A mixed integer linear programming-based optimization method for unmanned road sweeping vehicle configuration in urban road networks

Abstract

The present invention discloses a method for optimizing the configuration of unmanned sweeping vehicles in urban road networks based on mixed integer linear programming. The method performs mathematical modeling based on target road network information, technical parameters of sweeping vehicles, charging facility information, etc., to form a set of mixed integer linear programming models with the goal of minimizing the total cleaning cost. The total cost includes the driving time cost of the unmanned sweeping vehicle, the charging cost of the unmanned sweeping vehicle at the charging node, and the fixed purchase cost of the unmanned sweeping vehicle. The model constraints include the number of unmanned sweeping vehicles, path planning constraints, working state determination constraints, and power evolution constraints. After optimization, all sweeping vehicles start from the starting point and travel according to the path planning scheme obtained by optimization. During the driving process, they complete the cleaning work of all roads and perform charging operations as needed on the way, and finally return to the starting point to end the work.

Description

Technical Field

[0001] The present invention relates to a method for optimizing the configuration of unmanned sweeping vehicles in urban road networks based on mixed integer linear programming. The method is used to meet the needs of road cleaning in urban road networks and optimize the driving paths, configuration quantity, cleaning schemes, and charging schemes of the unmanned sweeping vehicles. The method belongs to the field of intelligent transportation technology. Background Art

[0002] Research on small cleaning vehicles has been underway since the 1980s. They not only reduce labor intensity but also improve cleaning efficiency. Currently, autonomous driving technologies have become a hot topic in intelligent transportation research, with an increasing number of applications in production and daily life, such as port handling, industrial park distribution, and mining. While intelligent cleaning technologies combining small cleaning vehicles with autonomous driving have emerged, existing autonomous cleaning vehicles are mostly for household use and can only be used in small, enclosed areas, resulting in significant limitations.

[0003] The use of unmanned road sweepers in open areas, such as residential roads and city streets, is an inevitable trend. Key challenges in improving system efficiency and reducing cleaning costs include how to allocate the number of unmanned road sweepers within the road network, determine their operating routes, and optimize their charging schemes. Therefore, developing a scientific and optimized method for configuring unmanned road sweepers in large open areas has important practical significance and application prospects. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for optimizing the configuration of unmanned sweeping vehicles in urban road networks based on mixed integer linear programming. The core idea of ​​this method is to construct a set of mixed integer linear programming models with the goal of minimizing the total cost (including the driving time cost of the unmanned sweeping vehicle, the charging cost of the unmanned sweeping vehicle at the charging node, and the fixed purchase cost of the unmanned sweeping vehicle). Based on the number of unmanned sweeping vehicles equipped with constraints, path planning constraints, working state determination constraints and power evolution constraints, the driving path, number of equipped vehicles, cleaning schemes and charging options of the unmanned sweeping vehicles are optimized to meet the cleaning needs of all sections in the road network. The model mainly includes a road network construction module for constructing a road network, an unmanned sweeping vehicle number equipped module, an unmanned sweeping vehicle driving path planning module, an unmanned sweeping vehicle working state determination module, an unmanned sweeping vehicle power evolution module and an objective function module for achieving the minimum total cost.

[0005] The technical solutions adopted in the present invention are as follows:

[0006] A method for optimizing the configuration of unmanned road sweepers in urban road networks based on mixed integer linear programming is proposed. The method is as follows:

[0007] c1: Given the target road network information to be cleaned, including: a set of road node I, a set of charging nodes S, S∈I, where the unmanned cleaning vehicle can charge; and a set A of all road segments in the road network;

[0008] c2: Determine the specific parameters of the unmanned cleaning vehicle, including the cleaning area per unit time in the cleaning state, the operating speed in the normal driving state, the maximum power, and the power consumption model; combine the specific parameters of the unmanned cleaning vehicle and the target road network information to be cleaned to obtain the set C of unmanned cleaning vehicles and the upper limit C of the number of unmanned cleaning vehicles max In the cleaning state, the time ct that the unmanned cleaning vehicle takes to pass through the road section ij is ij Under normal driving conditions, the time rt taken by the unmanned cleaning vehicle to pass through the road section ij is ij , in the cleaning state, the power CS consumed by the unmanned cleaning vehicle passing through the road section ij ij Under normal driving conditions, the amount of electricity consumed by the unmanned cleaning vehicle when passing through the road section ij is RS ij ;where, (i, j)∈a, a∈A;

[0009] C3: Build a mixed-integer linear programming model to optimize the configuration of the autonomous cleaning vehicles. The configuration includes the number of autonomous cleaning vehicles, the driving path of each autonomous cleaning vehicle, and the charging options for each autonomous cleaning vehicle. The optimization goal of the model is to minimize the total cost of completing the cleaning.

