CN116543564A

CN116543564A - Optimization method and system applied to traffic control

Info

Publication number: CN116543564A
Application number: CN202310825975.7A
Authority: CN
Inventors: 刘君; 李静林; 邹思思; 李永
Original assignee: Xintang Xintong Zhejiang Technology Co ltd
Current assignee: Xintang Xintong Zhejiang Technology Co ltd
Priority date: 2023-07-07
Filing date: 2023-07-07
Publication date: 2023-08-04
Anticipated expiration: 2043-07-07
Also published as: CN116543564B

Abstract

The invention discloses an optimization method and system applied to traffic control, and relates to the field of traffic control. Traffic control is defined as task allocation and path planning, with vehicles acting as agents to perform tasks that move from one location to another. ASP codes the task allocation and ordering problem, generates ASP program of the task, codes the vehicle routing and scheduling problem, and generates ASP program of the routing and scheduling. And expanding the two ASP programs by using differential constraint, and processing time and scheduling constraint to obtain an expanded ASP program. And solving the expanded ASP program by using a mixed ASP solver to obtain a scheduling solution meeting the constraint, and applying the scheduling solution to vehicle scheduling. The invention provides a traffic control optimization scheme with lower cost and higher efficiency, and improves urban traffic efficiency and smoothness.

Description

Optimization method and system applied to traffic control

Technical Field

The invention relates to the field of traffic control, in particular to an optimization method and system applied to traffic control.

Background

Traffic control is a complex problem that requires processing and analysis of a large amount of traffic information, traffic rules, and real-time traffic demands. The prior art mainly includes neural network-based methods and reinforcement learning-based methods.

Neural networks are a deep learning technique that has shown powerful performance in many areas. In the field of traffic control, a neural network can build a model capable of predicting traffic flow and planning a path by learning a large amount of historical traffic data. However, the training process of neural networks requires a significant amount of time and computational resources, and the adaptation of the trained model to new, unseen conditions may be limited. Furthermore, the black box nature of neural networks makes their decision process difficult to understand, which may in some cases raise security and reliability issues.

Reinforcement learning is a method of learning an optimal behavior strategy through interactions with an environment. In the traffic control problem, the reinforcement learning algorithm can learn the optimal traffic control strategy through continuous trial and error by establishing a simulation environment. However, reinforcement learning stability and convergence problems make it challenging to implement in practical applications. In addition, reinforcement learning also requires a significant amount of time and computational resources to learn, and in the face of complex constraints, may not give a solution that satisfies all constraints.

Existing neural networks and reinforcement learning methods have advantages in some respects, such as being able to process large-scale data, being able to learn complex nonlinear relationships, etc., but generally require higher training costs and training periods.

Disclosure of Invention

The invention aims to provide an optimization method applied to traffic control so as to solve the problems in the background art.

In order to achieve the above purpose, the invention adopts the following technical scheme:

an optimization method applied to traffic control, comprising the following steps:

s1, modeling an urban traffic network as a weighted directed graph, wherein vertexes in the graph represent intersections, directed edges represent roads, weights of the edges represent minimum running time through the road section, and the minimum running time is specifically expressed as G= (V, E, W), wherein V represents a vertex set, E represents an edge set, W represents a weight function, and G represents the weighted directed graph;

s2, defining traffic control problems as problems of task allocation and path planning, wherein a vehicle is regarded as an agent for executing tasks, and the tasks are defined as vehicles moving from one place to another place, and are specifically expressed as task sets T= { T1, T2, & gt, tn }, wherein each task ti represents that a vehicle moves from one place to another place;

s3, encoding task allocation and sequencing questions by using an Answer Set Program (ASP) to generate an ASP program (T) related to the tasks;

s4, encoding the vehicle routing and scheduling problems by using an Answer Set Program (ASP) to generate an ASP program (G, T) related to routing and scheduling;

s5, expanding the two ASP programs by using differential constraint, and processing time and scheduling related constraint to obtain an expanded ASP program;

s6: and solving the expanded ASP program by using a mixed ASP solver to obtain a scheduling solution S= solver (ASPprogram) meeting all constraints, and using the scheduling solution for scheduling the vehicle.

Preferably, the vehicle tasks include, but are not limited to, route planning from one location to another, vehicle speed adjustment, and parking spot selection.

