A collection and transportation vehicle scheduling method and system based on urban domestic waste classification

Abstract

The present invention discloses a collection and transportation vehicle scheduling method and system based on urban domestic waste classification, which statistically analyzes the historical waste volume in the service area and predicts the daily waste volume of different types; uses the K-means clustering algorithm to cluster the waste collection points in the entire area and divides them into multiple regions; statistically analyzes the working time and running time of different types of vehicles in the working area; plans and analyzes the configured quantity and routes of the vehicles, calculates the minimum configured quantity of the collection and transportation vehicles and the vehicle routes; compares the driving time required for the vehicles after vehicle configuration and route planning with the historical collection and transportation time, analyzes the efficiency of different planned routes by the collection and transportation vehicle configuration and route planning analysis sub-module, and determines the optimal vehicle collection and transportation route in the area according to the route efficiency. The present invention accurately calculates the transportation cost of the waste collection and transportation vehicles in the area, and intelligently analyzes the optimal configuration and routes of the collection and transportation vehicles for a large number of waste collection points within the scope.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent vehicle configuration, and particularly relates to a collection and transportation vehicle scheduling method and system based on urban domestic waste classification. Background Art

[0002] With the acceleration of the urbanization process and the growth of the population, the social and environmental impacts caused by urban domestic waste have been increasingly taken seriously by people. "Waste besieging the city" has become one of the important factors affecting urban development. How to effectively collect, dispose of and utilize the growing urban domestic waste has become a scientific problem to be solved urgently.

[0003] On the premise of the increasing amount of urban domestic waste, the collection and transportation system based on classification is an important link connecting the front and back ends of waste treatment. If a suitable classification collection and transportation method can be found and the collection and transportation scheduling can be optimized, it can not only effectively utilize the existing collection and transportation system and optimize the operation process, but also will not bring too much increase in collection and transportation costs. And the reverse logistics of urban waste (such as the timely recycling, transportation, treatment, etc. of waste) can effectively reduce pollution and reduce the impact on the living environment of residents. At the same time, a reasonable arrangement of the driving routes of waste collection and transportation vehicles can also achieve the purpose of saving vehicle purchase and collection costs, which is conducive to reducing the land occupation amount, minimizing the pollution of waste to the environment, and conducive to the realization of the sustainable development goal in many aspects such as the entire social resources, economy, and environment. Summary of the Invention

[0004] Object of the Invention: The object of the present invention is to provide a collection and transportation vehicle scheduling method and system based on urban domestic waste classification, which can intelligently analyze the optimal configuration and routes of collection and transportation vehicles for a large number of waste collection points within the scope.

[0005] Technical Solution: The collection and transportation vehicle scheduling method based on urban domestic waste classification described in the present invention specifically includes the following steps:

[0006] (1) Statistically analyze the historical waste volume in the service area, and predict the waste volumes of four different types of waste, namely perishable waste, recyclable waste, hazardous waste, and other waste, every day.

[0007] (2) Use the K-means clustering algorithm to cluster the waste collection points in the entire area and divide them into multiple regions.

[0008] (3) Statistically analyze the working time and running time of different types of vehicles in the working area.

[0009] (4) Plan and analyze the configuration quantity and routes of vehicles, and calculate the minimum configuration quantity of collection and transportation vehicles and the vehicle routes.

[0010] (5) Compare the time required for the vehicle to travel after vehicle configuration and route planning with the historical collection time, analyze the efficiency of different planned routes in the collection vehicle configuration and route planning analysis sub-module, determine the optimal collection route of the vehicle within the area according to the route efficiency, and form a scheduling plan.

[0011] Further, the implementation process of step (2) is as follows:

[0012] Suppose there are N randomly distributed and uneven garbage collection points, randomly select k points as the initial clustering centers among them, calculate and record the distances from the remaining points to the k initial centers, select the minimum distance and classify it into the class corresponding to the initial center, then calculate the average value in each class and update the initial center until the clustering criterion function converges and stops updating, and finally divide all the garbage collection points in the area into k areas.