[0010] c4: When solving the model, the driving path and road section information of the unmanned cleaning vehicles are used to obtain the operating status of each vehicle on each road section, including the cleaning state or normal driving state, and then determine the driving time and power changes of each vehicle.

[0011] Furthermore, the established mixed integer linear programming model includes:

[0012] The cleaning vehicle quantity allocation module is used to determine the total number of unmanned cleaning vehicles required for a given target road network to be cleaned;

[0013] The path planning module is used to generate the driving path of the unmanned cleaning vehicle. Each unmanned cleaning vehicle completes the entire cleaning process according to the optimized path plan;

[0014] The working state determination module is used to determine the working state of the unmanned cleaning vehicle on the road section ij during its k-th movement. If the road section ij has been cleaned before, then in the subsequent movement, the working state of the unmanned cleaning vehicle on the road section ij can only be the normal driving state; otherwise, the unmanned cleaning vehicle can choose the cleaning state or the normal driving state;

[0015] The power evolution module is used to calculate the power changes of the unmanned cleaning vehicles during their driving process, including the power changes of each vehicle after each movement on each road section, and the power changes of whether charging is carried out at the node;

[0016] And the objective function module is used to construct the objective function to minimize the total cost. The total cost includes the driving time cost of the unmanned cleaning vehicle, the charging power cost of the unmanned cleaning vehicle at the charging node, and the fixed cost of the unmanned cleaning vehicle.

[0017] Furthermore, the objective function constructed by the objective function module to achieve the minimum total cost is specifically:

[0018]

[0019] Where:

[0020] K: The number of times the unmanned cleaning car moves;

[0021] k: the number of moves, k∈K; the movement of the unmanned cleaning vehicle from node i to node j on the road section ij is called a move;

[0022] α: time value coefficient per unit time;

[0023] β: unit electricity cost;

[0024] γ: fixed cost of each unmanned cleaning vehicle;

[0025] CAR c : binary variable, CAR c =1 means the unmanned cleaning vehicle numbered c participates in the cleaning process, otherwise it does not participate;

[0026] SOC: Maximum power of the unmanned cleaning vehicle;

[0027] S i，k，c : The power information of the k-th mobile unmanned cleaning car c at the road section node i;

[0028] θ i，k，c : binary variable, when θ i，k，c =1, the unmanned cleaning car is charging at charging node i, otherwise it is not charging at node i;

[0029] Sθ i，k，c : Auxiliary variable, S i，k，c and θ i，k，c The product of

[0030] X a，k，c : Binary variable used to determine whether the unmanned cleaning car c is traveling on road section a during the kth movement, X a，k，c =1 means present, otherwise not present;

[0031] M: A large number to set a charging penalty, minimizing the number of charging times while still allowing the cleaning task to be completed with the available power.

[0032] is the travel time cost of the unmanned cleaning vehicle, β∑ i∈I ∑ c∈C ∑ k∈K (SOC-S i，k，c )*θ i，k，c is the charging cost of the unmanned cleaning vehicle at the charging node, γ*∑ c∈C CAR c is the fixed cost of the unmanned cleaning vehicle. i∈I ∑ c∈C ∑ k∈K θ i，k，c This is a charging penalty item set to reduce the number of charging times for the unmanned cleaning car.

[0033] Furthermore, the path planning module generates the driving path of the unmanned sweeping vehicle based on the following constraints. Each unmanned sweeping vehicle completes the entire cleaning process according to the optimized path plan. During each movement, the unmanned sweeping vehicle starts from node i on road section ij and arrives at node j, completing the corresponding task based on the working status of the movement: if the working status of the movement is cleaning, the unmanned sweeping vehicle will clean the road section; if the working status is normal driving, the unmanned sweeping vehicle will only need to drive through the road section normally:

[0034]

[0035]

[0036]

[0037]

[0038] Where: X ij，k，c =1 means it is present, otherwise it is not present; a∈AA is the set of all road sections in the road network, C is the set of unmanned cleaning vehicles, K is the set of the number of times the unmanned cleaning vehicle moves, and I is the set of road section nodes;

[0039] Constraint (1) means that during the entire cleaning process, each road section is guaranteed to be passed by an unmanned cleaning vehicle at least once; Constraint (2) means that each unmanned cleaning vehicle can only travel on one road section at a time; Constraint (3) means that for any unmanned cleaning vehicle, the flow balance in and out of any road section node must be guaranteed, so as to ensure that the entire driving path of each unmanned cleaning vehicle forms a closed loop; Constraint (4) means that for any unmanned cleaning vehicle, the end point of the previous movement is the starting point of the next movement, so as to ensure the continuity of the driving path of each unmanned cleaning vehicle.

