Crop planting strategy optimization method

Abstract

The invention discloses a crop planting strategy optimization method. The method comprises the following steps: 1, collecting data to obtain cultivated land types; 2, performing data preprocessing on the cultivated land types collected in the step 1, and performing secondary division according to planting seasons; establishing a crop label sequence for planting in each cultivated land; 3, according to the current unit area profit of the crops, a planting strategy model is established, the crops with the highest planting profit are selected to serve as the basis of state transition, and the target function is that the earnings of all cultivated lands are the maximum; 4, solving the planting strategy model established in the step 3; and 5, optimizing the model solved in the step 4, establishing a target planning model, introducing priori knowledge, and optimizing a result by using a heuristic algorithm to obtain a specific crop area of planting. According to the method, the country can better adapt to market changes in a complex agricultural environment, maximum utilization of agricultural resources is achieved, and then it is guaranteed that agricultural workers obtain more stable earnings.

Description

Technical Field

[0001] The invention belongs to the technical field of agricultural crop planting strategies, and specifically relates to a crop planting strategy optimization method. Background Art

[0002] The existing traditional rural small peasant economy model is becoming increasingly unsuitable for the current rural reality. It is of great significance to coordinate various resources based on the actual rural situation and formulate scientific planting strategies according to local conditions.

[0003] With the continuous improvement of technologies such as big data and artificial intelligence, technical support has been provided for the field of smart agriculture. These platforms can be used to count and analyze the current income and expenditure in the rural agricultural field. Based on this, appropriate mathematical models can be established to formulate scientific and reasonable crop planting strategies, so that rural arable land can be maximized and agricultural production can be made more intelligent and precise.

[0004] For example, the patent with publication number: CN112101868A discloses an integrated IoT cloud control system based on smart agriculture, including a coding system, an IoT detection system, a supply chain system, and a background management system. The system encodes agricultural product seeds one by one, and transmits the growth information of the encoded seeds and the logistics information of agricultural products to the background management system, so as to realize real-time monitoring and full tracking of agricultural products. This method ensures information transparency between consumers and producers and sellers, and reduces food safety risks. The defect is that it only considers the growth of seeds from the perspective of crops, but lacks the measurement of agricultural resources, market factors and various agricultural restrictions.

[0005] Another example is the patent with publication number: CN118095792A, which discloses a smart agricultural management system based on big data, including a data acquisition module, a data preprocessing module, a data analysis and decision-making module, and a public service module. The agricultural crop yield prediction model is constructed through the collected data to predict and analyze the agricultural crop yield. The agricultural crop yield obtained through the prediction and analysis is used to optimize the cultivated agricultural crop area, adjust the agricultural crop planting strategy, and achieve precise fertilization and irrigation. The defect is that it does not take into account the richness of actual cultivated land types in rural areas and the richness of crops that can be planted.

[0006] However, although existing technologies have achieved the intelligence and digitalization of agriculture to a certain extent, they are mostly based on the perspective of crop production, mainly focusing on the intelligent management of the same type of arable land in a small area, ignoring the diversity of rural arable land types and selectable crops.

[0007] In addition, current traditional agricultural planting strategies often rely on the experience and intuition of agricultural workers, and fail to effectively utilize big data analysis, especially the lack of overall understanding of agricultural resources, arable land restrictions and market factors, making it difficult to scientifically and effectively formulate the next stage of planting strategies based on the current agricultural conditions.

[0008] Therefore, in response to the above problems, an optimization method for crop planting strategies is needed. Summary of the invention

[0009] In order to overcome the above problems in the prior art, the purpose of the present invention is to provide a method for optimizing crop planting strategies, which comprehensively considers the types of rural cultivated land and the characteristics of crops that can be planted, and combines the current market economic status of crops and various cultivated land restrictions to formulate an economic maximization planting strategy optimization method and reasonably arrange the planting strategy for the next stage. It is conducive to better adapting to market changes in a complex agricultural environment, maximizing the use of agricultural resources, and ensuring that agricultural workers obtain more stable income.