Preferably, the weighted directed graph modeled by the urban traffic network further comprises real-time traffic information, wherein the real-time traffic information at least comprises traffic jams, accidents and road maintenance.

Preferably, the process in which the Answer Set Program (ASP) encodes task allocation and ordering questions, and vehicle routing and scheduling questions, further includes the input of real-time data to recalculate and update task allocation and path planning in the face of dynamic changes in the traffic network.

Preferably, the time and schedule related constraints in S5 include at least: the expected start time interval and end time interval of each task, and the dependency relationship among different tasks.

The invention also discloses an optimizing system for traffic control, which comprises:

the urban traffic network modeling module is used for modeling the urban traffic network into a weighted directed graph, wherein the vertexes in the graph represent intersections, the directed edges represent roads, and the weights of the edges represent the minimum running time of the road sections;

a task definition module for defining traffic control problems as a task allocation and path planning problem, wherein the vehicle is regarded as a proxy for executing the task, and the task is defined as the movement of the vehicle from one place to another place;

the ASP coding module is used for coding task allocation and sequencing problems by using an Answer Set Program (ASP), generating an ASP program (T) related to the tasks and coding vehicle routing and scheduling problems, and generating an ASP program (G, T) related to routing and scheduling;

the constraint expansion module is used for expanding the two ASP programs by using differential constraint, processing time and scheduling related constraint, and obtaining an expanded ASP program;

and the mixed ASP solver module is used for solving the expanded ASP program to obtain a scheduling solution S= solver (ASPprogram) meeting all constraints.

Preferably, the vehicle tasks defined by the task definition module include, but are not limited to, route planning from one location to another, vehicle speed adjustment, and parking spot selection.

Preferably, the urban traffic network modeling module further comprises real-time traffic information, and the real-time traffic information at least comprises traffic jams, accidents and road maintenance.

Preferably, the encoding process of the ASP encoding module further comprises the input of real-time data to recalculate and update the mission allocation and path planning in the face of dynamic changes in the traffic network.

Preferably, the constraint expansion module processes time and scheduling related constraints at least include: an expected start time interval and an end time interval for each task; dependency relationship between different tasks.

The invention has the advantages compared with the prior art that:

the calculation cost is low: the invention adopts a method based on an Answer Set Program (ASP) to optimize task allocation and path planning. ASP is a logical programming language and solver that performs problem solving based on logical rules and constraints. The computational cost of an ASP is typically lower compared to neural network training because it does not require extensive data training and model optimization.

The algorithm complexity is low: the algorithm complexity of the invention is relatively low, mainly because the ASP-based optimization method can solve the scheduling solution meeting the constraint condition in a short time. In contrast, neural network training typically requires long training times and significant computational resources, especially for complex network structures and large-scale data sets.

The interpretability is strong: the ASP method provides strong interpretability and can express the solving process and the result of the problem in the form of logic rules. This allows the present invention to better understand and analyze the generation process of the scheduling solution, identifying rules and constraints therein. While neural networks are typically a black box model, it is difficult to interpret the decision process and results.

Flexibility and adjustability: the ASP-based method has higher flexibility and adjustability, and can be adjusted and expanded according to the requirements of practical problems. The invention can add, modify or delete rules and constraints to meet different task allocation and path planning requirements. In contrast, neural networks typically require retraining or redesigning the network architecture to accommodate different problems.

Drawings

Fig. 1 is a flow chart of the method of the present invention.

Detailed Description

The following describes specific embodiments of the present invention with reference to the drawings.

As shown in fig. 1, the optimization method applied to traffic control of the present invention comprises the following steps:

s3, encoding task allocation and sequencing questions by using an Answer Set Program (ASP) to generate an ASP program (T) related to the tasks; specifically, the method comprises the following steps:

rules and constraints for determining task allocation and ordering problems: the requirements of task allocation and ordering problems are analyzed and the rules and constraints involved are determined. Including the starting and ending points of the tasks, the dependencies between the tasks, the time constraints of task execution, etc.

Logic defining an ASP program: logic portions of the ASP program are written using ASP language in accordance with rules and constraints of task allocation and ordering problems. The ASP language includes elements such as predicates, rules, and constraints for describing the logical structure and conditions of the problem. At this step, predicates need to be defined to represent the task, vehicle, dependency between tasks.