[0040] Furthermore, the cleaning vehicle quantity allocation module determines the unmanned cleaning vehicles actually participating in the cleaning work by using the following constraints, and then determines the total number of unmanned cleaning vehicles required:

[0041]

[0042] Where:

[0043] ∑ k∈K X ij，k，c : the sum of the number of times the unmanned cleaning car numbered c moves; M: a large number, X ij，k，c A binary variable used to determine whether the unmanned cleaning car c is traveling on the road section ij during the kth movement. ij，k，c =1 means present, otherwise not present; CAR c : binary variable, CAR c =1 means the unmanned cleaning vehicle numbered c participates in the cleaning process, otherwise it does not participate;

[0044] Constraint (5) indicates that if the number of moves of the unmanned cleaning car numbered c is positive, it is considered to have participated in the cleaning process and its fixed usage cost needs to be calculated; otherwise, it is considered not to have participated in the cleaning process and there is no need to purchase the car additionally.

[0045] Furthermore, the working state determination module determines the working state of the unmanned cleaning vehicle on the road section ij during the k-th movement through the following constraints: if the road section ij has been cleaned before, then in the subsequent movement, the working state of the unmanned cleaning vehicle on the road section ij can only be the normal driving state; otherwise, the unmanned cleaning vehicle can choose the cleaning state or the normal driving state.

[0046]

[0047]

[0048]

[0049]

[0050] Where:

[0051] Y ij，k，c : binary variable, Y ij，k，c =1 means that the unmanned cleaning vehicle numbered c is in the cleaning state when it completes the kth movement on the road section ij, and Y ij，k，c =0 means it is in normal driving state;

[0052] XY ij，k，c : Auxiliary variable, Y ij，k，c and X ij，k，c The product of ij，k，c A binary variable used to determine whether the unmanned cleaning car c is traveling on the road section ij during the kth movement. ij，k，c =1 means present, otherwise not present;

[0053] Constraints (6)-(8) are to set Y ij，k，c and X ij，k，c The product of linearization means that only when X ij，k，c = 1, that is, when the unmanned cleaning vehicle numbered c completes the kth movement on the road section ij, it is necessary to determine whether this movement is cleaning; cleaning the road section ij is equivalent to cleaning the road section ji, so constraint (9) indicates that for any road section ij, it must and only needs to be cleaned once during the entire cleaning process.

[0054] Furthermore, the power evolution module calculates the power changes of the unmanned cleaning vehicle during its travel, including the power changes of each vehicle after each movement on each road section, and the power changes of whether charging is performed at a node, using the following constraints:

[0055]

[0056]

[0057]

[0058]

[0059]

[0060]

[0061]

[0062] Where:

[0063] S i，k，c : The power information of the k-th mobile unmanned cleaning car c at the road section node i; θ i，k，c : binary variable, when θ i，k，c= 1, the unmanned cleaning car is charging at charging node i, otherwise it is not charging at node i; M: a large number, X ij，k，c A binary variable used to determine whether the unmanned cleaning car c is traveling on the road section ij during the kth movement. ij，k，c =1 means present, otherwise not present; Y ij，k，c : binary variable, Y ij，k，c =1 means that the unmanned cleaning vehicle numbered c is in the cleaning state when it completes the kth movement on the road section ij, and Y ij，k，c =0 means it is in normal driving state; SOC: the maximum power of the unmanned cleaning vehicle;

[0064] Constraint (10) specifies the power range of the unmanned cleaning car; Constraint (11) represents the power change from node i to node j when the unmanned cleaning car number c moves through the road section ij for the kth time; Constraints (12)-(14) are to set θ i，k，c and S i，k，c The product of is converted into a linear equation; Constraints (15) and (16) represent the change in power of the unmanned cleaning vehicle numbered c when it moves twice in the same road section node j (road section node j is the end point of the kth movement and the starting point of the k+1th movement), where constraint (15) is the change in power when node j is a charging node, and constraint (16) is the change in power when node j is not a charging node. The beneficial effects of the present invention are:

[0065] This invention provides a method for optimizing the configuration of unmanned road sweepers in urban road networks based on mixed-integer linear programming. A mixed-integer linear programming model is constructed to minimize the total cost of completing road sweeping tasks. This method optimizes the number of unmanned road sweepers deployed, their routes, cleaning schedules, and charging plans. The constructed mixed-integer linear programming model can be directly solved using a commercial numerical solver to obtain a globally optimal solution. This method simultaneously optimizes the cleaning plans of multiple unmanned road sweepers, enabling them to coordinate their work and rationally arranging their charging plans. This method can significantly improve road network cleaning efficiency and reduce cleaning costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 Schematic diagram of road cleaning network

[0067] Figure 2 Unmanned cleaning vehicle 1 moving sequence and working status diagram

[0068] Figure 3 Unmanned cleaning car 2 moving sequence and working status diagram

[0069] Figure 4 Unmanned cleaning car 3 moving sequence and working status diagram