[0010] In order to achieve the above object, the technical solution adopted by the present invention is:

[0011] A method for optimizing crop planting strategy comprises the following steps:

[0012] Step 1: Collect data and obtain the types of cultivated land;

[0013] Step 2: Pre-process the data of the cultivated land types collected in step 1 and divide them into two parts according to the planting seasons; establish a sequence of crop numbers that can be planted on each piece of cultivated land; and fully consider the diversity of crops;

[0014] Step 3: According to the current unit area profit of crops, a planting strategy model is established to select the crops with the highest planting profit as the basis for state transfer. The objective function is to maximize the benefits of all cultivated land.

[0015] According to the current agricultural market conditions, multi-year planting strategies are arranged, and the dynamic programming idea is particularly suitable for formulating multi-year crop planting strategies. Specifically, each year is regarded as a stage of dynamic programming, and the planting strategy of each stage represents the state of that stage. The decision is to determine which type of crops and the planting area of ​​each type of cultivated land. Considering the actual situation, the main constraint is that the total cultivated area of ​​each crop must be equal to the total cultivated area of ​​the village. The state transfer equation is based on the unit area profit of each plot at the current stage, and it is also necessary to consider the repeated cropping and the mutual influence between crops during crop planting (such as bean crops have a positive effect on the yield of crops in the next year). At each stage, according to the current unit area profit of the crop, the crop with the highest planting profit is selected as the basis for state transfer, and the objective function is to maximize the income of all cultivated land.

[0016] Step 4: Solve the planting strategy model established in step 3;

[0017] Step 5: Optimize the model solved in step 4, establish a target planning model, introduce prior knowledge, use heuristic algorithms to optimize the results, and obtain the specific crop area to be planted.

[0018] In the step 1, the cultivated land type includes agricultural resources, cultivation restrictions and market status data, which are used to understand the current agricultural economic situation of a certain rural area and facilitate scientific and reasonable arrangement of planting strategies;

[0019] The agricultural resource data include the type and area of ​​rural cultivated land in the region, the type of crops planted on each piece of cultivated land, the planting season and the yield;

[0020] Data on farmland restrictions include: farmland rotation, priority arrangements for specific crops, and farmland requirements for agricultural workers;

[0021] Market condition data include crop cultivation costs and average selling prices.

[0022] In the step 2, the secondarily divided cultivated land types are divided into flat dry land, terraced fields, irrigated land, ordinary greenhouses and smart greenhouses. Different types of cultivated land have different planting seasons. In addition, according to the types of crops and planting seasons collected in step 1;

[0023] Among them, flat dry land and terraced fields are single-season, irrigated land, ordinary greenhouses and smart greenhouses are double-season. For double-season cultivated land, it will be divided into two plots and arranged at the same time as the single-season cultivated land;

[0024] In addition, all crops that can be grown on each piece of arable land in each season are included in the arrangement and a crop number sequence is formulated.

[0025] The planting strategy model in step 3 is specifically as follows:

[0026] Step 1) Problem stage: The entire planting strategy problem is divided into multiple stages in chronological order. Each year is an independent decision-making stage. The problem to be solved in each stage is how to choose the crop with the highest profit at the current stage for each plot of land within the limited plot area M, so as to maximize the total income of the year.

[0027] Step 2) State definition: State T k It is expressed as the maximum profit of the planting strategy of the six plot types in the kth year, including the area of ​​crops planted in all plots, dp k(i,j) represents the maximum profit when the remaining crop area is j on plot type i in year k. The most profitable planting allocation is made for the same plot type in year k. At this time, the corresponding planting strategy on plot type i in year k-1 has been determined. According to the ineffectiveness of the state, the planting strategy of the next state, that is, the next year, is only restricted by the previous year.

