Garbage collection and transportation vehicle-oriented operation path planning method and system

Abstract

The application discloses a kind of operation path planning method and system for garbage collection and transportation vehicle, facing the operation area of garbage collection and transportation vehicle, set node set and vehicle set, establish mathematical model according to garbage collection and transportation rule;Respectively with total cost minimization, time window penalty cost minimization and carbon emission cost minimization as target, establish objective function, and set key constraint condition;For node set and vehicle set, establish multi-section natural number coding and decoding rule, use multi-strategy greedy population initialization method, set several strategies proportionally mixed, generate initial population for mathematical model;Variation evolution is carried out using crossover operator and mutation operator;After merging parent chromosome and child chromosome, the next generation population is screened out, continues to evolve until reaching maximum iteration number, output result.The application can efficiently handle complex constraints, and can quickly converge to high-quality optimal solution.

Description

Technical Field

[0001] This invention belongs to the field of interdisciplinary technology of smart environmental protection and intelligent transportation and logistics, specifically relating to a method and system for planning the operation path of garbage collection vehicles. Background Technology

[0002] With the acceleration of urbanization, the amount of household waste has surged, and the operational efficiency and management cost of waste collection and transportation systems have become core issues in urban governance. The waste collection and transportation process is essentially a complex vehicle routing problem, and in actual operation, multiple conflicting objectives such as economic cost, service timeliness, and carbon emission reduction must be considered simultaneously.

[0003] Current solutions still rely on human experience for job scheduling. This approach is difficult to handle complex constraints such as multiple vehicle types and multiple stations, resulting in limited optimization effects and high costs. Summary of the Invention

[0004] In view of the above-mentioned defects or improvement needs of the existing technology, the present invention proposes a method and system for operation path planning for garbage collection and transportation vehicles.

[0005] To achieve the above objectives, according to one aspect of the present invention, a method for planning the operational path of waste collection vehicles is provided, comprising the following steps: S1. Construct a multi-objective mixed-integer programming model: For the operating area of ​​garbage collection and transportation vehicles, set up a set of nodes and a set of vehicles, and establish a mathematical model based on garbage collection and transportation rules; Objective functions for three objectives—minimizing total cost, minimizing time window penalty cost, and minimizing carbon emission cost—are established, and key constraints are set. S2. Establish multi-segment natural number encoding and decoding rules for the node set and vehicle set: The coding rules specifically include: the chromosome is composed of multiple vehicle segments spliced ​​together end to end, and each vehicle segment represents the complete task of a vehicle; the complete task of a vehicle includes the vehicle number, vehicle type, and the node numbers passed through in sequence; The decoding rules specifically include: parsing based on chromosome structure to generate a job path; S3. A multi-strategy greedy population initialization method is adopted, which sets several strategies and mixes them in proportion to generate an initial population for the mathematical model. S4, Evolution: The parent chromosome is selected through a binary tournament based on the individual's non-dominance level and crowding. The selected parent chromosomes are mutated using crossover and mutation operators to generate offspring chromosomes; The offspring chromosomes are decoded and restored to specific vehicle paths, which are then substituted into the mathematical model to calculate the objective function for each target and verify the constraints. S5. Merge the parent and offspring chromosomes, and use NSGA-II's unique fast non-dominated sorting and crowding calculation to select the next generation population. Continue to evolve through S4 until the maximum number of iterations is reached, and output the results.

[0006] According to the above scheme, the nodes in the node set include departure points, collection points, and transfer points; each node is numbered. The vehicle collection includes multiple vehicles, each with a vehicle type; each vehicle and vehicle type is numbered separately. The numbering structure for each vehicle depot is as follows: departure point number, vehicle type number, waste collection point sequence, transfer station number, and departure point number.

[0007] According to the above scheme, during decoding, each vehicle segment is read sequentially to generate a specific operation path; at the same time, a capacity feasibility check is performed. If adding collection points in sequence leads to overloading, a transfer station is inserted before the collection point for unloading, generating a new operation path.

[0008] According to the above scheme, the objective function with the goal of minimizing total cost consists of fixed vehicle departure costs and variable transportation costs related to distance. In the objective function aimed at minimizing the time window penalty cost, the time window penalty cost is obtained by multiplying the number of collection points that fail to provide services within the required time window by the penalty coefficient; In the objective function aimed at minimizing carbon emission costs, the carbon emission cost is obtained by multiplying the sum of fuel consumption of all vehicles to complete the full task by the carbon emission coefficient.