[0070] Figure 5 Road cleaning plan DETAILED DESCRIPTION

[0071] The present invention provides a method for optimizing the configuration of unmanned road sweeping vehicles in urban road networks based on mixed integer linear programming, comprising the following steps:

[0072] c1: Given the target road network information to be cleaned, including the specific road section in the road network, the length of the road section, the charging nodes of the unmanned sweeping vehicles in the road network, and the charging efficiency of the unmanned sweeping vehicles at the charging nodes;

[0073] c2: Determine the specific parameters of the unmanned cleaning vehicle, including the area cleaned per unit time in the cleaning state, the operating speed in the normal driving state, the maximum power consumption, and the power consumption model. Combine the specific parameters of the unmanned cleaning vehicle with the target road network information to be cleaned to determine the cleaning time, normal driving time, cleaning power consumption, and normal driving power consumption for each road section.

[0074] Steps c1 and c2 are the road network construction process, which can be performed using a road network construction module. The obtained parameters are expressed as follows:

[0075] I: road segment node set;

[0076] S: a set of charging nodes where the unmanned cleaning vehicle can be charged, S∈I;

[0077] A: the set of all road segments in the road network;

[0078] C: Collection of unmanned cleaning carts;

[0079] C max : The upper limit of the number of unmanned cleaning vehicles;

[0080] i: road segment node, i∈I;

[0081] j: road segment node, j∈I;

[0082] n: road segment node, n∈I;

[0083] a: represents the road section, a∈A;

[0084] ij: represents the road section, ij∈A; ij and a have the same meaning and can both be used to represent a road section; since cleaning is not direction-specific, road section ji is equivalent to road section ij;

[0085] c: the number of the unmanned cleaning vehicle, c∈C;

[0086] ct ij : The time it takes for the unmanned cleaning vehicle to pass through the road section ij in the cleaning state;

[0087] rt ij : The time it takes for the unmanned cleaning vehicle to pass through the road section ij under normal driving conditions;

[0088] CS ij : The amount of electricity consumed by the unmanned cleaning vehicle when passing through road section ij in the cleaning state;

[0089] RS ij : The amount of electricity consumed by the unmanned cleaning vehicle when passing through road section ij under normal driving conditions.

[0090] c3: Establish a mixed integer linear programming model to solve the configuration optimization plan of the unmanned cleaning vehicles; the configuration plan includes the number of unmanned cleaning vehicles, the driving path of each unmanned cleaning vehicle, and the charging option of each unmanned cleaning vehicle. The model optimization goal is to minimize the total cost of completing cleaning; specifically, the established mixed integer linear programming model includes an unmanned cleaning vehicle number configuration module, an unmanned cleaning vehicle driving path planning module, an unmanned cleaning vehicle working status determination module, an unmanned cleaning vehicle power evolution module and an objective function module for achieving the minimum total cost. The total cost of completing cleaning includes the unmanned cleaning vehicle's driving time cost, the unmanned cleaning vehicle's charging power cost at the charging node, and the unmanned cleaning vehicle's fixed cost.

[0091] Specifically, each module can take the following forms:

[0092] 1.1、Objective function module

[0093] The following expression minimizes the total cost of completing a cleaning task, including the autonomous cleaning vehicle's travel time, the cost of charging the vehicle at the charging node, and the autonomous cleaning vehicle's fixed cost. Furthermore, a penalty term is added to the objective function to minimize the number of times the vehicle needs to be charged while still completing the cleaning task.

[0094]

[0095] Where:

[0096] K: The number of times the unmanned cleaning car moves;

[0097] k: the number of moves, k∈K; the unmanned cleaning vehicle moving from node i to node j on road section ij is called a move, so k is increasing;

[0098] α: time value coefficient per unit time;

[0099] β: unit electricity cost;

[0100] γ: fixed cost of each unmanned cleaning vehicle;

[0101] CAR c: binary variable, CAR c =1 means the unmanned cleaning vehicle numbered c participates in the cleaning process, otherwise it does not participate;

[0102] SOC: Maximum power of the unmanned cleaning vehicle;

[0103] S i，k，c : The power information of the k-th mobile unmanned cleaning car c at the road section node i;

[0104] θ i，k，c : binary variable, when θ i，k，c =1, the unmanned cleaning car is charging at charging node i, otherwise it is not charging at node i;

[0105] Sθ i，k，c : Auxiliary variable, S i，k，c and θ i，k，c The product of

[0106] X a，k，c : Binary variable used to determine whether the unmanned cleaning car c is traveling on road section a during the kth movement, X a，k，c =1 means present, otherwise not present;

[0107] M: A large number, set according to the specific situation of the problem, and is required to be much larger than the final objective function value. In this invention, 9999 can be used to set a charging penalty. Under the premise that the existing power can complete the cleaning task, the number of charging times is minimized;

[0108] is the travel time cost of the unmanned cleaning vehicle, β∑ i∈I ∑ c∈C ∑ k∈K (SOC-S i，k，c )*θ i，k，c is the charging cost of the unmanned cleaning vehicle at the charging node, γ*∑ c∈C CAR c is the fixed cost of the unmanned cleaning vehicle. i∈I ∑ c∈C ∑ k∈K θ i，k，c This is a charging penalty item set to reduce the number of charging times for the unmanned cleaning car.