[0028] Step 3) State transition equation: The state transition equation describes the process of how to transfer from one state to the next stage state. In each stage k, according to the unit area profit of the currently arable crops, the crop with the highest profit is selected.

[0029] Specifically:

[0030]

[0031] Among them, J' is the set of crops that do not include those planted in the previous year, c is the specific crop, j' is the planting area of ​​crop c, and lr is the profit corresponding to the area of ​​j' of crop c. Therefore, dp k (i, j) represents the maximum profit when the remaining planting area of ​​the i-th plot is j in the k-th year.

[0032] The step 4 breaks down each stage of the dynamic programming problem into a 0-1 knapsack problem, and its specific principle is:

[0033] The planting strategy at each stage needs to consider whether these crops should be planted at the current stage and how to arrange the planting to maximize the profit. It is necessary to arrange n kinds of crops to be planted on the total area of ​​M arable land in the village, among which the planting area of ​​the jth crop is s j , the profit per unit area of ​​this crop is v j , specifically:

[0034] v j =b j ×m j -c j

[0035] There are only two possibilities for whether to plant a certain crop: either plant it or not plant it; therefore, it can be transformed into a 0-1 backpack problem, that is, put some items into a backpack and find the solution that can put the most valuable items, where the capacity of the backpack is the total area of ​​arable land available M, and each crop is regarded as an item in the backpack, and its "weight" is the planting area s of the crop. j , the value is its unit area benefit v j , the goal is to maximize the total revenue.

[0036] Based on the above, the following questions are transformed:

[0037] Given M>0, s j >0,v j >0, where 1≤j≤n, solve for an n-dimensional 0-1 vector (x1,x2,…,x n ), when x j =1, it means planting the jth crop, otherwise it is not planted. The goal is to maximize profit, that is, to satisfy:

[0038]

[0039] Based on the non-aftereffect of dynamic programming, the initial planting strategy of the current stage is obtained by using the 0-1 knapsack problem and the greedy strategy in each stage, and the specific crop area planted can be obtained.

[0040] The optimization model of step 5 is as follows:

[0041]

[0042] Among them, α is the number of divided arable land blocks, β represents the total number of crop labels, and x ijk is a 0-1 variable, Z k is the total net income in the kth year, b ijk is the selling price, v ijk is the unit area profit of planting the jth crop on the i-th plot in the k-th year, s ijk is the actual yield of the jth crop planted on the ith plot of land in the kth year, M is the total area of ​​arable land, j is the crop number, and j' is the number of the legume crop.

[0043] Beneficial effects of the present invention:

[0044] 1. Through the data collection step, the present invention can obtain data such as agricultural resources and market conditions in a certain area, including the type and area of ​​arable land, sales price of crops, etc. It comprehensively considers the types of existing rural economic arable land and the diversity of selectable crops, effectively utilizes big data analysis, and facilitates a comprehensive and comprehensive grasp of the current agricultural development status of the region.

[0045] 2. The present invention can arrange the planting strategy for each year in the future by establishing a planting strategy model. Agricultural workers can use this model to arrange a crop planting plan that maximizes profits and achieves maximum utilization of agricultural resources.

[0046] 3. In the model solving step, the present invention refines the dynamic programming problem into a 0-1 knapsack problem at each stage, and solves it using a greedy strategy based on the lack of aftereffect of dynamic programming, ensuring that the planting strategy at each stage can meet various restrictions while maximizing the benefits. The obtained planting strategy is conducive to the village to better adapt to market changes in a complex agricultural environment, thereby maximizing the use of agricultural resources and achieving more robust benefits.

[0047] 4. The present invention performs a model optimization step after the model solving step, establishes a target planning model, introduces prior knowledge, and uses a heuristic algorithm to solve the problem, which can not only ensure that the optimal planting strategy is obtained, but also accelerate the convergence speed of the heuristic algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A schematic diagram of a process flow of a crop planting strategy optimization method provided by an example of the present invention.

[0049] Figure 2 Schematic diagram of the data collection steps provided for examples of the present invention.