[0009] According to the above scheme, the key constraints include: The amount of garbage carried by a single vehicle must not exceed its maximum capacity; The total amount of waste transported to the transfer station each day must not exceed its processing capacity. Vehicle operation time is during facility opening hours; Each collection point is served by only one vehicle per trip; The vehicle departs from a certain departure point, completes its mission, and returns to the same departure point. The flow conservation constraint means that the flow of vehicles between collection points remains balanced.

[0010] According to the above scheme, the strategies in S3 include: First strategy: Completely random, accounting for 'a'; Second strategy: Prioritize closest distance, accounting for b%; Third strategy: Prioritize the largest demand, accounting for c%; Fourth strategy: Prioritize the earliest time window, accounting for d; Where a+b+c+d=100%; and b>c≥d≥a.

[0011] According to the above scheme, the crossover operator includes: Segment exchange crossover: Several complete vehicle segments are randomly selected from two parent chromosomes and directly exchanged in the offspring chromosomes; Sequential crossover: Randomly select a continuous gene region from one parent and retain it to the offspring; fill in the unoccupied clearance point positions in the offspring from the clearance point sequence of the other parent in sequence.

[0012] According to the above scheme, the crossover operator also includes crossover based on edge recombination: crossover of several segments of the parent generation, and exchange the complete closed loop of a certain vehicle among the offspring generation, while preserving the internal order of each segment.

[0013] According to the above scheme, the mutation operator includes: Inversion mutation: Randomly select two cut sites on a chromosome and invert the clearance point sequence between the cut sites; Vehicle type and transfer station probability variation: Select a vehicle segment by roulette, and then vary its vehicle type with a certain probability, or vary its transfer station with another probability.

[0014] According to another aspect of the present invention, a work path planning system for garbage collection vehicles is provided, comprising: A multi-objective mixed-integer programming model construction module is used to set up a set of nodes and a set of vehicles for the operating area of ​​garbage collection and transportation vehicles, and to establish a mathematical model based on garbage collection and transportation rules. Objective functions for the three objectives are established, namely minimizing total cost, minimizing time window penalty cost, and minimizing carbon emission cost, and key constraints are set. The encoding and decoding module is used to establish multi-segment natural number encoding and decoding rules for the node set and vehicle set. The encoding rules specifically include: the chromosome is composed of multiple vehicle segments spliced ​​together end to end, and each vehicle segment represents the complete task of a vehicle; the complete task of a vehicle includes the vehicle number, vehicle type, and the node numbers passed through in sequence; the decoding rules specifically include: parsing based on the chromosome structure to generate the job path. The initial population generation module is used to generate an initial population for the mathematical model by using a multi-strategy greedy population initialization method, which sets several strategies and mixes them in proportion. The evolution module is used to select parent chromosomes based on the non-dominance level and crowding degree of individuals through a binary tournament; the selected parent chromosomes are mutated using crossover and mutation operators to generate offspring chromosomes; the offspring chromosomes are decoded to restore the specific vehicle paths, and substituted into the mathematical model to calculate the objective function of each objective and verify the constraints. The iteration module is used to merge the parent chromosome and the offspring chromosome, and select the next generation population through NSGA-II's unique fast non-dominated sorting and crowding calculation. The next generation population continues to evolve through the evolution module until the maximum number of iterations is reached, and the results are output.

[0015] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects: By establishing a mathematical model, the operational logic of waste collection and transportation is deeply integrated; a multi-segment natural number encoding and decoding strategy is adopted, enabling the chromosome to fully and naturally express all decision variables in waste collection and transportation scheduling, and to be well utilized in subsequent evolutionary processes; by setting several strategies to be mixed in proportion to generate the initial population, the quality of the initial population is improved, laying the foundation for quickly finding the optimal solution; by designing crossover and mutation operators, the structural constraints of the path problem are adapted. The method of this invention can efficiently handle complex constraints and quickly converge to a high-quality optimal solution. Attached Figure Description

[0016] Figure 1 This is a flowchart of a method provided in an embodiment of the present invention.

[0017] Figure 2 This is a schematic diagram of chromosome encoding and decoding provided in an embodiment of the present invention.

[0018] Figure 3 This is a schematic diagram of paragraph swapping and crossing provided in an embodiment of the present invention.