[0109] 1.2 Path Planning Module

[0110] The following constraints generate the driving paths of the autonomous sweeping vehicles, and each vehicle completes the entire cleaning process according to the optimized path plan. During each movement, the autonomous sweeping vehicle departs from node i on road segment ij and arrives at node j, completing the corresponding task based on the working status of that movement (if the working status of the movement is cleaning, the autonomous sweeping vehicle cleans the road segment; if the working status is normal driving, the autonomous sweeping vehicle simply drives through the road segment normally).

[0111]

[0112]

[0113]

[0114]

[0115] Where:

[0116] Constraint (1) means that during the entire cleaning process, each road section is guaranteed to be passed by an unmanned cleaning vehicle at least once; Constraint (2) means that each unmanned cleaning vehicle can only travel on one road section at a time; Constraint (3) means that for any unmanned cleaning vehicle, the flow balance in and out of any road section node must be guaranteed, so as to ensure that the entire driving path of each unmanned cleaning vehicle forms a closed loop; Constraint (4) means that for any unmanned cleaning vehicle, the end point of the previous movement is the starting point of the next movement, so as to ensure the continuity of the driving path of each unmanned cleaning vehicle.

[0117] 1.3. Module for the number of cleaning trolleys

[0118] The following constraints are used to determine the number of unmanned cleaning vehicles actually involved in the cleaning work, and then determine the total number of unmanned cleaning vehicles required:

[0119]

[0120] Where:

[0121] ∑ k∈K X ij，k，c : The sum of the number of times the unmanned cleaning vehicle numbered c moves;

[0122] Constraint (5) indicates that if the number of moves of the unmanned cleaning car numbered c is positive, it is considered to have participated in the cleaning process and its fixed usage cost needs to be calculated; otherwise, it is considered not to have participated in the cleaning process and there is no need to purchase the car additionally.

[0123] 1.4. Working status determination module

[0124] The working state of the unmanned sweeping vehicle on road section ij during its k-th movement is determined by the following constraints: if road section ij has been cleaned before, then in subsequent movements, the working state of the unmanned sweeping vehicle on road section ij can only be the normal driving state; otherwise, the unmanned sweeping vehicle can choose the cleaning state or the normal driving state.

[0125]

[0126]

[0127]

[0128]

[0129] Where:

[0130] Y ij，k，c : binary variable, Y ij，k，c =1 means that the unmanned cleaning vehicle numbered c is in the cleaning state when it completes the kth movement on the road section ij, and Y ij，k，c =0 means it is in normal driving state;

[0131] XY ij，k，c : Auxiliary variable, Y ij，k，c and X ij，k，c The product of

[0132] Constraints (6)-(8) are to set Y ij，k，c and X ij，k，c The product of linearization means that only when X ij，k，c = 1, that is, when the unmanned cleaning vehicle numbered c completes the kth movement on the road section ij, it is necessary to determine whether this movement is cleaning; cleaning the road section ij is equivalent to cleaning the road section ji, so constraint (9) indicates that for any road section ij, it must and only needs to be cleaned once during the entire cleaning process.

[0133] 1.5. Power Evolution Module

[0134] The following constraints are used to calculate the changes in the power consumption of the unmanned cleaning vehicle during its driving process, including the changes in the power consumption of each vehicle after each movement on each road section, and the changes in the power consumption when charging at the node.

[0135]

[0136]

[0137]

[0138]

[0139]

[0140]

[0141]

[0142] Where:

[0143] Sθ i，k，c : Auxiliary variable, S i，k，c and θ i，k，c The product of

[0144] Constraint (10) specifies the power range of the unmanned cleaning car; Constraint (11) represents the power change from node i to node j when the unmanned cleaning car number c moves through the road section ij for the kth time; Constraints (12)-(14) are to set θ i，k，c and S i，k，c The product of is converted into a linear one; constraints (15) and (16) represent the change in the amount of electricity when the unmanned cleaning vehicle numbered c moves twice in the same road section node j (the road section node j is the end point of the k-th movement and the starting point of the k+1-th movement), where constraint (15) is the change in the amount of electricity when node j is a charging node, and constraint (16) is the change in the amount of electricity when node j is not a charging node.