ASP rules and constraints are defined: in the logic portion of the ASP program, rules and constraints are written to describe the requirements of the task allocation and ordering problem. These rules and constraints will use the predicates previously defined, in conjunction with logical operators and the reasoning mechanism of the ASP to represent the constraints of the problem. For example, rules may be defined to ensure that each task is assigned to only one vehicle, or constraints may be defined to limit the chronological order in which the vehicle performs the tasks.

Adding initial facts and problem instances: according to the specific task allocation and ordering problem instances, add initial facts and problem instances to the ASP program. These initial facts may be the start status of the task, the vehicle, and other relevant information. These facts will provide the starting point and context for the ASP solver so that the solver can find solutions that satisfy the constraints.

Saving the ASP program: the written ASP program is saved as a text file for subsequent use and input to the ASP solver.

S4, encoding the vehicle routing and scheduling problems by using an Answer Set Program (ASP) to generate an ASP program (G, T) related to routing and scheduling; it may also include:

rules and constraints for determining vehicle routing and scheduling problems: the requirements of the vehicle routing and scheduling problem are analyzed and the rules and constraints involved are determined. Including the starting and ending locations of the vehicles, topology information of the traffic network, collision avoidance between vehicles, etc. A clear definition and understanding of these rules and constraints is ensured.

Logic defining an ASP program: logic portions of the ASP program are written using ASP language in accordance with rules and constraints of the vehicle routing and scheduling problem. The ASP language includes elements such as predicates, rules, and constraints for describing the logical structure and conditions of the problem. At this step, predicates need to be defined to represent vehicles, traffic networks, relationships between vehicles, and the like.

ASP rules and constraints are defined: in the logic portion of the ASP program, rules and constraints are written to describe the requirements of the vehicle routing and scheduling problem. These rules and constraints will use the predicates previously defined, in conjunction with logical operators and the reasoning mechanism of the ASP to represent the constraints of the problem. For example, rules may be defined to ensure that the path of each vehicle is legal, avoid collisions between vehicles, and so forth.

Adding initial facts and problem instances: according to the specific vehicle routing and scheduling problem instances, add initial facts and problem instances to the ASP procedure. These initial facts may be the starting state of the vehicle, the topology of the traffic network, and other relevant information. These facts will provide the starting point and context for the ASP solver so that the solver can find solutions that satisfy the constraints.

The following are some possible predicate definition examples:

vehicle predicate (Vehicle predicate): attributes and states of the vehicle are defined. For example, vehicle (V) may be represented as being present using vehicle (V). The predicate may be extended to represent attributes of a starting location, an ending location, a current location, etc. of the vehicle.

Traffic network predicate (Road predicate): attributes and relationships of roads in the traffic network are defined. For example, road (R) may be used to represent the presence of road R. The predicate may be extended to represent attributes of a start intersection, an end intersection, a transit time, etc. of the road.

Intersection predicate (Intersection predicate): attributes and relationships of intersections in the traffic network are defined. For example, intersection (I) may be used to indicate the presence of intersection I. The predicate may be extended to represent attributes of adjacency, capacity, etc. of the intersection.

Path predicate (Path predicate): path information of the vehicle is defined. For example, path (P, V) may be used to represent path P of vehicle V. The predicate may be extended to represent attributes of a start intersection, an end intersection, a passing road, etc. of the path.

Dependency predicate (Dependency predicate): dependency relationships between tasks are defined. For example, dependency (T1, T2) may be used to indicate that task T1 depends on task T2. The predicate may be extended to represent attributes of a start location, an end location, etc. of the task.

S5, expanding the two ASP programs by using differential constraint, and processing time and scheduling related constraint to obtain an expanded ASP program; more specifically:

determining the need for differential constraints: the time and scheduling constraints involved in task allocation and routing scheduling problems are analyzed. Including vehicle arrival time limitations, task execution time limitations, intersection maximum capacity limitations, and the like. Ensuring a clear definition and understanding of these constraints.

Differential variables are introduced: to introduce differential constraints, differential variables need to be introduced for the relevant points in time and time intervals. The differential variable represents a difference in relative time or a difference in time interval. For example, a variable d [ i, j ] may be introduced to represent the time required for the vehicle to traverse the road (i, j), or a variable d [ t1, t2] may be introduced to represent the time difference between the task t1 and the task t2.

