Intersection geometric design and signal control collaborative optimization method based on simulation optimization

Abstract

The invention discloses an intersection geometric design and signal control collaborative optimization method based on simulation optimization, and the method comprises the steps: building an intersection optimization model and a simulation model, enabling the input of the intersection optimization model to be intersection traffic demands and geometric conditions, and enabling the output of the intersection optimization model to be intersection geometric design and signal control related decision variables; the output of the simulation model is an operation evaluation index of the corresponding scheme; wherein the simulation model automatically adjusts a decision variable of an intersection design scheme into a geometric structure and a signal timing parameter in the simulation model, simulation is operated to obtain an optimized target value, the optimized target value serves as an adaptive value of a particle swarm optimization algorithm to update the decision variable of the current scheme, and the updated variable value is returned to the intersection optimization model. And scheme transmission and optimization are realized. According to the method, the diversity of simulation output results is fully exerted, the optimization target can be freely replaced on the basis of not modifying the upper layer model, and the flexibility and optimization efficiency of the intersection optimization model are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent city road control, and in particular to a method for collaborative optimization of intersection geometry design and signal control based on simulation optimization. Background Art

[0002] The plane intersection is the bottleneck node of urban road traffic. In order to improve the utilization efficiency of the time and space resources of the intersection, an unconventional intersection using "left turn by using the lane" has been proposed. However, the current design method mainly follows the design method of the general plane intersection. The pre-stop line and pre-signal unique to the left turn intersection by using the lane are mainly selected based on experience. There is no targeted optimization setting method, and no invention patents for such methods have been retrieved.

[0003] As a key node of urban roads, the optimization design of planar intersections has become an important part of improving the efficiency of urban road traffic and meeting the development needs of modern urban traffic. In order to improve the adaptability of intersection optimization design methods to real-world scenarios and improve the flexibility and scalability of the model, a collaborative optimization model of intersection geometry design and signal control based on simulation optimization was proposed. However, the current intersection optimization method focuses on the analysis of intersection traffic operation characteristics and single-element design, which is limited in practical application and the optimization effect is difficult to guarantee. Therefore, in the face of complex real-world operating environments and diverse design modes, we embed the simulation model into the optimization framework to break the limitations of traditional optimization methods. In addition, no invention patents for this method have been retrieved so far.

[0004] After searching the literature on the prior art, it is found that the relevant methods for intersection optimization mainly include the following aspects:

[0005] 1. Intersection optimization method based on logical process. Specifically including intersection geometry layout, lane function division, signal phase sequence, cycle duration, green light time allocation for each flow direction, etc., which are optimized step by step according to the process. Representative works include "Urban Traffic Control" and "Traffic Management and Control".

[0006] 2. Model-based intersection optimization method. By establishing a mathematical programming model, the intersection is optimized based on the analytical expression of the operation evaluation index (optimization objective function). Representative methods include the invention patents "Pre-stop line and pre-signal setting method for left-turn intersection" (ZL201610164786.X), "Continuous flow intersection left-turn non-motor vehicle traffic design method" (ZL201810155206.X), etc.

[0007] 3. Intersection optimization method based on simulation. Based on the geometric layout of the intersection, a geometric model is constructed in the simulation, and then the intersection design is optimized through multi-scheme comparison. Representative works include "Traffic Application Software" and "Research on Dynamic Cooperative Optimization of Variable Guide Lanes and Signal Timing at Intersections".

[0008] Method 1 is a conventional design method, which currently has relatively mature technical results, but it is difficult to obtain the optimal solution.

[0009] Method 2 relies on the analytical expression of the operation evaluation index (optimization objective function), but the calculation model of the intersection evaluation index is closely related to the design mode and control method. Different geometric scenarios, different optimization methods, and different index calculation models are different, making accurate modeling and solving difficult.