[0050] Figure 3 A flowchart of solving the planting strategy for each stage based on the greedy strategy idea is provided for an example of the present invention. DETAILED DESCRIPTION

[0051] The present invention will be further described in detail below in conjunction with the accompanying drawings.

[0052] A method for optimizing crop planting strategy comprises the following steps:

[0053] Step 1: Data Collection

[0054] It includes collecting data on agricultural resources, farmland restrictions and market conditions. Among them, agricultural resource data includes: the type and area of ​​rural farmland in the region; the types of crops that can be planted on each piece of farmland, the planting season and yield, etc. Farmland restriction data includes: farmland rotation planting, priority arrangement of special types of crops, and farmland demand of agricultural workers, etc. Market condition data includes: crop planting costs, average sales prices, etc.

[0055] Step 2: Data preprocessing

[0056] Considering the actual situation, the types of cultivated land in rural areas may be flat dry land, terraced fields, irrigated land, ordinary roc and smart greenhouses, and the planting seasons of different types of cultivated land are different, so the cultivated land types are divided twice according to the planting seasons. In addition, a sequence of crop numbers that can be planted on each piece of cultivated land is established based on the secondarily divided cultivated land, taking into full account the diversity of crops.

[0057] Step 3: Establishing the planting strategy model

[0058] The present invention arranges multi-year planting strategies according to the existing agricultural market conditions, and the dynamic programming concept is particularly suitable for formulating multi-year crop planting strategies, specifically: each year is regarded as a stage of dynamic programming, and the planting strategy of each stage represents the state of that stage. The decision is to determine what type of crops to plant on each type of cultivated land and the planting area. Considering the actual situation, the main constraint is that the total cultivated area of ​​each crop must be equal to the total cultivated area of ​​the village. The state transfer equation is based on the unit area profit of each plot at the current stage, and it is also necessary to consider the mutual influence between repeated cropping and crops during crop planting (such as bean crops have a positive effect on the yield of crops in the next year). In other words, at each stage, according to the current unit area profit of the crop, the crop with the highest planting profit is selected as the basis for state transfer. The objective function is to maximize the profit of all cultivated land. Specifically:

[0059] Step 1) Problem stage: The entire planting strategy problem is divided into multiple stages in chronological order. Each year is an independent decision-making stage. The problem to be solved in each stage is how to select the most profitable crop for planting at the current stage in each plot of land within the limited plot area M, so as to maximize the total income of the year.

[0060] Step 2) State definition: State T k It is expressed as the maximum profit of the planting strategy of the six plot types in the kth year, including the area of ​​crops planted in all plots. k (i,j) represents the maximum profit when the remaining crop area on plot type i is j in the kth year. The planting allocation with the highest profit is made for the same plot type in the kth year. At this time, the corresponding planting strategy on plot type i in the k-1th year has been determined. According to the non-aftereffect of the state, the planting strategy of the next state, that is, the next year, is only restricted by the previous year.

[0061] Step 3) State transition equation: The state transition equation describes the process of how to transfer from one state to the next stage state. In each stage k, according to the unit area profit of the currently arable crops, the crop with the highest planting profit is selected. Specifically:

[0062]

[0063] Among them, J' is the set of crops that do not include those planted in the previous year, c is the specific crop, j' is the planting area of ​​crop c, and lr is the profit corresponding to the area of ​​j' of crop c. Therefore, dp k (i, j) represents the maximum profit when the remaining planting area of ​​the i-th plot is j in the k-th year.

[0064] Step 4: Model solution

[0065] First, each stage of step three is refined into a 0-1 knapsack problem.