Defining a differential constraint: differential variables are used to define time and scheduling related constraints according to the requirements of the problem. For example, d [ i, j ] <=t may be defined to limit the time for the vehicle to traverse the road (i, j) to not exceed T, or d [ T1, T2] > =d [ T2, T3] +t may be defined to indicate that task T2 must begin at least T after task T3.

Extending the ASP program: based on the existing ASP program, differential variables and differential constraints are introduced. These variables and constraints are added to the ASP program (ASPprogram (T)) for task allocation and path planning and the ASP program (ASPprogram (G, T)) for vehicle routing and scheduling. This results in an extended ASP procedure that includes time and scheduling related differential constraints.

Adjusting a solver: depending on the addition of differential constraints, appropriate adjustments to the hybrid ASP solver may be required to handle extended ASP programs that contain differential constraints. Ensuring that the solver can properly solve the problem containing the differential constraints and generate a scheduling solution that satisfies all the constraints.

S6: and solving the expanded ASP program by using a mixed ASP solver to obtain a scheduling solution S= solver (ASPprogram) meeting all constraints, and using the scheduling solution for scheduling the vehicle. More specifically, it may include:

preparing an ASP solver: a suitable hybrid ASP solver, such as clinco or DLV, is selected for solving the extended ASP program. Ensuring that the solver has been properly installed and configured and can handle ASP programs that contain differential constraints.

Building input: the expanded ASP program is provided as input to an ASP solver. This may be by saving the ASP program as a text file or passing it to the solver through a program interface.

Starting a solver: the hybrid ASP solver is operated and the expanded ASP program is passed as input to the solver. Depending on the particular solver, this may be implemented by a command line parameter or a programming interface.

The solving process comprises the following steps: the solver will attempt to find a scheduling solution that satisfies all constraints. It derives and searches through automatic reasoning and search algorithms to find variable assignments that satisfy all constraints.

Analysis results: once the solver finds the scheduling solution, it will return the results of the solution. This is typically presented in terms of assignment of variables in an ASP program. The results may be parsed to extract information about task allocation, path planning and scheduling.

Scheduling an application: the obtained scheduling solution is applied to the scheduling of the vehicle. And determining task allocation and path planning of each vehicle according to the analysis result, and executing scheduling according to constraint conditions. This may include assigning tasks to specific vehicles, planning routes and times of travel, coordinating with other traffic participants, and the like.

Example 1 below assumes a city with five intersections, denoted as vertices V1, V2, V3, V4, V5, respectively. The five intersections are connected through each road to form an urban traffic network. The network can be modeled as a weighted directed graph G in which vertices represent intersections, directed edges represent roads, and the weights of the edges represent the minimum travel time through the road segment. The weight calculation takes the length of the road, traffic rules (such as speed limit), the number of lanes, traffic signals and real-time traffic information (such as traffic jams, accidents, road maintenance and the like) into account.

Such an urban traffic network is represented as follows:

vertex: v1, V2, V3, V4, V5;

edges: (V1, V2), (V2, V3), (V3, V1), (V2, V4), (V4, V5), (V5, V1);

weight:

W(V1,V2)=10,W(V2,V3)=15,W(V3,V1)=20,W(V2,V4)=5,W(V4,V5)=8,W(V5,V1)=12；

these weights represent the minimum time (in minutes) required for a vehicle to travel from one intersection to another under ideal road conditions. These weights may be dynamically adjusted based on real-time traffic information.

Then, there are the following tasks:

task 1: vehicle1 moves from V1 to V3

Task 2: vehicle2 moves from V2 to V5

Task 3: the vehicle3 moves from V4 to V1

In encoding these tasks into the ASP program, the start and end positions of each task and the corresponding time intervals need to be considered, as well as the possible dependencies between the respective tasks. For example, if the target locations of two tasks are the same intersection, then to avoid congestion, we can set that they cannot arrive at the same time, which forms a dependency. In the case of more vehicles, we can set that only a certain number of vehicles below can reach the intersection at the same time, so as to avoid the intersection from being jammed.