[0013] Further, the implementation process of step (3) is as follows:

[0014] Mark the working time consumed in the first measurement, measure the time consumed by the collection vehicle driving on the road section, mark the running time consumed in the second measurement, compare the marked working time and running time of each collection vehicle, set the running time comparison value as (0, t n , when the difference between the working time and running time of each collection vehicle is greater than or equal to t n , then re-measure, analyze the re-measured data, mark it after meeting the comparison value threshold, and send the marked data to the garbage collection vehicle configuration and optimal path analysis module.

[0015] Further, step (4) includes the following steps:

[0016] (41) Set the minimum configuration number of the collection vehicle currently analyzed and the algorithm for vehicle routes for conditional constraints. N is the set of garbage collection points, where N = 0 represents the starting point, K is the set of collection vehicles, S is the set of collection points visited by a certain vehicle. Set the model parameters according to this mathematical representation. Set the maximum load capacity of the vehicle as Q, set the distance between collection points i and j as d ij , set the transportation cost per unit distance of vehicle k as C k , set the fixed cost of vehicle k as f k , set the left time window for arriving at collection point i as a i , set the right time window for arriving at collection point i as b i , set the time window width of collection point i as W ik ;

[0017] (42) Set the minimum number of collection vehicles configured within the garbage collection service area as minn;

[0018] (43) It is set that each collection point can only be served by one vehicle and can only be visited once. According to the formula: and constraints are imposed;

[0019] (44) It is set that for each collection point except the starting point and the ending point, the number of vehicles entering and leaving each collection point is equal. The collection vehicle entering a certain collection point must also leave from that point. According to the formula: flow conservation constraints are imposed;

[0020] (45) It is set that each vehicle starts from the garbage vehicle concentration point and must return to the garbage vehicle concentration point after service. According to the formula and constraints are imposed;

[0021] (46) The starting time when all collection vehicles arrive at each collection point and the ending time when they leave each collection point are defined. Among them, according to the formula: constraints are imposed on the earliest arrival time and the latest departure time of all collection vehicles at each collection point;

[0022] (47) It is set that there is a capacity limit for the collection vehicle to ensure that the load of each vehicle does not exceed the vehicle limit. According to the formula constraints are imposed;

[0023] (48) For x ijk and y ijk are defined as 0-1 integer decision variables:

[0024]

[0025]

[0026] Furthermore, step (5) includes the following steps:

[0027] (51) State transition rule: In the ant colony algorithm, when each ant is searching for the optimal path, it independently selects the next access point according to the amount of pheromone remaining on the path and the heuristic information η from point i to point j; the time window compliance of the vehicle accessing the collection point, the vehicle capacity constraint, and the vehicle capacity utilization rate are comprehensively considered in the state transition probability; the state transition rule formula is:

[0028]

[0029] where l represents all possible values; τ ijDenotes the pheromone concentration on edge (i, j), that is, the trail intensity of edge (i, j); α (α≥0) is the information heuristic factor, representing the relative importance of pheromone. The larger its value, the more inclined the ant is to choose the path that more ants have walked through; η ij Denotes the heuristic pheromone on edge (i, j):

[0030] η ij = 1 / d ij

[0031] where d ij is the distance from point i to j; β (β≥0) is the expected heuristic factor, representing the relative importance of visibility. The larger its value, the more inclined to choose the path with a shorter distance; k ij Denotes the capacity utilization rate of the vehicle:

[0032] k ij =(q i +q j ) / Q

[0033] where λ (λ≥0) is the relative importance of the vehicle capacity utilization rate; ε ij is the time window matching value:

[0034] ε ij = 1 / (|s i -e i |+|s i +l i |)

[0035] where s i denotes the time when the vehicle arrives at collection point i; θ (θ≥0) is the relative importance of time matching; p is a random number within the interval [0, 1]; p0 is a fixed value of the algorithm parameter, and its value range is (0, 1);

[0036] When the pheromone concentration, distance, vehicle capacity utilization rate, and time window matching degree between collection points i and j are greater, the proportion of their product is greater. Therefore, the probability of choosing collection point j is also greater;

[0037] (52) Pheromone update rule: After each ant completes a tour, it will update the pheromone on the path. The pheromone concentration update adopts a global update strategy. After all ants construct paths, additional pheromone is injected into all generated feasible paths and the path with the minimum cost. The update formula is as follows:

[0038]