[0019] Figure 4 This is a schematic diagram of a sequential crossing provided in an embodiment of the present invention.

[0020] Figure 5 This is a schematic diagram of a flipping mutation provided in an embodiment of the present invention.

[0021] Figure 6 This is a schematic diagram illustrating the probability variation of vehicle type and transfer station provided in an embodiment of the present invention.

[0022] Figure 7 This is a map showing the distribution of waste collection points and facilities provided in an embodiment of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0024] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0025] This invention aims to solve the problem of rapid dispatching of garbage collection and transportation vehicles.

[0026] To achieve the above objectives, according to one aspect of the present invention, this embodiment provides a method for planning the operational path of garbage collection vehicles, such as... Figure 1 As shown, it includes the following steps: S1. Construct a multi-objective mixed-integer programming model: For the operating area of ​​waste collection vehicles, a node set and a vehicle set are set up, and a mathematical model is established based on waste collection rules. The nodes in the node set include departure points, collection points, and transfer points; each node is numbered. The vehicle set includes multiple vehicles, each with a vehicle type; each vehicle and vehicle type is numbered.

[0027] The node set V includes nodes such as departure points, collection points, and transfer stations. The departure point set P is denoted by node p; the collection point set W is denoted by node w; and the transfer station set F is denoted by node f.

[0028] Let K be a set of vehicles and A be a set of other waste collection and transportation vehicles, where A = {A1, A2} represents two types of other waste collection and transportation vehicles, A1 is a 1.5-ton boat-type vehicle and A2 is a 3-ton compactor vehicle.

[0029] The following are the other parameters involved: i and j represent waste collection nodes; Let i be the distance traveled from i to j; The average daily output of garbage at garbage collection point i; Let $\mathbf$ be the amount of garbage collected by the $k$ vehicle in a single trip at collection point $i$. The maximum garbage loading capacity of the k-th vehicle; Let be the actual load of the k-th vehicle during its journey from node i to j; This represents the maximum daily waste capacity that the transfer station can receive. The time it takes for the vehicle to arrive at node v; The time when the vehicle leaves node v; Left time window for departure point; Right time window for departure point; For the left time window of the transfer station; For the right time window of the transfer station; Let be the time consumed by the k-th vehicle collecting garbage at point i. Let f be the unloading time of the kth vehicle at transfer station f; Let k be the distance traveled by vehicle k. The basic dispatch cost for vehicle k; Let β be the unit distance transportation cost for vehicle k; β be the penalty cost for not meeting the time window; γ be the unit distance transportation cost for vehicle k; β be the penalty cost for not meeting the time window. This is the carbon emission factor for fuel combustion, typically taken as 2.621 kg / L; This represents the fuel consumption when traveling from point i to point j. This refers to the unit fuel consumption of a vehicle when it is unloaded. This refers to the unit fuel consumption when the vehicle is fully loaded. This indicates whether vehicle k has traveled from node i to node j; 1 indicates yes, 0 indicates no. This indicates whether vehicle k is being collected at node w; 1 indicates yes, 0 indicates no. To indicate whether vehicle k is in use, record 1 if in use and 0 if not in use; The value is 1 if the vehicle arrives at the collection point w within the required timeframe, and 0 if it does.

[0030] Objective functions for the three objectives—minimizing total cost, minimizing time window penalty cost, and minimizing carbon emission cost—are established, and key constraints are set.

[0031] The objective function specifically includes: Objective 1: Minimize total cost F1 F1 =

[0032] The first item is the fixed vehicle dispatch cost, and the second item is the variable transportation cost related to distance. This indicates whether vehicle k has traveled from node i to node j; 1 indicates yes, and 0 indicates no.

[0033] Objective 2: Minimize the time window penalty cost F2 F2 = β

[0034] The penalty coefficient β is used to penalize collection points that fail to provide services within the required time window.

[0035] Objective 3: Minimize carbon emission costs F3 F3 = γ

[0036] in, The calculation takes into account the effect of load: , E represents the fuel consumption under no-load conditions, while E represents the fuel consumption under full load conditions.