[0066] The planting strategy at each stage needs to consider whether these crops should be planted at the current stage and how to arrange the planting to maximize the profit. It is necessary to arrange n kinds of crops to be planted on the total area of ​​M arable land in the village, among which the planting area of ​​the jth crop is s j , the profit per unit area of ​​this crop is v j , specifically:

[0067] v j =b j ×m j -c j

[0068] There are only two possibilities for whether to plant a certain crop: either plant it or not. In the current problem, there will be some situations that do not meet the planting area, so consider planting together. Therefore, it can be transformed into a 0-1 backpack problem, that is, put some items into a backpack and find the solution that can put the most valuable items. The capacity of the backpack is the total area of ​​available arable land M, and each crop is regarded as an item in the backpack, and its "weight" is the planting area s of the crop. j , the value is its unit area benefit v j The goal is to maximize the total revenue. Based on the above ideas, it is transformed into the following questions:

[0069] Given M>0, s j >0,v j >0, where 1≤j≤n, solve for an n-dimensional 0-1 vector (x1,x2,…,x n ), when x j =1, it means planting the jth crop, otherwise it is not planted. The goal is to maximize profit, that is, to satisfy:

[0070]

[0071] Based on the non-aftereffect of dynamic programming, the initial planting strategy of the current stage is obtained by using the 0-1 knapsack problem and the greedy strategy in each stage, and the specific crop area planted can be obtained.

[0072] Step 5: Model optimization

[0073] By establishing a target programming model, introducing prior knowledge, and optimizing the results using a heuristic algorithm, the specific crop area to be planted can be obtained. The specific optimization model is as follows:

[0074]

[0075] Among them, α is the number of divided arable land blocks, β represents the total number of crop labels, and x ijk is a 0-1 variable, Zk is the total net income in the kth year, b ijk is the selling price, v ijk is the unit area profit of planting the jth crop on the i-th plot in the k-th year, s ijk is the actual yield of the jth crop planted on the ith plot of land in the kth year, M is the total area of ​​arable land, j is the crop number, and j' is the number of the legume crop.

[0076] Example:

[0077] A method for optimizing crop planting strategies, as shown in the attached Figure 1 , including the following steps:

[0078] Step 1: Data Collection

[0079] As attached Figure 2 The data collection steps include collecting agricultural resources, farmland restrictions and market conditions data. Among them, agricultural resource data include: the type and area of ​​rural farmland in the region; the types of crops that can be planted on each piece of farmland, the planting season and yield, etc. The farmland restriction data include: farmland rotation planting, priority arrangement of special types of crops, and farmland demand of agricultural workers, etc. The market condition data include: crop planting cost, average selling price, etc.

[0080] Step 2: Data preprocessing

[0081] In the example of the present invention, the types of cultivated land in the village may be flat dry land, terraces, irrigated land, ordinary greenhouses and smart greenhouses. Due to the different planting seasons of cultivated land: some cultivated land can only be planted in a single season, while some cultivated land can be planted in two seasons. In order to unify this issue, the irrigated land, ordinary greenhouses and smart greenhouses that can be used to plant two crops are divided into D and D'; E and E'; F and F' according to the season. Arid land, terraces and hillsides are still A, B and C. In this way, when arranging the planting strategy for each year, it is only necessary to arrange single-season crops in each plot.

[0082] For irrigated land, if rice is planted, it will be arranged in the first quarter D, and D' must not be arranged with crops, ensuring that the irrigated land can be planted with rice in a single season. If rice is not planted, D' will only arrange three crops: Chinese cabbage, white radish and red radish.

[0083] Different types of farmland can grow different crops, so a crop number sequence is established:

[0084] j={j A ,j B ,...,j F1 ,j F2}={(1,2,...,15),(16,17,...,30),...,(90,91,...,107),(107,108,...,125)}

[0085] Among them, j is the crop number sequence that can be selected for each plot, including the crop numbers that can be selected for each terrain from A to F2.