[0039] where: When (i, j) belongs to the feasible path, Otherwise it is 0; when (i, j) belongs to the path with the minimum cost, Otherwise it is 0; D(L k (t)) represents the length of the t-th feasible path, and D(L * ) represents the length of the path with the minimum cost;

[0040] (53) Introduce a negative feedback mechanism: Limit the pheromone level within [τ min , τ max to prevent the pheromone concentration gap between paths from being too large and prompt the algorithm to jump out of the local optimum; Introduce the idea of pheromone smoothing. When the pheromone concentrations are very different, perform a weighted average of the minimum and maximum pheromone values of each edge to relatively reduce the pheromone concentration difference and better generate new search paths. The update formula is as follows:

[0041]

[0042] (54) Premature convergence judgment: Based on the aggregation degree of individuals in the population and the fact that the optimal solution remains unchanged or changes very little after multiple iterations, judge whether the algorithm converges prematurely; The specific judgment is as follows:

[0043] Set the minimum population fitness variance The population fitness variance σ 2 , and the formula is:

[0044]

[0045] f = max{1, max{|f i - f avg |}}

[0046] where n is the number of ants; f i is the i-th degree value; f avg is the current average fitness value of the population; Limit the size of σ 2 by the changing value of f. When , it is considered that the individuals in the population are severely aggregated, the algorithm enters premature convergence, select some relatively good solutions from the initial solutions and perform chaotic optimization again, otherwise continue to the next process;

[0047] Calculate the number of consecutive non-changing iterations of the global optimal value. When the preset number value is reached, it means that the algorithm evolves slowly and gets stuck. Select some relatively good solutions from the initial solutions and perform chaotic optimization again, otherwise continue to the next process;

[0048] (55) Further optimization of the solution: Calculate the loading capacity of each vehicle in the initial path, and find the vehicle with the smallest loading capacity; Add the collection points served last by other vehicles to the service path of the vehicle with the smallest loading capacity, determine whether the constraint conditions are satisfied and improve the objective function to construct an initial solution of higher quality; Considering the time window constraints, use two-point exchange and 2-opt processes for local search in the route improvement, and optimize the obtained optimal path.

[0049] Based on the same inventive concept, the present invention also provides a collection and transportation vehicle scheduling system based on urban domestic waste classification, including: a garbage volume statistical prediction module, a different region clustering and zoning module, a different type vehicle running time statistical module, and a garbage collection and transportation vehicle configuration and optimal path analysis module;

[0050] The garbage volume statistical prediction module is used to count the quantities of different types of garbage in the service area and predict the garbage volume according to the historical statistics;

[0051] The different region clustering and zoning module uses the K-means clustering algorithm to locate and cluster all garbage collection points in the region;

[0052] The different situation vehicle running time statistical module marks the working time measured for the first time; The running time statistical sub-module is used to measure the running time of the collection and transportation vehicle on the road section, mark the running time measured for the second time, compare the marked working time and running time of each collection and transportation vehicle, and set the running time comparison value to (0, t n , when the difference between the working time and the running time of each collection and transportation vehicle is greater than or equal to t n , re-measurement is carried out, and the data obtained from the re-measurement is analyzed and marked after meeting the comparison value threshold, and the marked data is sent to the garbage collection and transportation vehicle configuration and optimal path analysis module;

[0053] The garbage collection and transportation vehicle configuration and optimal path analysis module is used to plan and analyze the configuration quantity and path of the vehicle, calculate the minimum configuration quantity of the vehicle and the vehicle path; The different vehicle path efficiency analysis sub-module is used to compare the total mileage of the path that the vehicle needs to travel after the vehicle configuration and path planning with the total mileage in the service area, analyze the efficiency of different planned paths, and thus determine the optimal path of the collection and transportation vehicle in the collection and transportation area according to the path efficiency, so as to save the working cost.

[0054] Beneficial effects: Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention accurately measures the transportation cost of garbage collection and transportation vehicles in the region and intelligently analyzes the optimal configuration and routes of garbage collection and transportation vehicles for a large number of garbage collection points within the scope. Description of the Drawings

[0055] Figure 1 This is the flow chart of the present invention. Detailed implementation manners

[0056] The present invention will be further described in detail below with reference to the accompanying drawings.