[0145] c4: Call the CPLEX solver to solve the mixed-integer linear programming model to determine the number of autonomous cleaning vehicles in use, their routes, cleaning status, battery life, and charging status. The model solver uses the autonomous cleaning vehicles' routes and road section information to determine the operating status of each vehicle on each road section, including whether it is cleaning or driving normally. This in turn determines the driving time and battery life of each vehicle.

[0146] Let's take a certain area road network as an example (such as Figure 1 ). There are 10 road nodes and 13 road sections in the road network.

[0147] The implementation steps are as follows:

[0148] Step 1: Determine the specific parameters of the unmanned cleaning vehicle, including the cleaning area per unit time in the cleaning state, the operating speed in the normal driving state, the maximum power, the power consumption model, etc., as shown in the following table;

[0149] Table 1 Parameters of unmanned cleaning vehicle

[0150]

[0151] Step 2: Determine the specific information of the target road network, including the specific road sections in the road network, the length of the road sections, and the charging efficiency of the unmanned vehicle charging nodes and unmanned sweeping vehicles at the charging nodes in the road network;

[0152] Table 2 Road network information

[0153]

[0154]

[0155] Step 3: Determine the cleaning time, normal driving time, cleaning power consumption, and normal driving power consumption for each road section;

[0156] Table 3 Road section information parameters

[0157]

[0158]

[0159] Step 4: Call the CPLEX solver to solve the mixed integer linear programming model to obtain the number of vehicles in use, driving routes, cleaning status, power changes, and charging status of the unmanned cleaning vehicles.

[0160] Table 4 Specific parameter determination of mixed integer linear programming model

[0161]

[0162] The starting point of the cleaning process is set to the road node 1. The solution results show that in order to clean the target road network, a total of 3 unmanned cleaning vehicles are used. The driving paths, working status of each movement, power information and charging status of the 3 unmanned cleaning vehicles are shown in Tables 5, 6 and 7. Figure 2-4 shown.

[0163] The No. 1 unmanned sweeping car made a total of 11 movements. The first movement was from section node 1 to section node 4. The working status was normal driving, that is, not cleaning. Since the end point 4 of section 14 is not a charging point, the car was not charged. The remaining 10 movements were similar to the first movement; the No. 2 unmanned sweeping car made a total of 7 movements. The third movement was from section node 5 to section node 6. The working status was cleaning, which proves that section 56 has been cleaned. The end point 6 of section 56 is a charging point, but the car chose not to charge, so the car was not charged. The remaining 6 movements were similar to the third movement; the No. 3 unmanned sweeping car made 4 movements. The first movement was from section node 1 to section node 2. The working status was cleaning, which proves that section 12 has been cleaned. Since the end point 2 of section 12 is not a charging point, the car was not charged. The remaining 3 movements were similar to the first movement.

[0164] Table 5 Cleaning plan for unmanned cleaning vehicle 1

[0165]

[0166]

[0167] Table 6 Cleaning plan for unmanned cleaning car 2

[0168]

[0169] Table 7 Cleaning plan for unmanned cleaning car 3

[0170]

[0171] Table 8 and Figure 5 The specific information of each road section during the cleaning process is introduced in detail. For example, for road sections (1, 2) and (2, 1), since cleaning does not distinguish between directions, the two expressions mean the same road section, among which (1, 2, 1, normal driving) means that road section 12 is passed by car No. 2 for the first time, and car No. 2 is moving for the first time. The working status of the car is normal driving, that is, not cleaning; (2, 3, 1, cleaning status) means that road section 12 is passed by car No. 3 for the second time, and car No. 3 is moving for the first time, and its working status is cleaning; (3, 2, 7, normal driving) means that road section 12 is passed by car No. 2 for the third time, and car No. 2 is moving for the seventh time, and its working status is normal driving, that is, not cleaning; (4, 1, 11, normal driving) means that road section 12 is passed by car No. 1 for the fourth time, and car No. 1 is moving for the eleventh time, and its working status is normal driving, that is, not cleaning. The specific information of other sections is the same as above.

[0172] Table 8 Specific information of road sections during cleaning process

[0173]

[0174]

[0175] The method of the present invention simultaneously optimizes the cleaning plans of multiple unmanned cleaning vehicles, allowing the multiple unmanned cleaning vehicles to work in a coordinated manner and reasonably arranges the charging plans of the vehicles, which can greatly improve the cleaning efficiency of the road network and save cleaning costs.

[0176] Matters not covered by the present invention are known technologies.

[0177] The above embodiments are intended only to illustrate the technical concepts and features of the present invention. Their purpose is to enable those skilled in the art to understand the contents of the present invention and implement them accordingly. They are not intended to limit the scope of protection of the present invention. Any equivalent changes or modifications made in accordance with the spirit of the present invention are intended to be covered by the scope of protection of the present invention.