[0037] Key constraints include: The amount of garbage carried by a single vehicle must not exceed its maximum capacity; The total amount of waste transported to the transfer station each day must not exceed its processing capacity. Vehicle operation time is during facility opening hours; Each collection point is served by only one vehicle per trip; The vehicle departs from a certain departure point, completes its mission, and returns to the same departure point. The flow conservation constraint means that the flow of vehicles between collection points remains balanced. This means that the number of vehicles arriving at and serving each collection point is the same, that is, the "number of vehicles flowing into" each collection point is equal to the "number of vehicles flowing out", thereby ensuring the continuity and rationality of vehicle routes.

[0038] S2. Establish multi-segment natural number encoding and decoding rules for the node set and vehicle set: The coding rules specifically include: the chromosome is composed of multiple vehicle segments spliced ​​together end to end, and each vehicle segment represents the complete task of a vehicle; the complete task of a vehicle includes the vehicle number, vehicle type, and the node numbers passed through in sequence.

[0039] In some embodiments, chromosome coding is as follows Figure 2 As shown in (a), a chromosome (a complete scheduling scheme) is represented by a one-dimensional array of natural numbers. The chromosome is composed of multiple vehicle segments joined end to end, and each vehicle segment represents a complete task for one vehicle.

[0040] The structure of each vehicle depot is: [Departure Point Number, Vehicle Type Number, Collection Point Sequence, Transfer Station Number, Departure Point Number]. Example numbering rule: Departure Point {0,1,2,3,4}, Vehicle Type {5,6,...}, Transfer Station {7,8,9}, Collection Points are numbered consecutively starting from 10.

[0041] Example chromosome: {0,5,10,12,15,7,0, 2,6,11,13,14,9,2, 4,5,16,17,18,19,8,4}. This chromosome indicates that a total of 3 vehicles were dispatched. The first vehicle departed from departure point 0, using a type 5 vehicle, serving collection points 10, 12, and 15 in sequence, then unloading at transfer station 7, and finally returning to departure point 0.

[0042] The decoding rules specifically include: parsing based on chromosome structure and generating a job path.

[0043] In some embodiments, chromosome decoding is as follows Figure 2 As shown in (b), the analysis is performed directly based on the chromosome structure, without the need for complex mapping. Each vehicle segment is read sequentially to generate the specific driving path: 0→10→12→15→7→0.

[0044] During the decoding process, a capacity feasibility check is performed simultaneously: if adding collection points in sequence leads to overload, a transfer station is inserted before that point for unloading, generating a new path such as 0→10→12→7→15→0, to ensure that the solution is always feasible.

[0045] S3. A multi-strategy greedy population initialization method is adopted, which sets several strategies and mixes them in proportion to generate an initial population for the mathematical model.

[0046] To improve the quality of the initial population, this invention abandons completely random initialization and uses a proportional mixture of four strategies to generate the initial solution.

[0047] First strategy: Completely random, accounting for 20%, to ensure population diversity.

[0048] The second strategy prioritizes the shortest distance, accounting for 40%, and generates solutions with compact paths and low transportation costs.

[0049] The third strategy is to prioritize the highest demand, allocating 20% ​​of the total to this strategy. Prioritizing services for high-demand locations helps improve vehicle load capacity.

[0050] Fourth strategy: Prioritize the earliest time window, accounting for 20%, and prioritize serving points with tight time windows to reduce the risk of time default.

[0051] This hybrid strategy generates an initial population that combines quality and diversity. This step directly utilizes the model's goal orientation and achieves a good balance in diversity, path length, loading rate, and time compliance, laying a high-quality starting point for subsequent evolution.

[0052] It should be noted that the proportions of the above four strategies can be dynamically adjusted according to the characteristics of the actual problem, rather than being fixed values.

[0053] S4, Evolution, specifically includes: S401. Through a binary tournament, the parent chromosome is selected based on the individual's non-dominance level and crowding.

[0054] S402. The selected parent chromosomes are mutated using crossover and mutation operators to generate offspring chromosomes.

[0055] The probabilistic Pc operation applies segment swapping or sequential crossover and flip mutation or vehicle / transfer station probabilistic mutation. These operators fully utilize the structural information inherent in the multi-segment encoding of S2, ensuring that offspring inherit the superior characteristics of their parents while enabling effective feasibility exploration.

[0056] Crossover operators include: Paragraph swapping and crossing: such as Figure 3As shown, several complete vehicle segments are randomly selected from the two parent chromosomes (such as the second segment of parent P1 and the third segment of parent P2) and directly exchanged in the offspring chromosome. The segment exchange and crossover preserves the complete and feasible sub-path structure in the parent. After the exchange, there may be duplicate or missing clearance points, so a "patch and delete" repair operation is required to ensure that each node is served once and only once.

