z 0mh ; In the formula, f1 represents the operating cost, b represents the picking cost per unit time; e represents the driving cost per unit time; r represents the fixed cost of each vehicle; Y jl represents the total picking time of picking team l for picking agricultural product j; Z mnk If vehicle h visits node m and then node n, then Z mnh =1; otherwise, Z mnh =0;v mn represents the travel time from node m to node n; Z 0mh If vehicle h departs from the farm and goes to customer m, then Z 0mh =1; otherwise, Z 0mh =0; m,n∈Q, represents the node index, where Q={0}∪J, represents the set of all nodes, 0 represents the farm, J={1,2,…,q}, represents the customer set, and q represents the total number of customers; h∈K, represents the vehicle index, where K={1,2,…,k}, represents the vehicle set, and k represents the number of vehicles; l∈M, represents the picking team index, where M={1,2,…,g}, represents the picking team set, and g represents the number of picking teams; j∈P, represents the agricultural product index, where P={0,1,2,…,p}, represents the agricultural product set, where 0 represents the virtual agricultural product, and p represents the number of agricultural product types; P\{0} represents the agricultural product set that does not contain 0; The maximized customer satisfaction function is expressed by minimizing freshness loss and delivery delay: minf2=∑ m∈J ψ m +∑ m∈J ψ m ; Where f2 represents freshness loss and delivery delay; ψ m Get the maximum freshness loss in agricultural products on behalf of customer m; ξ m represents the delivery delay caused by visiting customer m; m∈J, J={1,2,…,q}, represents the customer set, and q represents the total number of customers; Based on the constraints and the objective function, a dual-objective optimization model for integrated scheduling of fresh agricultural product picking and distribution is established; The co-evolutionary algorithm is used to solve the dual-objective optimization model to obtain the optimal solution set, and agricultural products are dispatched according to the picking and distribution information in the optimal solution set.

2. The integrated dispatching method for picking and distributing fresh agricultural products based on the collaborative evolution algorithm according to claim 1 is characterized in that: The dual-objective optimization model is expressed as: ∑ n∈Q ∑ m∈J ∑ j∈P\{0} Z nmh ·d mj ≤δ,h∈K; Where i,j∈P represents the agricultural product index, where P = {0,1,2,…,p}, represents the agricultural product set, where 0 represents the virtual agricultural product and p represents the number of agricultural product types; P\{0} represents the agricultural product set that does not contain 0; m,n∈Q represents the node index, where Q = {0}∪J, represents the set of all nodes, 0 represents the farm, J = {1,2,…,q}, represents the customer set, and q represents the total number of customers; h∈ K represents the vehicle index, where K = {1, 2, ..., k}, represents the vehicle set, and k represents the number of vehicles; l∈M represents the picking team index, where M = {1, 2, ..., g}, represents the picking team set, and g represents the number of picking teams; d mj The demand of customer m for agricultural product j; R mj If customer m needs agricultural product j, then R mj =1; otherwise , R mj =0;θ j represents the freshness loss rate of agricultural product j; t lj represents the picking time of agricultural product j by picking team l; jl represents the fatigue rate of picking team l picking agricultural product j; L m represents the delivery time required by customer m; v mn Represents the travel time from node m to node n; v 0m represents the travel time from node m to the farm; δ represents the maximum load of each vehicle; B stands for infinity; U jl If picking team l picks agricultural product j, then U jl =1; otherwise , U jl =0;X ijl If picking team l picks agricultural product i and then picks agricultural product j, then X ijl =1; otherwise, X ijl =0; W mh If vehicle h serves customer m, then W mh =1; otherwise, W mh =0; Z mnh If vehicle h visits node m and then node n, then Z mnh =1; otherwise, Z mnh =0; Z 0mh If vehicle h departs from the farm and goes to customer m, then Z 0mh =1; otherwise, Z 0mh =0;D h represents the time when vehicle h departs from the farm; A m represents the time when the vehicle arrives at customer m; S j represents the picking start time of agricultural product j; Y jl represents the total picking time of picking team l for picking agricultural product j; C j represents the harvesting completion time of agricultural product j; represents the freshness loss of agricultural product j when it is delivered to customer m; ψ m Get the maximum freshness loss of agricultural products on behalf of customer m; t′ jl represents the actual picking time of picking agricultural product j by picking team l; ξ m The delivery delay caused by visiting customer m.

3. The integrated dispatching method for picking and distributing fresh agricultural products based on the collaborative evolution algorithm according to claim 1 is characterized in that: The collaborative evolution algorithm comprises: S51: Initialize algorithm parameters, including population size N, local search probability α, local search iteration number β, number of subproblems in the decomposition method T, Q table, initial state, and maximum number of iterations; S52: randomly generate an initial population, including solutions represented by N integer codes; S53: Evaluate the target values ​​of N solutions in the population and store the non-dominated solution set in an external archive; S54: An evolutionary framework for selecting populations based on their current state; S55: The population performs evolution and updates the external archive; S56: external archive performs evolution and updates the population; S57: Calculate the reward value based on the change of state and update the Q table; S58: If the maximum number of iterations is reached, terminate, otherwise execute S54; S59: Finally output the optimal solution set obtained from the external archive.

4. A fresh agricultural product picking and distribution integrated scheduling system based on a collaborative evolutionary algorithm, characterized in that: The system is used to implement the integrated scheduling method for picking and distributing fresh agricultural products based on the collaborative evolution algorithm as described in any one of claims 1 to 3, specifically comprising: The constraint setting module is used to set constraints for the integrated scheduling problem of picking and delivering fresh agricultural products, including farmers receiving customer orders, arranging picking teams to pick agricultural products in an optimized picking order, and arranging vehicles to deliver agricultural products from farms to customers according to optimized routes and finally returning to farms; Function definition module, used to define the agricultural product freshness loss function and the picking team work efficiency function; An objective function definition module is used to define two objective functions, including minimizing the operating cost function and maximizing the customer satisfaction function, where the operating cost includes the picking cost, transportation cost and vehicle use cost, and the customer satisfaction is related to the freshness of the delivered agricultural products and the delivery time; A dual-objective optimization model establishment module, used to establish a dual-objective optimization model for integrated scheduling of fresh agricultural product picking and distribution based on the constraint conditions and the objective function; The solution and scheduling module is used to solve the dual-objective optimization model using a co-evolutionary algorithm to obtain an optimal solution set, and to schedule agricultural products according to the picking and delivery information in the optimal solution set.

5. A computer storage medium, characterized in that: The computer storage medium stores computer software agricultural products, and the computer software agricultural products include a number of instructions for enabling a computer device to execute the integrated scheduling method for picking and distributing fresh agricultural products based on a collaborative evolutionary algorithm as described in any one of claims 1 to 3.

Images