[0057] As Figure 1 shown, the present invention provides a collection and transportation vehicle scheduling method based on urban domestic waste classification, including the following steps:

[0058] Step 1: Statistically analyze the historical waste volume in the service area, and predict the waste volumes of four different types of waste, namely perishable waste, recyclable waste, hazardous waste, and other waste, on a daily basis.

[0059] Statistically analyze the daily collection volumes of recyclable waste, perishable waste, hazardous waste, and other waste in the collection and transportation area in the past 3 years, and use the time series model to accurately estimate the collection volumes of the four types of waste on the next day respectively.

[0060] Step 2: Use the K-means clustering algorithm to cluster the waste collection points in the entire area and divide them into multiple regions.

[0061] Suppose there are N unevenly distributed waste collection points. Randomly select k points as the initial clustering centers, calculate and record the distances from the remaining points to the k initial centers, select the smallest distance and classify it into the class corresponding to the initial center, then calculate the average value in each class and update the initial center until the clustering criterion function converges and stops updating. Finally, divide all the waste collection points in the area into k regions.

[0062] Step 3: Statistically analyze the working time and running time of different types of vehicles in the working area.

[0063] Use the working time statistical sub-module of the collection and transportation vehicle to measure the total working time of the collection and transportation vehicle, and mark the initially measured working time; use the running time statistical sub-module to measure the running time of the collection and transportation vehicle on the directed road section, and mark the second measured running time. Compare the marked working time and running time of each collection and transportation vehicle, and set the running time comparison value as (0, t n , when the difference between the working time and running time of each collection and transportation vehicle is greater than or equal to t n , re-measurement will be carried out. After analyzing the re-measured data to meet the comparison value threshold, mark it, and send the marked data to the waste collection and transportation vehicle configuration and optimal path analysis module.

[0064] Use the collection vehicle configuration and route planning analysis sub-module to plan and analyze the configuration quantity and route of the vehicles, calculate the minimum configuration quantity of the vehicles and the vehicle routes; use the different vehicle route efficiency analysis sub-module to compare the total mileage of the routes that the vehicles need to travel after vehicle configuration and route planning with the total mileage within the working area, analyze the efficiency of different planned routes, so as to determine the optimal route of the collection vehicles in the area according to the route efficiency, thereby saving the working cost.

[0065] Step 4: Plan and analyze the configuration quantity and route of the vehicles, and calculate the minimum configuration quantity of the collection vehicles and the vehicle routes.

[0066] (4.1) Set the minimum configuration quantity of the collection vehicles and the vehicle routes under current analysis to be conditionally constrained by an algorithm. Let N be the set of garbage collection points, where N = 0 represents the starting point, K be the set of collection vehicles, and S be the set of collection points visited by a certain vehicle. Set the model parameters according to this mathematical representation. Set the maximum load capacity of the vehicle to be Q, set the distance between collection points i and j to be d ij , and set the transportation cost per unit distance of vehicle k to be C k , and set the fixed cost of vehicle k to be f k , and set the left time window for arriving at collection point i to be a i , and set the right time window for arriving at collection point i to be b i , and set the time window width of collection point i to be W ik .

[0067] (4.2) Set the minimum number of collection vehicles configured inside the garbage collection service area to be minn.

[0068] (4.3) Set that each collection point can only be served by one vehicle and can only be visited once. According to the formula: and carry out the constraints.

[0069] (4.4) Set that for each collection point except the starting point and the ending point, the number of vehicles entering and leaving each collection point is equal. The collection vehicle must also leave from a certain collection point when it enters. According to the formula: carry out the flow conservation constraint.

[0070] (4.5) Set that each vehicle starts from the garbage vehicle concentration point and must return to the garbage vehicle concentration point after service. According to the formula and carry out the constraints.

[0071] (4.6) Define the starting time when all collection vehicles arrive at each collection point and the ending time when they leave each collection point. Among them, according to the formula: Constraints are imposed on the earliest arrival time and the latest departure time of all collection vehicles at each collection point.

[0072] (4.7) Set the capacity limit of the collection vehicle to ensure that the load of each vehicle does not exceed the vehicle limit. According to the formula Conduct constraints.