[0086] Step 3: Establishing the planting strategy model

[0087] Each stage may be N (N≥1) years. The present invention arranges multi-year planting strategies according to the existing agricultural market conditions, and the dynamic programming concept is particularly suitable for formulating multi-year crop planting strategies. Specifically, each year is regarded as a stage of dynamic programming. The planting strategy of each stage represents the state of that stage. The decision is to determine which type of crops to plant on each type of cultivated land and the planting area. Considering the actual situation, the main constraint is that the total cultivated area of ​​each crop must be equal to the total cultivated area of ​​the village. The state transfer equation is based on the unit area profit of each plot at the current stage. At the same time, it is also necessary to consider the mutual influence between repeated cropping and crops during crop planting (such as bean crops have a positive effect on the yield of crops in the next year). In other words, at each stage, according to the current unit area profit of the crop, the crop with the highest planting profit is selected as the basis for state transfer. The objective function is to maximize the income of all cultivated land. Specifically:

[0088] Step 1) Problem stage: The entire planting strategy problem is divided into multiple stages in chronological order, each stage corresponds to one year: 2024 to 2030, a total of 7 years, take k = 1, 2, ..., 7. Each year is an independent decision-making stage. The problem to be solved in each stage is how to choose the most profitable crop for each plot of land within the limited plot area M, so as to maximize the total income of the year.

[0089] Step 2) State definition: State T k It is expressed as the maximum profit of the planting strategy of the six plot types in the kth year, including the area of ​​crops planted in all plots. k (i,j) represents the maximum profit when the remaining crop area on plot type i is j in the kth year. The planting allocation with the highest profit is made for the same plot type in the kth year. At this time, the corresponding planting strategy on plot type i in the k-1th year has been determined. According to the non-aftereffect of the state, the planting strategy of the next state, that is, the next year, is only restricted by the previous year.

[0090] Step 3) State transition equation: The state transition equation describes the process of how to transfer from one state to the next stage state. In each stage k, according to the unit area profit of the currently arable crops, the crop with the highest planting profit is selected. Specifically:

[0091]

[0092] Among them, J' is the set of crops that do not include those planted in the previous year, c is the specific crop, j' is the planting area of ​​crop c, and lr is the profit corresponding to the area of ​​j' of crop c. Therefore, dp k (i, j) represents the maximum profit when the remaining planting area of ​​the i-th plot is j in the k-th year.

[0093] Step 4: Model solution

[0094] First, refine each stage of step three into a 0-1 knapsack problem:

[0095] The planting strategy at each stage needs to consider whether these crops should be planted at the current stage and how to arrange the planting to maximize the profit. It is necessary to arrange n kinds of crops to be planted on the total area of ​​M arable land in the village, among which the planting area of ​​the jth crop is s j , the profit per unit area of ​​this crop is v j , specifically:

[0096] v j =b j ×m j -c j

[0097] There are only two possibilities for whether to plant a certain crop: either plant it or not. In the current problem, there will be some situations that do not meet the planting area. Consider joint planting. Therefore, it can be transformed into a 0-1 backpack problem, that is, put some items into a backpack and find the solution that can put the most valuable items. The capacity of the backpack is the total area of ​​available arable land M, and each crop is regarded as an item in the backpack, and its "weight" is the planting area s of the crop. j , the value is its unit area benefit v j The goal is to maximize the total revenue. Based on the above ideas, it is transformed into the following questions:

[0098] Given M>0, s j >0,v j >0, where 1≤j≤n, solve for an n-dimensional 0-1 vector (x1,x2,…,x n ), when x j =1, it means planting the jth crop, otherwise it is not planted. The goal is to maximize profit, that is, to satisfy:

[0099]

[0100] Furthermore, as attached Figure 3Based on the no-aftereffect property of dynamic programming, the initial planting strategy of the current stage is obtained by using the 0-1 knapsack problem and the greedy strategy in each stage, and the specific crop area planted can be obtained.