The ASP is then used to encode the routing and scheduling problems. Factors that may also be considered by this encoding process include traffic flow at each intersection, route selection, adjustment of vehicle speed, etc. For example, for task1 (vehicle 1 moves from V1 to V3), there are V1- > V2- > V3 and V1- > V3 possible routes, which need to be selected to be optimal according to the current traffic conditions and road weights.

Next, the ASP program is extended using differential constraints. In this step, constraints related to time and scheduling need to be handled. I.e. the start and end positions of each task and the corresponding time intervals as described above, and also the possible dependencies between the individual tasks.

And finally, solving the expanded ASP program by using a mixed ASP solver to obtain a scheduling solution meeting all constraints. This solution will provide an explicit route and schedule for each vehicle to most efficiently accomplish its task.

Notably, this process can accept real-time data input to accommodate changes in traffic conditions. For example, if the roads of V2 through V3 are temporarily closed due to an accident, the road weights and task codes may be updated quickly and then the scheduling problem re-solved to find new routes for all affected vehicles.

Here is given an ASP code in the simplest case:

% definition intersection and road

intersection(v1). intersection(v2). intersection(v3). intersection(v4). intersection(v5).

road(v1, v2). road(v2, v3). road(v3, v1). road(v2, v4). road(v4, v5). road(v5, v1).

% definition weight

weight(v1, v2, 10). weight(v2, v3, 15). weight(v3, v1, 20).

weight(v2, v4, 5). weight(v4, v5, 8). weight(v5, v1, 12).

% definition predicate

task(T) :- member(T, [task1, task2, task3]).

vehicle(V) :- member(V, [vehicle1, vehicle2, vehicle3]).

start_loc(task1, v1).

end_loc(task1, v3).

start_loc(task2, v2).

end_loc(task2, v5).

start_loc(task3, v4).

end_loc(task3, v1).

start_time(task1, 0).

end_time(task1, 60).

start_time(task2, 10).

end_time(task2, 50).

start_time(task3, 20).

end_time(task3, 80).

% rules and constraints

:- vehicle(V), start_loc(T1, L), end_loc(T2, L), task(T1), task(T2), T1 != T2, assign_vehicle(T1, V), assign_vehicle(T2, V).

:- assign_vehicle(T1, V1), assign_vehicle(T2, V2), vehicle(V1), vehicle(V2), T1 != T2, start_loc(T1, L), end_loc(T1, L), start_loc(T2, L), end_loc(T2, L).

% definition path planning

path(V1, V2, [V1, V2]) :- road(V1, V2).

path(V1, V3, [V1|Path]) :- road(V1, V2), path(V2, V3, Path).

% route scheduling

route_schedule(V, Path) :- assign_vehicle(T, V), start_loc(T, S), end_loc(T, E), path(S, E, Path).

% solution

#show route_schedule/2.

The meaning of each part of the above code is explained line by line as follows:

interwork (v 1), interwork (v 2), interwork (v 3), interwork (v 4), interwork (v 5): predicates of intersections, i.e., intersections in cities, are defined. v1, v2, v3, v4, v5 each represent five intersections.

Head (v 1, v 2), head (v 2, v 3), head (v 3, v 1), head (v 2, v 4), head (v 4, v 5), head (v 5, v 1): predicates of roads, i.e., road connection relationships between intersections, are defined. Roads between intersections can be represented by road predicates.

weight (v 1, v2, 10), weight (v 2, v3, 15), weight (v 3, v1, 20), weight (v 2, v4, 5), weight (v 4, v5, 8), weight (v 5, v1, 12): the weight of the road, i.e. the minimum travel time through the road, is defined. The weights of different roads are represented by weight predicates weight.

task (T): -number (T, [ task1, task2, task3 ]): predicates of tasks, i.e., tasks that need to be completed, are defined. Different tasks are represented by task predicates.

Vehicle (V): -membrane (V, [ Vehicle1, vehicle2, vehicle3 ]). Predicates of the vehicle, i.e., vehicles that are available for task execution, are defined. Different vehicles are represented by a vehicle predicate.

start_loc (task 1, v 1). End_loc (task 1, v 3).: the starting and ending positions of the task, i.e. the starting and ending points of the task, are defined. The start and end positions of different tasks are represented by start_loc and end_loc predicates.

start_time (task 1, 0). End_time (task 1, 60): the start time and end time of the task, i.e. the start and stop time of the task, are defined. The start and end times of different tasks are represented by start_time and end_time predicates.