[0073] (4.8) Define x ijk and y ijk as 0-1 integer decision variables:

[0074]

[0075]

[0076] Step 5: Compare the travel time required for the vehicle after vehicle configuration and route planning with the historical collection time, analyze the efficiency of different planned routes by the vehicle collection configuration and route planning analysis sub-module, determine the optimal route for vehicle collection in the area according to the route efficiency, and form a scheduling plan.

[0077] Use the different collection vehicle route efficiency analysis sub-module to compare the travel time required for the vehicle after vehicle configuration and route planning with the historical collection time, analyze the efficiency of different planned routes by the vehicle collection configuration and route planning analysis sub-module, so as to determine the optimal route for vehicle collection in the area according to the route efficiency, and save the cost of mechanized operation of garbage collection in the area. Specifically, it includes the following steps:

[0078] (5.1) State transition rule. In the ant colony algorithm, when each ant is looking for the optimal path, it independently selects the next visited point according to the amount of pheromone remaining on the path and the heuristic information η from point i to point j. To ensure the directionality of collection point selection and the diversity of search, this algorithm will comprehensively consider the time window compliance of vehicle access to collection points, vehicle capacity constraints, and vehicle capacity utilization rate in the state transition probability. The state transition rule formula is:

[0079]

[0080] Among them: l represents all possible values; τ ij represents the pheromone concentration on edge (i,j), that is, the track strength of edge (i,j); α (α≥0) is the information heuristic factor, representing the relative importance of pheromone. The larger its value, the more ants tend to choose the path that more ants have walked; η ij represents the heuristic pheromone on edge (i,j), and its expression is:

[0081] η ij = 1 / d ij

[0082] where: d ij is the distance from point i to j; β (β ≥ 0) is the desired heuristic factor, representing the relative importance of visibility. The larger its value, the more inclined to choose a path with a shorter distance; k ij represents the capacity utilization rate of the vehicle, and its expression is:

[0083] k ij =(q i +q j ) / Q

[0084] where: λ (λ ≥ 0) is the relative importance of the vehicle capacity utilization rate; ε ij is the time window matching value, and its expression is:

[0085] ε ij =1 / (|s i -e i |+|s i +l i |)

[0086] where: s i represents the time when the vehicle arrives at collection point i; θ (θ ≥ 0) is the relative importance of time matching; p is a random number within the interval [0, 1]; p0 is a fixed value of an algorithm parameter, and its value range is (0, 1).

[0087] It can be seen from the formula that when the pheromone concentration between collection points i and j is higher, the distance is shorter, the capacity utilization rate of the vehicle is higher, and the time window matching degree is higher, the proportion of their product is larger. Therefore, the probability of selecting collection point j is also larger.

[0088] (5.2) Pheromone update rule. After each ant completes a tour, it updates the pheromone on the path. In this paper, the pheromone concentration update adopts a global update strategy. After all ants construct paths, additional pheromone is injected into all generated feasible paths and the path with the minimum cost. This is a positive feedback process. The update formula is as follows:

[0089]

[0090] where: when (i, j) belongs to the feasible path, otherwise it is 0; when (i, j) belongs to the path with the minimum cost, otherwise it is 0; D(L k (t)) represents the length of the t-th feasible path, and D(L * ) represents the length of the path with the minimum cost.

[0091] (5.3) Max-Min Ant Colony System and pheromone smoothing idea. The selection mechanism of the ant colony algorithm belongs to the positive feedback mechanism, that is, the probability of selecting good paths is relatively large, and this mechanism will inevitably lead to these paths having more competitive advantages in subsequent selections. When the algorithm falls into the local optimum, according to the traditional selection mechanism, it is very difficult to jump out of the local optimum, resulting in premature convergence of the algorithm.