[0057] Sequential intersection: such as Figure 4 As shown, a continuous gene region is randomly selected from one parent P1 and retained in the offspring; from the clearance point sequence of the other parent P2, unoccupied clearance point positions in the offspring are filled in sequentially. Sequential crossover, while preserving part of the structure of P1, introduces the access order information of P2 to explore new path combinations.

[0058] As a preferred solution, the crossover operator can also include edge-based crossover: crossover several segments of the parent generation, swapping the complete closed loop of a vehicle among the offspring, preserving the internal order of each segment, ensuring that capacity and metadata such as departure and return points are not corrupted, and ensuring the feasibility of the path. Specific operation steps include: 1) Selecting parent chromosomes: First, select two parent chromosomes, P1 and P2. 2) Randomly selecting vehicle segments: Randomly select several vehicle segment indices. 3) Swapping segments: Swap these two segments in the offspring. 4) Repair operation: Check for duplicate or missing clearance points, fill in the empty spaces with excess clearance points, or remove redundant duplicate points.

[0059] In some embodiments, the mutation operator includes: Inversion mutation: such as Figure 5 As shown, by randomly selecting two cut points in a chromosome and reversing the cleanup point sequence between the cut points, a local perturbation within the path can be achieved, which may reveal a better access order.

[0060] Vehicle type and transfer station probability variation: such as Figure 6 As shown, a vehicle segment is selected using a roulette wheel selection method, and its vehicle type is mutated with a certain probability (e.g., 0.6), or its transfer station is mutated with another probability (e.g., 0.4). Taking P1 as an example, if vehicle type mutation is performed and the first segment of the path is selected for mutation using the roulette wheel selection method, the result is as follows. Figure 6 As shown in (a); if the transfer station mutation is performed and the roulette wheel selects the third mutation, the result is as follows. Figure 6 As shown in (b).

[0061] By exploring different resource (vehicle type, unloading point) configuration schemes while keeping the clearing order unchanged, the probability variation of vehicle type and transfer station enhances the algorithm's search capability at the configuration level.

[0062] S403. The offspring chromosomes are decoded using S2 to reconstruct specific vehicle paths, which are then substituted into the mathematical model to calculate the objective function (F1, F2, F3) for each target and verify the constraints. This is direct feedback and guidance from the model to the algorithm's search results.

[0063] S5. Merge the parent and offspring chromosomes and use NSGA-II's unique fast non-dominated sorting and crowding calculation to select the next generation of the population. This process ensures that the population evolves towards the Pareto front (i.e., the optimal balance between cost, time, and carbon emissions) and maintains the diversity of the solution set. Continue the evolution through S4 until the maximum number of iterations is reached, and output the result, which is the optimal solution.

[0064] According to another aspect of the present invention, a work path planning system for garbage collection vehicles is provided, comprising: A multi-objective mixed-integer programming model construction module is used to set up a set of nodes and a set of vehicles for the operating area of ​​garbage collection and transportation vehicles, and to establish a mathematical model based on garbage collection and transportation rules. Objective functions for the three objectives are established, namely minimizing total cost, minimizing time window penalty cost, and minimizing carbon emission cost, and key constraints are set. The encoding and decoding module is used to establish multi-segment natural number encoding and decoding rules for the node set and vehicle set. The encoding rules specifically include: the chromosome is composed of multiple vehicle segments spliced ​​together end to end, and each vehicle segment represents the complete task of a vehicle; the complete task of a vehicle includes the vehicle number, vehicle type, and the node numbers passed through in sequence; the decoding rules specifically include: parsing based on the chromosome structure to generate the job path. The initial population generation module is used to generate an initial population for the mathematical model by using a multi-strategy greedy population initialization method, which sets several strategies and mixes them in proportion. The evolution module is used to select parent chromosomes based on the non-dominance level and crowding degree of individuals through a binary tournament; the selected parent chromosomes are mutated using crossover and mutation operators to generate offspring chromosomes; the offspring chromosomes are decoded to restore the specific vehicle paths, and substituted into the mathematical model to calculate the objective function of each objective and verify the constraints. The iteration module is used to merge the parent chromosome and the offspring chromosome, and select the next generation population through NSGA-II's unique fast non-dominated sorting and crowding calculation. The next generation population continues to evolve through the evolution module until the maximum number of iterations is reached, and the results are output.