Vehicle (V), start_loc (T1, L), end_loc (T2, L), task (T1), task (T2), T1 |=t2, assignment_vehicle (T1, V), assignment_vehicle (T2, V): the constraint prohibits the same vehicle from simultaneously executing different tasks with the same starting or ending position.

Path (V1, V2, [ V1, V2 ]) road (V1, V2.): predicates for path planning are defined, and paths are generated from one intersection to another in a recursive manner.

route_schedule (V, path): -assignment_vehicle (T, V), start_loc (T, S), end_loc (T, E), path (S, E, path): predicates for route scheduling are defined, and specific routes for each vehicle are generated through task allocation and path planning.

# show route_schedule/2.). The result to be output, i.e., the result of route_schedule/2, i.e., the specific route of each vehicle is specified.

Through the codes and the solver, the solving of the vehicle route scheduling problem can be performed, and the specific route of each vehicle is output.

In the following embodiment 2, the invention is of particular interest in how to avoid traffic congestion during scheduling. This example uses the same urban traffic network model and vehicle mission as in example 1 in steps S1 and S2. However, in this example, the present invention is different in the handling of traffic congestion.

And S3, encoding the task allocation and sequencing questions by using an Answer Set Program (ASP) to generate an ASP program (T) related to the tasks. In this step, we consider traffic congestion information in addition to the basic task allocation and ordering codes. In particular, if traffic congestion is predicted to occur in some road segments, assignment of tasks to these road segments will be avoided as much as possible.

And S4, encoding the vehicle routing and scheduling problems by using an Answer Set Program (ASP) to generate an ASP program (G, T) related to routing and scheduling. In this step we also introduce a special constraint that a route should not be selected if its predicted traffic congestion level exceeds a certain threshold.

And S5, expanding the two ASP programs by using differential constraint, and processing time and scheduling related constraint to obtain an expanded ASP program. In this example, time and scheduling related constraints may take into account the impact of traffic congestion in particular. In particular, we will consider the extra time delay that traffic congestion may cause and introduce it as an extra differential constraint.

And S6, solving the expanded ASP program by using a mixed ASP solver to obtain a scheduling solution S= solver (ASPprogram) meeting all constraints. In the calculation result, due to the consideration of traffic jam, the obtained solution should be able to effectively avoid the predicted traffic jam area, thereby improving the overall traffic efficiency.

In embodiment 3, we consider another approach to avoid congestion:

to make this embodiment more specific, we assume that there is a specific traffic congestion area in the urban traffic network, where traffic congestion occurs during specific time periods, such as early peak and late peak.

(1) First, we need to model the urban traffic network as a weighted directed graph, which is the same as step S1 in example 1. In this process we need to add an extra weight to each edge of the traffic congestion area. This weight may be set according to the severity of the traffic congestion, as well as the period of the traffic congestion. Specifically, if the degree of traffic congestion of a road is high during a specific period of time, the weight of the road is high.

(2) Next, we need to define the traffic control problem as a task allocation and path planning problem, and this process is the same as step S2 in embodiment 1. In this process we need to take into account the impact of traffic congestion. In particular, if a vehicle's mission requires it to pass through a traffic congestion area, we need to do special handling for this mission. For example, this task may be broken up into multiple subtasks, each of which bypasses the traffic congestion area.

(3) Then, it is necessary to encode the task allocation and ranking questions using an Answer Set Program (ASP), which is the same as step S3 in embodiment 1. In this process, we need to define a special ASP rule for each task that ensures that the task does not cause the vehicle to pass through the traffic congestion area, or to pass through the congestion area only during off-peak hours.

(4) In performing the encoding of the vehicle routing and scheduling problem, it is necessary to encode the problem using an Answer Set Program (ASP), which is the same as step S4 in embodiment 1. In this process, we need to set a special ASP rule for each path, which can ensure that the vehicle does not pass through the traffic congestion area for a certain period of time.

(5) Next, we need to extend the above two ASP procedures using differential constraints, processing time and scheduling related constraints, this procedure being the same as step S5 in embodiment 1. In this process we need to consider the period of traffic congestion, as well as the situation of the traffic congestion area.