[0092] Therefore, a negative feedback mechanism is introduced into the algorithm. Drawing on the idea of the Max-Min Ant System, the pheromone level is restricted to [τ min , τ max , preventing the pheromone concentration gap between paths from being too large and prompting the algorithm to jump out of the local optimum. In addition, this paper also introduces the idea of pheromone smoothing. When the pheromone concentrations vary greatly, the minimum and maximum pheromones of each edge are weighted and averaged, making the pheromone concentration difference relatively smaller and better generating new search paths. The update formula is as follows:

[0093]

[0094] (5.4) Premature convergence judgment. There are mainly two criteria for judging premature convergence of the algorithm: the aggregation degree of individuals in the population; the optimal solution remains unchanged or changes very little after multiple iterations. The specific judgment is as follows:

[0095] Set the minimum population fitness variance The population fitness variance σ 2 , and the formula is:

[0096]

[0097] f = max{1, max{|f i - f avg |}}

[0098] where: n is the number of ants; f i is the i-th degree value; f avg is the current average fitness value of the population. By continuously changing the value of f, the size of σ 2 is restricted. When , it is considered that the individuals in the population are severely aggregated, and the algorithm enters premature convergence. Some relatively good solutions in the selected initial solutions are selected and chaotic optimization is performed again, otherwise the next process continues.

[0099] Calculate the number of consecutive non-changing iterations of the global optimal value. When the preset number value is reached, it means that the algorithm evolves slowly and falls into stagnation. Some relatively good solutions in the selected initial solutions are selected and chaotic optimization is performed again, otherwise the next process continues.

[0100] (5.5) Further optimization of the solution. To prevent the reversal of the service order of the generated paths, the algorithm seeks a better value by balancing the vehicle loading capacity in the improvement between routes. Its essence is the process of improving the initial solution obtained after the ants construct the paths. First, calculate the loading capacity of each vehicle in the initial path, find the vehicle with the smallest loading capacity, and then add the collection points that are served last by other vehicles to the service path of the vehicle with the smallest loading capacity. Determine whether the constraint conditions are met and improve the objective function, thereby constructing a higher-quality initial solution. Considering the time window constraint, local search is performed on the improvement within the route using the two-point exchange and 2-opt processes to optimize the obtained optimal path.

[0101] Based on the same inventive concept, the present invention also provides a collection and transportation vehicle scheduling system based on urban domestic waste classification, including a garbage volume statistical prediction module, a different-region clustering and zoning module, a running time statistical module for different types of vehicles, and a garbage collection and transportation vehicle configuration and optimal path analysis module; wherein:

[0102] The garbage volume statistical prediction module is used to count the quantities of different types of garbage in the service area and predict the garbage volume according to the historical statistics.

[0103] The different-region clustering and zoning module uses the K-means clustering algorithm to locate and cluster all garbage collection points in the region.

[0104] The running time statistical module for different situations marks the working time measured for the first time; the running time statistical sub-module is used to measure the running time of the collection and transportation vehicle on the road section, mark the running time measured for the second time, compare the marked working time and running time of each collection and transportation vehicle, and set the running time comparison value as (0, t n , when the difference between the working time and the running time of each collection and transportation vehicle is greater than or equal to t n , re-measurement is performed, and the data obtained from the re-measurement is analyzed and marked after meeting the comparison value threshold, and the marked data is sent to the garbage collection and transportation vehicle configuration and optimal path analysis module.

[0105] The garbage collection and transportation vehicle configuration and optimal path analysis module is used to plan and analyze the configuration quantity and path of the vehicles, calculate the minimum configuration quantity of the vehicles and the vehicle paths; the different vehicle path efficiency analysis sub-module is used to compare the total mileage of the paths that the vehicles need to travel after the vehicle configuration and path planning with the total mileage in the service area, analyze the efficiency of different planned paths, and thus determine the optimal path of the collection and transportation vehicles in the collection and transportation area according to the path efficiency, so as to save the working cost.

[0106] It is obvious to those skilled in the art that the present invention is not limited to the details of the above-described exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention. Therefore, in all respects, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, it is intended to embrace all changes falling within the meaning and scope of the equivalent elements of the claims in the present invention.