Claims

1. A method for optimizing the configuration of unmanned road sweeping vehicles in urban road networks based on mixed integer linear programming, characterized in that: The method is specifically: c1: Given the target road network information to be cleaned, including: a set of road node I, a set of charging nodes S, S∈I, where the unmanned cleaning vehicle can charge; and a set A of all road segments in the road network; c2: Determine the specific parameters of the unmanned cleaning vehicle, including the cleaning area per unit time in the cleaning state, the operating speed in the normal driving state, the maximum power, and the power consumption model; combine the specific parameters of the unmanned cleaning vehicle and the target road network information to be cleaned to obtain the set C of unmanned cleaning vehicles and the upper limit C of the number of unmanned cleaning vehicles max In the cleaning state, the time ct that the unmanned cleaning vehicle takes to pass through the road section ij is ij Under normal driving conditions, the time rt taken by the unmanned cleaning vehicle to pass through the road section ij is ij , in the cleaning state, the power CS consumed by the unmanned cleaning vehicle passing through the road section ij ij Under normal driving conditions, the amount of electricity consumed by the unmanned cleaning vehicle when passing through the road section ij is RS ij ;where, (i,j)∈a, a∈A; C3: Build a mixed-integer linear programming model to optimize the configuration of the autonomous cleaning vehicles. The configuration includes the number of autonomous cleaning vehicles, the driving path of each autonomous cleaning vehicle, and the charging options for each autonomous cleaning vehicle. The optimization goal of the model is to minimize the total cost of completing the cleaning. c4: When solving the model, the driving path and road section information of the unmanned cleaning vehicles are used to obtain the operating status of each vehicle on each road section, including cleaning status or normal driving status, and then determine the driving time and power consumption of each vehicle; The established mixed integer linear programming model includes: The cleaning vehicle quantity allocation module is used to determine the total number of unmanned cleaning vehicles required for a given target road network to be cleaned; The path planning module is used to generate the driving path of the unmanned cleaning vehicle. Each unmanned cleaning vehicle completes the entire cleaning process according to the optimized path plan; The working state determination module is used to determine the working state of the unmanned cleaning vehicle on the road section ij during its k-th movement. If the road section ij has been cleaned before, then in the subsequent movement, the working state of the unmanned cleaning vehicle on the road section ij can only be the normal driving state; otherwise, the unmanned cleaning vehicle can choose the cleaning state or the normal driving state; The power evolution module is used to calculate the power changes of the unmanned cleaning vehicles during their driving process, including the power changes of each vehicle after each movement on each road section, and the power changes of whether charging is carried out at the node; The objective function module is used to construct an objective function to minimize the total cost. The total cost includes the travel time cost of the unmanned cleaning vehicle, the charging power cost of the unmanned cleaning vehicle at the charging node, and the fixed cost of the unmanned cleaning vehicle. The cleaning vehicle quantity allocation module determines the number of unmanned cleaning vehicles actually involved in the cleaning work by using the following constraints, and then determines the total number of unmanned cleaning vehicles required: Where: ∑ k∈K X ij,k,c : the sum of the number of times the unmanned cleaning car numbered c moves; M: a large number, X ij,k,c A binary variable used to determine whether the unmanned cleaning car c is traveling on the road section ij during the kth movement. ij,k,c =1 means present, otherwise not present; CAR c : binary variable, CAR c =1 means the unmanned cleaning vehicle numbered c participates in the cleaning process, otherwise it does not participate; Constraint (5) indicates that if the number of moves of the unmanned cleaning car numbered c is positive, it is considered to have participated in the cleaning process and its fixed usage cost needs to be calculated; otherwise, it is considered not to have participated in the cleaning process and there is no need to purchase the car additionally.

2. The method according to claim 1, wherein: The objective function constructed by the objective function module to achieve the minimum total cost is specifically: Where: K: The number of times the unmanned cleaning car moves; k: the number of moves, k∈K; the movement of the unmanned cleaning vehicle from node i to node j on the road section ij is called a move; α: time value coefficient per unit time; β: unit electricity cost; γ: fixed cost of each unmanned cleaning vehicle; CAR c : binary variable, CAR c =1 means the unmanned cleaning vehicle numbered c participates in the cleaning process, otherwise it does not participate; SOC: Maximum power of the unmanned cleaning vehicle; S i,k,c : The power information of the k-th mobile unmanned cleaning car c at the road section node i; θ i,k,c : binary variable, when θ i,k,c =1, the unmanned cleaning car is charging at charging node i, otherwise it is not charging at node i; Sθ i,k,c : Auxiliary variable, S i,k,c and θ i,k,c The product of X a,k,c : Binary variable used to determine whether the unmanned cleaning car c is traveling on road section a during the kth movement, X a,k,c =1 means present, otherwise not present; M: A large number to set a charging penalty, minimizing the number of charging times while still allowing the cleaning task to be completed with the available power. is the travel time cost of the unmanned cleaning vehicle, β∑ i∈I ∑ c∈C ∑ k∈K (SOC-S i,k,c )*θ i,k,c is the charging cost of the unmanned cleaning vehicle at the charging node, γ*∑ c∈ C CAR c is the fixed cost of the unmanned cleaning vehicle; M*∑ i∈I ∑ c∈C ∑ k∈K θ i,k,c This is a charging penalty item set to reduce the number of charging times for the unmanned cleaning car.