(6) Finally, we need to solve the expanded ASP program using a hybrid ASP solver to obtain a scheduling solution that satisfies all constraints, which is the same as step S6 in embodiment 1. In this process, the solver needs to be able to find a solution that ensures that all vehicles can complete their tasks at a predetermined time while avoiding passing through the traffic congestion area for a specific period of time.

The optimization system for traffic control of the present invention may be implemented in a server, and specifically includes:

(1) Urban traffic network modeling module: the module is realized on a hardware server, and the urban traffic network is modeled into a weighted directed graph through a preloaded software algorithm. Specifically, the server acquires urban traffic network data including intersection positions, road information, and the like by connecting to an urban traffic information center. By processing the data, the modeling module generates a weighted directed graph representing the urban traffic network.

(2) The task definition module: the module is also implemented on a hardware server, and the traffic control problem is defined as a task allocation and path planning problem through a pre-installed software algorithm. The server acquires task information including a vehicle position, a destination, and the like by connecting to the vehicle dispatching center.

(3) An ASP coding module: the module is implemented on a hardware server, and encodes task allocation and ordering questions using a pre-installed Answer Set Program (ASP) software library to generate an ASP program for tasks, and encodes vehicle routing and scheduling questions to generate an ASP program for routing and scheduling.

(4) Constraint expansion module: the module also realizes that the two ASP programs are expanded by using a preloaded differential constraint processing software library on a hardware server, and the related constraints of time and scheduling are processed to obtain an expanded ASP program.

(5) Hybrid ASP solver module: the module is realized on a hardware server, and a preloaded mixed ASP solver software library is used for solving the expanded ASP program to obtain a scheduling solution meeting all constraints.

The specific configuration of the server hardware may be selected according to the size of the urban traffic network, the complexity of the task, and the real-time processing requirements, for example, a server with a high-performance CPU, a large memory, and a high-speed network connection may be selected.

The server may be a stand-alone physical server or a cloud server. If the cloud server is selected, the computing requirements can be dynamically adjusted by utilizing the elastic resources of cloud computing so as to meet the traffic control requirements of different time periods.

Software modules may be developed using programming languages such as c++, python, etc., depending on development requirements and performance requirements. The ASP software library may be selected to be either open-source or commercial, for example, an open-source ASP parsing library such as clinco may be selected.

The various modules in the system may communicate over a network, as well as through inter-process communication on a server.

The foregoing is only a preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art, who is within the scope of the present invention, should be covered by the protection scope of the present invention by making equivalents and modifications to the technical solution and the inventive concept thereof.

Claims

1. An optimization method applied to traffic control is characterized by comprising the following steps:

2. The optimization method applied to traffic control according to claim 1, wherein the vehicle tasks include, but are not limited to, route planning from one location to another, vehicle speed adjustment, and parking spot selection.

3. The optimization method for traffic control according to claim 1, wherein the weighted directed graph modeled by the urban traffic network further comprises real-time traffic information including at least traffic congestion, accidents, road maintenance.

4. The optimization method applied to traffic control according to claim 1, wherein the process of encoding task allocation and ordering problems and vehicle routing and scheduling problems by the Answer Set Program (ASP) further comprises the input of real-time data to recalculate and update task allocation and path planning in the face of dynamic changes in the traffic network.

5. The optimization method applied to traffic control according to claim 1, wherein the time and schedule related constraints in S5 include at least: the expected start time interval and end time interval of each task, and the dependency relationship among different tasks.

6. An optimization system for traffic control, comprising:

7. The optimization system for traffic control according to claim 6, wherein the vehicle tasks defined by the task definition module include, but are not limited to, route planning from one location to another, vehicle speed adjustment, and parking spot selection.

8. The optimization system for traffic control according to claim 6, wherein said urban traffic network modeling module further comprises real-time traffic information including at least traffic congestion, accidents, road maintenance.

9. The optimization system for use in traffic control according to claim 6, wherein the encoding process of the ASP encoding module further comprises the input of real-time data to recalculate and update mission allocation and path planning in the face of dynamic changes in the traffic network.

10. The optimization system for traffic control according to claim 6, wherein the time and schedule related constraints processed by the constraint expansion module include at least: an expected start time interval and an end time interval for each task; dependency relationship between different tasks.