[0101] Step 5: Model optimization

[0102] By establishing a target programming model, introducing prior knowledge, and optimizing the results using a heuristic algorithm, the specific crop area to be planted can be obtained. The specific optimization model is as follows:

[0103]

[0104] Among them, x ijk is a 0-1 variable, Z k is the total net income in the kth year, b ijk is the selling price, v ijk is the unit area profit of planting the jth crop on the i-th plot in the k-th year, s ijk is the actual yield of the jth crop planted on the ith plot of land in the kth year, M is the total area of ​​arable land, j is the crop number, and j' is the number of the legume crop.

[0105] The algorithm involved in the present invention can be implemented by Matlab software, including data import, solving the initial solution of each stage based on the greedy strategy, and optimizing by using the simulated annealing algorithm.

[0106] The specific operation steps are as follows: first, the collected data, including information such as crop number sequence, unit price, and planting area, are imported into the Matlab environment. The Mat file storage mechanism of Matlab is used to efficiently load data into the workspace. Then, the state matrix, planting strategy matrix, etc. are initialized, and the planting strategy of the previous year is used as the initial reference for the planting strategy of the next year. A 0-1 backpack model is constructed to clarify the state variables of the problem and its state transition equation, and decide whether to continue to plant a certain crop in the second year. The core link of the algorithm is to use the greedy strategy to find the initial solution, follow the basic idea of ​​the greedy algorithm, traverse the numbering sequence of crops, and make accurate decisions at each decision stage based on the preset greedy criterion, that is, give priority to crops with the highest unit area income. Subsequently, the target programming equation is defined, with the goal of maximizing the overall income, combined with key factors such as planting resource constraints and market demand, to ensure that the algorithm has a high practical application value. Then, the initial solution obtained by the greedy algorithm is used as input, combined with rich prior knowledge, and the simulated annealing algorithm is introduced for optimization. The simulated annealing function is used to set the core parameters such as the initial temperature and annealing rate. During the simulated annealing process, the algorithm will strictly follow the Metropolis criterion and accept the neighboring solutions of the current solution with a certain probability, even if the benefits of these solutions are lower than the current solution. This strategy effectively avoids the risk of the algorithm falling into the local optimal solution and significantly improves the global search capability. Through continuous iteration, the algorithm will output the optimal planting strategy that meets expectations. Through this sophisticated algorithm design and implementation, it not only helps farmers maximize their benefits, but also provides a solid scientific basis for the rational allocation of agricultural resources.

[0107] In summary, with the help of the above-mentioned technical scheme of the present invention, by obtaining the existing agricultural resources, arable land restrictions and market conditions in rural areas of a certain region, a dynamic programming model of planting strategy can be constructed, which can be further converted into a 0-1 knapsack problem, combined with a greedy strategy solution, and optimized using a heuristic algorithm. The specific planting strategy for each subsequent year can be obtained, thus realizing intelligent decision-making on crop planting.

Claims

1. A method for optimizing crop planting strategy, characterized in that: The steps include: Step 1: Collect data and obtain the types of cultivated land; Step 2: Pre-process the data of the cultivated land types collected in step 1, and divide them into two parts according to the planting seasons; and establish a sequence of crop numbers that can be planted on each piece of cultivated land; Step 3: According to the current unit area profit of crops, a planting strategy model is established to select the crops with the highest planting profit as the basis for state transfer. The objective function is to maximize the benefits of all cultivated land. Step 4: Solve the planting strategy model established in step 3; Step 5: Optimize the model solved in step 4, establish a target planning model, introduce prior knowledge, use heuristic algorithms to optimize the results, and obtain the specific crop area to be planted.

2. A crop planting strategy optimization method according to claim 1, characterized in that: In the step 1, the cultivated land type includes agricultural resources, cultivation restrictions and market status data, which are used to understand the current agricultural economic status of a certain rural area; The agricultural resource data include the type and area of ​​rural cultivated land in the region, the type of crops planted on each piece of cultivated land, the planting season and the yield; Data on farmland restrictions include: farmland rotation, priority arrangements for specific crops, and farmland requirements for agricultural workers; Market condition data include crop cultivation costs and average selling prices.