[0065] The following simulation analysis uses a specific example to further demonstrate the effectiveness of the present invention.

[0066] Data Source for the Case Study: This paper selects a certain road section as the study area for the experiment. The experimental data includes 79 other waste (household waste) collection points, 1 departure point, and 3 waste transfer stations. A total of 39 vehicles of 2 types can be dispatched: 1.5-ton boat-type trucks and 3-ton compactor trucks. The location data of each waste collection point, collection and transportation facility, and terminal treatment facility (transfer stations and treatment plants, etc.) are all internal actual data obtained through on-site surveys of enterprises. The location distribution map of each waste collection point, departure point, and transfer station is shown below. Figure 7 As shown in Table 1. The horizontal and vertical axes represent the location coordinates, respectively. Vehicle information is shown in Table 1.

[0067] Table 1

[0068] Simulation results analysis: Experiments were conducted using the improved NSGA-II algorithm. Through systematic analysis of the Pareto front, three of the most representative non-dominated optimal solutions were selected from the obtained optimal solutions.

[0069] Option 1 has the lowest total cost of 9784.2 yuan, with a time window penalty cost of 123 yuan and a total carbon emission cost of 391.62 kg. The 1.5-ton boat-type vehicle consumes 41.11L of fuel, and the 3-ton compressed air vehicle consumes 89.56L of fuel. A total of 14 vehicles are used, including 7 1.5-ton vehicles and 7 3-ton vehicles. See Table 2 for the complete routes, departure times, and distances corresponding to the specific vehicle types.

[0070] Table 2

[0071] Option 2 has the lowest carbon emission cost, with a total cost of 10,220.84 yuan, a time window penalty cost of 161 yuan, and a total carbon emission cost of 312.7 kg. The 1.5-ton boat-type vehicle consumes 25.97 L of fuel, and the 3-ton compressed air vehicle consumes 115.7 L of fuel. A total of 12 vehicles are used, including 4 1.5-ton vehicles and 8 3-ton vehicles. See Table 3 for detailed vehicle scheduling.

[0072] Table 3

[0073] Option 3 has the lowest time window penalty cost, with a total cost of 10,946.27 yuan, a time window penalty cost of 50.12 yuan, and a total carbon emission cost of 369.2 kg. The 1.5-ton boat-type vehicle consumes 51.64 L / 100km of fuel, and the 3-ton compressed air vehicle consumes 87.15 L / 100km. A total of 15 vehicles are used, including 8 1.5-ton vehicles and 7 3-ton vehicles. See Table 4 for detailed vehicle scheduling.

[0074] Table 4

[0075] Comparative Analysis with Existing Solutions: Through a systematic analysis of the calculated Pareto front solution set, the three solutions exhibit differences in core indicators. Solution 1, with its more compact route planning, achieves an average load factor of 89%, reducing empty mileage by 17% compared to Solution 3. Solution 2, by increasing the utilization rate of 3-ton compactor trucks, effectively reduces carbon emissions per unit distance by 0.23 L / km, but the high basic dispatch cost of 3-ton trucks increases the total cost by 4.5%. Solution 3 increases vehicle deployment, sacrificing route continuity, but effectively reduces time window defaults through off-peak scheduling, lowering the default rate to 1.2%.

[0076] Optimal Value Analysis: To verify the effectiveness of the improved NSGA-II algorithm, its optimized scheme is compared with the original historical scheduling data for this road segment. Since the company's historical scheduling scheme was entirely based on manual experience and did not consider carbon emission costs and time window penalty costs, the scheme with the lowest total cost (Scheme 1) is selected for comparison with the historical scheme. The total cost of the historical scheme was 16714.39 yuan, while the total cost of Scheme 1 was reduced to 9784.2 yuan, a cost reduction of 41.45%. The historical scheme used an average of 23 vehicles per day, while Scheme 1 only used 14 vehicles, increasing resource utilization by 39.13%. Specific optimization results are shown in Table 5.