Claims

1. A collection and transportation vehicle scheduling method based on urban domestic waste classification, characterized in that, It includes the following steps: (1) Statistically analyze the historical garbage volume in the service area and predict the garbage volumes of four different types, namely perishable garbage, recyclable garbage, hazardous waste, and other waste, on a daily basis; (2) Use the K-means clustering algorithm to cluster the garbage collection points in the entire area and divide them into multiple regions; (3) Statistically analyze the working time and running time of different types of vehicles in the working area; (4) Plan and analyze the configured number and routes of the vehicles, and calculate the minimum configured number of collection and transportation vehicles and the vehicle routes; (5) Compare the time required for the vehicles to travel after vehicle configuration and route planning with the historical collection and transportation time, analyze the efficiency of different planned routes by the collection and transportation vehicle configuration and route planning analysis sub-module, determine the optimal collection and transportation route of the vehicles in the area according to the route efficiency, and form a scheduling plan; The implementation process of step (2) is as follows: Suppose there are N garbage collection points that are unevenly distributed. Randomly select k points as the initial clustering center points, calculate and record the distances from the remaining points to the k initial center points, select the minimum distance and classify it into the class corresponding to the initial center point, then calculate the average value in each class and update the initial center point until the clustering criterion function converges and stops updating. Finally, divide all the garbage collection points in the area into k regions; The implementation process of step (3) is as follows: Mark the working time consumed in the first measurement, measure the time consumed by the collection and transportation vehicle driving on the road section, mark the running time consumed in the second measurement, compare the working time consumed and the running time consumed by each collection and transportation vehicle after marking, and set the running time comparison value to (0, t n , when the difference between the working time consumed and the running time consumed by each collection and transportation vehicle is greater than or equal to t n , re-measurement is carried out. After the data obtained from the re-measurement is analyzed and meets the comparison value threshold, it is marked, and the marked data is sent to the garbage collection and transportation vehicle configuration and optimal path analysis module; Step (4) includes the following steps: (41) Set the minimum number of configured collection vehicles for the current analysis and use an algorithm for vehicle routing for conditional constraints. N is the set of waste collection points, where N = 0 represents the starting point. K is the set of collection vehicles, and S is the set of collection points visited by a certain vehicle. Set the model parameters according to this mathematical representation. Set the maximum load capacity of the vehicle to Q, and set the distance between collection points i and j to d ij , set the transportation cost per unit distance of vehicle k to c k , set the fixed cost of vehicle k to f k , set the left time window for arriving at collection point i to a i , set the right time window for arriving at collection point i to b i , set the time window width of collection point i to w ik ; (42) Set the minimum configured number of collection and transportation vehicles in the garbage collection and transportation service area as min n; (43) Set that each collection point can only be served by one vehicle and can only be visited once. According to the formula: and constrain; (44) Set each collection point except the starting point and the ending point, such that the number of vehicles entering and leaving each collection point is equal. The collection vehicle must also leave from the collection point it enters. According to the formula: Perform flow conservation constraints; (45) It is set that each vehicle starts from the garbage vehicle concentration point and must return to the garbage vehicle concentration point after service, and constraints are carried out according to the formulas and ; (46) Define the starting time when all collection vehicles arrive at each collection point and the ending time when they leave each collection point. Among them, according to the formula: Constrain the earliest arrival time and the latest departure time of all collection vehicles at each collection point; (47) Set the capacity limit of the collection and transportation vehicles to ensure that the load of each vehicle does not exceed the vehicle limit, and perform constraints according to the formula ; (48) Define for x ijk and y ijk as 0-1 integer decision variables: Step (5) includes the following steps: (51) State transition rule: In the ant colony algorithm, when each ant is looking for the optimal path, it independently selects the next access point according to the amount of pheromone remaining on the path and the heuristic information η from point i to point j. The time window compliance, vehicle capacity constraint, and vehicle capacity utilization rate of the vehicle accessing the collection point are comprehensively considered in the state transition probability. The state transition rule formula is: where l represents all possible values; τ ij represents the pheromone concentration on edge (i, j), that is, the trail intensity of edge (i, j); α (α ≥ 0) is the information heuristic factor, indicating the relative importance of pheromone. The larger its value, the more inclined the ant is to choose the path that more ants have walked; η ij represents the heuristic pheromone on edge (i, j): η ij = 1 / d ij where d ij is the distance from point i to j; β (β≥0) is the expected heuristic factor, representing the relative importance of visibility. The larger its value, the more inclined to choose a path with a shorter distance; k ij represents the capacity utilization rate of the vehicle: k ij = (q i + q j ) / Q where λ (λ≥0) is the relative importance of the vehicle capacity utilization rate; ε ij is the time window matching value: ε ij = 1 / (|s i - e i | + |s i + l i |) where s i represents the time when the vehicle arrives at collection point i; θ (θ≥0) is the relative importance of time matching; p is a random number within the interval [0,1]; p0 is a fixed value of an algorithm parameter, and its value range is (0,1); When the pheromone concentration, distance, vehicle capacity utilization rate, and time window compliance between collection points i and j are greater, the proportion of their product is greater. Therefore, the probability of selecting collection point j is also greater; (52) Pheromone update rule: After each ant completes a tour, it will update the pheromone on the path. The pheromone concentration update adopts a global update strategy. After all ants construct the paths, additional pheromone is injected into all generated feasible paths and the path with the minimum cost. The update formula is as follows: Where: when (i, j) belongs to the feasible path, otherwise it is 0; when (i, j) belongs to the path with the minimum cost, otherwise it is 0; D(L k (t)) represents the length of the t-th feasible path, D(L * ) represents the length of the path with the minimum cost; (53) Introduce a negative feedback mechanism: limit the pheromone level within [τ min , τ max , prevent the pheromone concentration gap between paths from being too large, and prompt the algorithm to jump out of the local optimum; introduce the idea of pheromone smoothing. When the pheromone concentrations differ greatly, perform a weighted average of the minimum and maximum pheromone values of each edge to relatively reduce the pheromone concentration difference and better generate new search paths. The update formula is as follows: (54) Premature convergence judgment: Based on the aggregation degree of individuals in the population and the fact that the optimal solution does not change or changes very little after multiple iterations, judge whether the algorithm converges prematurely. The specific judgment is as follows: Set the minimum population fitness variance Population fitness variance σ 2 , and the formula is: f = max{1, max{|f i - f avg |}} where n is the number of ants; f i is the i-th degree value; f avg is the current average fitness value of the population; the size of σ is restricted by the continuously changing value of f 2 When it is considered that the individuals in the population are severely aggregated, and the algorithm enters premature convergence. Some relatively good solutions in the selected initial solutions are optimized by chaos again, otherwise the next process is continued; Calculate the number of consecutive non-changing iterations of the global optimal value. When the preset number value is reached, it means that the algorithm evolves slowly and falls into stagnation. Select some relatively good solutions in the initial solution and perform chaotic optimization again, otherwise continue to the next process; (55)Further optimization of the solution: Calculate the loading capacity of each vehicle in the initial path, and find the vehicle with the smallest loading capacity; Add the collection points served last by other vehicles to the service path of the vehicle with the smallest loading capacity, judge whether the constraint conditions are satisfied and improve the objective function to construct an initial solution of higher quality; Considering the time window constraint, local search is performed on the improvement within the route using the two-point exchange and 2-opt processes to optimize the obtained optimal path.