3. A crop planting strategy optimization method according to claim 2, characterized in that: In the step 2, the secondarily divided cultivated land types are divided into flat dry land, terraced fields, irrigated land, ordinary greenhouses and smart greenhouses. Different types of cultivated land have different planting seasons. In addition, according to the types of crops and planting seasons collected in step 1; Among them, flat dry land and terraced fields are single-season, irrigated land, ordinary greenhouses and smart greenhouses are double-season. For double-season cultivated land, it will be divided into two plots and arranged at the same time as the single-season cultivated land; In addition, all crops that can be grown on each piece of arable land in each season are included in the arrangement and a crop number sequence is formulated.

4. The crop planting strategy optimization method according to claim 1, characterized in that: The planting strategy model in step 3 is specifically as follows: Step 1) Problem stage: The entire planting strategy problem is divided into multiple stages in chronological order. Each year is an independent decision-making stage. The problem to be solved in each stage is how to choose the crop with the highest profit at the current stage for each plot of land within the limited plot area M, so as to maximize the total income of the year. Step 2) State definition: State T k It is expressed as the maximum profit of the planting strategy of the six plot types in the kth year, including the area of ​​crops planted in all plots, dp k (i,j) represents the maximum profit when the remaining crop area is j on plot type i in year k. The most profitable planting allocation is made for the same plot type in year k. At this time, the corresponding planting strategy on plot type i in year k-1 has been determined. According to the ineffectiveness of the state, the planting strategy of the next state, that is, the next year, is only restricted by the previous year. Step 3) State transfer equation: The state transfer equation describes the process of how to transfer from one state to the next stage state. In each stage k, the crop with the highest profit is selected based on the unit area profit of the current crops that can be planted.

5. A crop planting strategy optimization method according to claim 4, characterized in that: Specifically: Among them, J' is the set of crops that do not include those planted in the previous year, c is the specific crop, j' is the planting area of ​​crop c, lr is the profit corresponding to the area of ​​j' planted with crop c, dp k (i, j) represents the maximum profit when the remaining planting area of ​​the i-th plot is j in the k-th year.

6. A crop planting strategy optimization method according to claim 5, characterized in that: The step 4 breaks down the problem of each stage of dynamic programming into a 0-1 knapsack problem, specifically: The planting area of ​​the jth crop is s j , the profit per unit area of ​​this crop is v j , specifically: v j =b j ×m j -c j There are only two possibilities for whether to plant a certain crop: either plant it or not plant it; therefore, it can be transformed into a 0-1 knapsack problem, that is, put some items into a backpack and find the solution that can put the most valuable items, where the capacity of the backpack is the total area of ​​arable land available M, and each crop is regarded as an item in the backpack, and its "weight" is the planting area s of the crop. j , the value is its unit area benefit v j , the goal is to maximize the total revenue.

7. A crop planting strategy optimization method according to claim 6, characterized in that: Based on the above, the following questions are transformed: Given M>0, s j >0,v j >0, where 1≤j≤n, solve for an n-dimensional 0-1 vector (x1,x2,…,x n ), when x j =1, it means planting the jth crop, otherwise it is not planted, with the goal of maximizing profit, that is, satisfying: Based on the non-aftereffect of dynamic programming, the initial planting strategy of the current stage is obtained by using the 0-1 knapsack problem and the greedy strategy in each stage, and the specific crop area planted can be obtained.

8. The crop planting strategy optimization method according to claim 1, characterized in that: The optimization model of step 5 is as follows: Among them, α is the number of divided arable land blocks, β represents the total number of crop labels, and x ijk is a 0-1 variable, Z k is the total net income in the kth year, b ijk is the selling price, v ijk is the unit area profit of planting the jth crop on the i-th plot in the k-th year, s ijk is the actual yield of the jth crop planted on the ith plot of land in the kth year, M is the total area of ​​arable land, j is the crop number, and j' is the number of the legume crop.

Images