[0077] Table 5

[0078] Algorithm Comparison Analysis: To verify the superiority of the improved NSGA-II algorithm, under the same experimental conditions and with the same experimental parameters, it was compared with the basic NSGA-II algorithm. The comparison of various indicators is shown in Table 6. It can be seen that the improved NSGA-II algorithm is far superior to the basic NSGA-II algorithm in terms of convergence speed, solution quality, and computational efficiency. This is because the improved NSGA-II algorithm uses a multi-strategy greedy initialization, which can significantly improve the quality of the initial solution and accelerate the convergence process; at the same time, the two crossover and two mutation methods improve the exploration efficiency of the solution space. Therefore, the improved NSGA-II algorithm largely balances exploration and development, improving the solution speed while calculating a high-quality Pareto front.

[0079] Table 6

[0080] Simulation comparisons revealed that, compared to historical manual methods, the proposed solution reduced total costs by 41.45%, decreased vehicle usage by 39.13%, and increased the average vehicle load factor to 89.1%. This directly demonstrates the significant advantages of the proposed solution in terms of economy and resource utilization.

[0081] Compared with the basic NSGA-II algorithm under the same experimental conditions, as shown in Table 6, the method of this invention has a faster convergence speed: the number of generations required for convergence is reduced from 611 to 327, a reduction of 46.5%; higher computational efficiency: the solution time is shortened from 2840 seconds to 1205 seconds, a reduction of 57.6%; better solution set quality: the feasible solution generation rate is increased from 67.3% to 92.7%; and the Hypervolume (HV) index, which characterizes the overall quality of the solution set, is improved from 0.632 to 0.741, an improvement of 17.2%.

[0082] These data fully demonstrate the effectiveness of the present invention in improving the encoding, initialization, and genetic operators for enhancing the convergence and robustness of the algorithm.

[0083] This invention has strong feasibility assurance: multi-segment coding and structure-aware genetic operators fundamentally ensure that the generated scheduling scheme meets all business constraints (capacity, time window, path closure, etc.), solving the problem of low feasible solution rate of traditional algorithms, and making the algorithm results directly applicable to actual operation.

[0084] This invention provides a set of uniformly distributed Pareto optimal solutions (e.g., the lowest total cost solution, the lowest carbon emission solution, and the solution with the least time default), rather than a single solution, providing decision-makers with a rich space of trade-offs and achieving comprehensive optimization of economy, environment, and service quality.

[0085] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0086] It should be noted that, depending on the implementation needs, the various steps described in this application can be broken down into more steps, or two or more steps or parts of the steps can be combined into new steps to achieve the purpose of this invention.

[0087] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for planning operational paths for garbage collection vehicles, characterized in that: Includes the following steps: S1. Construct a multi-objective mixed-integer programming model: For the operating area of ​​garbage collection and transportation vehicles, set up a set of nodes and a set of vehicles, and establish a mathematical model based on garbage collection and transportation rules; Objective functions for three objectives—minimizing total cost, minimizing time window penalty cost, and minimizing carbon emission cost—are established, and key constraints are set. S2. Establish multi-segment natural number encoding and decoding rules for the node set and vehicle set: The coding rules specifically include: the chromosome is composed of multiple vehicle segments spliced ​​together end to end, and each vehicle segment represents the complete task of a vehicle; the complete task of a vehicle includes the vehicle number, vehicle type, and the node numbers passed through in sequence; The decoding rules specifically include: parsing based on chromosome structure to generate a job path; S3. A multi-strategy greedy population initialization method is adopted, which sets several strategies and mixes them in proportion to generate an initial population for the mathematical model. S4, Evolution: The parent chromosome is selected through a binary tournament based on the individual's non-dominance level and crowding. The selected parent chromosomes are mutated using crossover and mutation operators to generate offspring chromosomes; The offspring chromosomes are decoded and restored to specific vehicle paths, which are then substituted into the mathematical model to calculate the objective function for each target and verify the constraints. S5. Merge the parent and offspring chromosomes, and use NSGA-II's unique fast non-dominated sorting and crowding calculation to select the next generation population. Continue to evolve through S4 until the maximum number of iterations is reached, and output the results.

2. The method for planning operation paths for garbage collection vehicles according to claim 1, characterized in that: The nodes in the node set include departure points, collection points, and transfer points; each node is numbered. The vehicle collection includes multiple vehicles, each with a vehicle type; each vehicle and vehicle type is numbered separately. The numbering structure for each vehicle depot is as follows: departure point number, vehicle type number, waste collection point sequence, transfer station number, and departure point number.