2. A collection and transportation vehicle scheduling system based on urban domestic waste classification using the method described in claim 1, characterized in that, Including: Garbage volume statistical prediction module, different area clustering and zoning module, different type vehicle operation time-consuming statistical module, and garbage collection vehicle configuration and optimal path analysis module; The garbage volume statistical prediction module is used to count the quantities of different types of garbage in the service area and predict the garbage volume based on historical statistics; The different area clustering and zoning module uses the K-means clustering algorithm to locate and cluster all garbage collection points in the area; The vehicle operation time-consuming statistics module for different situations marks the working time-consuming measured for the first time; the operation time-consuming statistics sub-module is used to measure the time-consuming of the collection vehicle driving on the road section, mark the operation time-consuming measured for the second time, compare the marked working time-consuming and operation time-consuming of each collection vehicle, and set the operation time-consuming comparison value as (0, t n , when the difference between the working time-consuming and the operation time-consuming of each collection vehicle is greater than or equal to t n , re-measurement is carried out. After the data obtained from the re-measurement is analyzed and meets the comparison value threshold, it is marked, and the marked data is sent to the garbage collection vehicle configuration and optimal path analysis module; The garbage collection vehicle configuration and optimal path analysis module is used to plan and analyze the configuration quantity and path of the vehicle, and calculate the minimum configuration quantity of the vehicle and the vehicle path; The different vehicle path efficiency analysis sub-module is used to compare the total mileage that the vehicle needs to travel after vehicle configuration and path planning with the total mileage in the service area, analyze the efficiency of different planned paths, and thus determine the optimal path of the collection vehicle in the collection area according to the path efficiency to save the working cost.

Images