3. The method according to claim 1, wherein: The path planning module generates the driving path of the unmanned cleaning vehicle based on the following constraints. Each unmanned cleaning vehicle completes the entire cleaning process according to the optimized path plan. During each movement, the unmanned cleaning vehicle starts from node i on road section ij and arrives at node j. It completes the corresponding task based on the working status of the movement: if the working status of the movement is cleaning, the unmanned cleaning vehicle cleans the road section; if the working status is normal driving, the unmanned cleaning vehicle only needs to drive through the road section normally: Where: X ij,k,c =1 means it is present, otherwise it is not present; a∈AA is the set of all road sections in the road network, C is the set of unmanned cleaning vehicles, K is the set of the number of times the unmanned cleaning vehicle moves, and I is the set of road section nodes; Constraint (1) means that during the entire cleaning process, each road section is guaranteed to be passed by an unmanned cleaning vehicle at least once; Constraint (2) means that each time an unmanned cleaning vehicle moves, it can only travel on one road section; Constraint (3) means that for any unmanned cleaning vehicle, the flow balance in and out of any road section node must be guaranteed, so as to ensure that the entire driving path of each unmanned cleaning vehicle forms a closed loop; Constraint (4) means that for any unmanned cleaning vehicle, the end point of the previous movement is the starting point of the next movement, so as to ensure the continuity of the driving path of each unmanned cleaning vehicle.

4. The method according to claim 1, wherein: The working state determination module determines the working state of the unmanned cleaning vehicle on the road section ij during its k-th movement according to the following constraints: if the road section ij has been cleaned before, then in the subsequent movement, the working state of the unmanned cleaning vehicle on the road section ij can only be the normal driving state; otherwise, the unmanned cleaning vehicle can choose the cleaning state or the normal driving state; Where: Y ij,k,c : binary variable, Y ij,k,c =1 means that the unmanned cleaning vehicle numbered c is in the cleaning state when it completes the kth movement on the road section ij, and Y ij,k,c =0 means it is in normal driving state; XY ij,k,c : Auxiliary variable, Y ij,k,c and X ij,k,c The product of ij,k,c A binary variable used to determine whether the unmanned cleaning car c is traveling on the road section ij during the kth movement. ij,k,c =1 means present, otherwise not present; Constraints (6)-(8) are to set Y ij,k,c and X ij,k,c The product of linearization means that only when X ij,k,c = 1, that is, when the unmanned cleaning vehicle numbered c completes the kth movement on the road section ij, it is necessary to determine whether this movement is cleaning; cleaning the road section ij is equivalent to cleaning the road section ji, so constraint (9) indicates that for any road section ij, it must and only needs to be cleaned once during the entire cleaning process.

5. The method according to claim 1, wherein: The power evolution module calculates the power changes of the unmanned cleaning vehicles during their travels using the following constraints, including the power changes of each vehicle after each movement on each road section, and the power changes of whether charging is performed at a node: Where: Sθ i,k,c : Auxiliary variable, S i,k,c and θ i,k,c The product of i,k,c : The power information of the k-th mobile unmanned cleaning car c at the road section node i; θ i,k,c : binary variable, when θ i,k,c =1, the unmanned cleaning car is charging at charging node i, otherwise it is not charging at node i; M :A large number, X ij,k,c A binary variable used to determine whether the unmanned cleaning car c is traveling on the road section ij during the k-th movement. ij,k,c =1 means present, otherwise not present; Y ij,k,c : binary variable, Y ij,k,c =1 means that the unmanned cleaning vehicle numbered c is in the cleaning state when it completes the kth movement on the road section ij, and Y ij,k,c =0 means it is in normal driving state; SOC: Maximum power of the unmanned cleaning vehicle; Constraint (10) specifies the power range of the unmanned cleaning car; Constraint (11) represents the power change from node i to node j when the unmanned cleaning car number c moves through the road section ij for the kth time; Constraints (12)-(14) are to set θ i,k,c and S i,k,c The product of is converted into a linear one; constraints (15) and (16) represent the change in the amount of electricity when the unmanned cleaning vehicle numbered c moves twice in the same road section node j (the road section node j is the end point of the k-th movement and the starting point of the k+1-th movement), where constraint (15) is the change in the amount of electricity when node j is a charging node, and constraint (16) is the change in the amount of electricity when node j is not a charging node.

Images