3. The method for planning operation paths for garbage collection vehicles according to claim 2, characterized in that: During decoding, each vehicle segment is read sequentially to generate a specific operation path; at the same time, a capacity feasibility check is performed. If adding collection points in sequence leads to overloading, a transfer station is inserted before the collection point for unloading, generating a new operation path.

4. The method for planning operation paths for garbage collection vehicles according to claim 1, characterized in that: In the objective function that aims to minimize total cost, the total cost consists of fixed vehicle departure costs and distance-related variable transportation costs. In the objective function aimed at minimizing the time window penalty cost, the time window penalty cost is obtained by multiplying the number of collection points that fail to provide services within the required time window by the penalty coefficient; In the objective function aimed at minimizing carbon emission costs, the carbon emission cost is obtained by multiplying the sum of fuel consumption of all vehicles to complete the full task by the carbon emission coefficient.

5. The method for planning operation paths for garbage collection vehicles according to claim 1, characterized in that: Key constraints include: The amount of garbage carried by a single vehicle must not exceed its maximum capacity; The total amount of waste transported to the transfer station each day must not exceed its processing capacity. Vehicle operation time is during facility opening hours; Each collection point is served by only one vehicle per trip; The vehicle departs from a certain departure point, completes its mission, and returns to the same departure point. The flow conservation constraint means that the flow of vehicles between collection points remains balanced.

6. The method for planning operation paths for garbage collection vehicles according to claim 1, characterized in that: The strategies in S3 include: First strategy: Completely random, accounting for 'a'; Second strategy: Prioritize closest distance, accounting for b%; Third strategy: Prioritize the largest demand, accounting for c%; Fourth strategy: Prioritize the earliest time window, accounting for d; Where a+b+c+d=100%; and b>c≥d≥a.

7. The method for planning operation paths for garbage collection vehicles according to claim 2, characterized in that: Crossover operators include: Segment exchange crossover: Several complete vehicle segments are randomly selected from two parent chromosomes and directly exchanged in the offspring chromosomes; Sequential crossover: Randomly select a continuous gene region from one parent and retain it to the offspring; fill in the unoccupied clearance point positions in the offspring from the clearance point sequence of the other parent in sequence.

8. The method for planning operation paths for garbage collection vehicles according to claim 7, characterized in that: Crossover operators also include edge-reorganization-based crossover: crossover several segments of the parent generation, swapping the complete closed loop of a vehicle among the offspring, while preserving the internal order of each segment.

9. The method for planning operation paths for garbage collection vehicles according to claim 2, characterized in that: Mutation operators include: Inversion mutation: Randomly select two cut sites on a chromosome and invert the clearance point sequence between the cut sites; Vehicle type and transfer station probability variation: Select a vehicle segment by roulette, and then vary its vehicle type with a certain probability, or vary its transfer station with another probability.

10. A path planning system for waste collection vehicles, characterized in that: include: A multi-objective mixed-integer programming model construction module is used to set up a set of nodes and a set of vehicles for the operating area of ​​garbage collection and transportation vehicles, and to establish a mathematical model based on garbage collection and transportation rules. Objective functions for the three objectives are established, namely minimizing total cost, minimizing time window penalty cost, and minimizing carbon emission cost, and key constraints are set. The encoding and decoding module is used to establish multi-segment natural number encoding and decoding rules for the node set and vehicle set. The encoding rules specifically include: the chromosome is composed of multiple vehicle segments spliced ​​together end to end, and each vehicle segment represents the complete task of a vehicle; the complete task of a vehicle includes the vehicle number, vehicle type, and the node numbers passed through in sequence; the decoding rules specifically include: parsing based on the chromosome structure to generate the job path. The initial population generation module is used to generate an initial population for the mathematical model by using a multi-strategy greedy population initialization method, which sets several strategies and mixes them in proportion. The evolution module is used to select parent chromosomes based on the non-dominance level and crowding degree of individuals through a binary tournament; the selected parent chromosomes are mutated using crossover and mutation operators to generate offspring chromosomes; the offspring chromosomes are decoded to restore the specific vehicle paths, and substituted into the mathematical model to calculate the objective function of each objective and verify the constraints. The iteration module is used to merge the parent chromosome and the offspring chromosome, and select the next generation population through NSGA-II's unique fast non-dominated sorting and crowding calculation. The next generation population continues to evolve through the evolution module until the maximum number of iterations is reached, and the results are output.

Images