[0010] Method 3 uses a simulation model to calculate and evaluate operating indicators, but the existing method does not dynamically modify the intersection geometry parameters in the simulation, resulting in only multiple schemes for intersection geometry layout being compared, making it difficult to obtain the optimal scheme. Summary of the invention

[0011] In view of the shortcomings existing in the prior art, the purpose of the present invention is to provide a method for collaborative optimization of intersection geometry design and signal control based on simulation optimization. By directly outputting the operation evaluation index of the intersection through simulation, there is no need to remodel it and construct an analytical expression of the evaluation index when facing different intersection scenarios, thus avoiding the problem of inaccurate model construction and optimizing the effect. Only by outputting the diversity of optimization targets through simulation, the intersection geometry and signal control are collaboratively optimized for different optimization targets. In order to achieve the above-mentioned purpose and other advantages according to the present invention, a method for collaborative optimization of intersection geometry design and signal control based on simulation optimization is provided, comprising:

[0012] Building an intersection optimization model and a simulation model, wherein the intersection optimization model inputs the intersection traffic demand and geometric conditions, and outputs the intersection geometry design and signal control-related decision variables;

[0013] The input of the simulation model is the parameters of the signal control related decision variables, and the output is the operation evaluation index of the corresponding scheme;

[0014] Among them, the simulation model automatically adjusts the decision variables of the intersection design scheme to the geometric structure and signal timing parameters in the simulation model, runs the simulation to obtain the optimization target value, and uses this value as the fitness value of the particle swarm optimization algorithm to update the decision variables of the current scheme, and returns the updated variable values ​​to the intersection optimization model, thereby realizing scheme transfer and optimization.

[0015] Preferably, the intersection optimization model includes an objective function, which is obtained through simulation, and any measurable operating indicators and combinations thereof can be used as the objective function, including queue length, carbon emissions and delays.

[0016] Preferably, the intersection optimization model includes constraints, and the constraints include geometric design constraints and signal timing constraints;

[0017] The geometric design constraints include total number of lanes constraint, minimum number of lanes constraint, prevention of internal conflict of import traffic flow, minimum number of flow directions constraint and exit lane number constraint;

[0018] The signal timing constraints include cycle duration constraints, green light start time constraints, minimum green light duration constraints, lane signal setting constraints, optional signal phase constraints, green light display order constraints and clearing time constraints.

[0019] Preferably, the decision variables related to signal control include the number of import lanes, variables characterizing the functional division of lanes, signal cycle duration, the start time and duration of the green light for each direction; the decision variables related to signal control are written into the simulation model in real time through Python, the optimization target obtained by simulation calculation is passed into the algorithm as the fitness value, the algorithm updates the variables, and the interaction between the intersection optimization model and the simulation model is realized.

[0020] Preferably, the simulation model building includes building a simulation basic model, and the specific process is as follows:

[0021] Build the road geometry design of the intersection in SUMO, input the number of entrance and exit lanes and the entrance lane flow;

[0022] Enter the traffic flow in each direction of the intersection and the signal control strategy, including phase sequence, green light start time, duration and signal cycle length;

[0023] Set the relevant parameters of SUMO simulation operation, including saturation flow rate, lane width, headway and simulation time step.

[0024] Preferably, each element of the simulation basic model is written into a code file, which can be read and written by Python, and the simulation file can be modified in real time. After the simulation basic model is built, it is connected with the optimization algorithm, as follows:

[0025] (1) Run the simulation basic model and calculate the optimization index value.

[0026] (2) The index value obtained by simulation is connected with the optimization algorithm and fed back to the optimization algorithm as the adaptation value.

[0027] (3) The optimization algorithm updates the corresponding geometric and signal timing parameters, which are read and written into the SUMO simulation file using the Python tool, that is, transmitted to the SUMO simulation model.

[0028] (4) Repeat (1) to (3) until the algorithm converges or the maximum number of iterations is reached.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] 1. The present invention provides a collaborative optimization method for intersection geometry design and signal control based on simulation optimization.

[0031] 2. The method of the present invention breaks away from the reliance of traditional optimization methods on the operational evaluation analytical model. The simulation output results are used to replace the traditional analytical objectives, so that the model can adapt to various traffic designs and control modes and is suitable for diverse traffic environments.

[0032] 3. The simulation-based optimization framework proposed in this invention can flexibly optimize intersections under different traffic policy guidance. It simplifies the intersection optimization modeling process under different geometric conditions, quickly realizes the comparison of optimization schemes, and supports real-time dynamic optimization in complex traffic scenarios, providing strong support for future multi-mode collaborative optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 A flowchart of a method for collaborative optimization of intersection geometry design and signal control based on simulation optimization according to the present invention;

[0034] Figure 2 A schematic diagram of conflicts to be avoided within the import traffic flow according to the intersection geometry design and signal control collaborative optimization method based on simulation optimization according to the present invention;

[0035] Figure 3 A phase sequence diagram of a case scenario of a collaborative optimization method for intersection geometry design and signal control based on simulation optimization according to the present invention;

[0036] Figure 4 It is a simulation-based optimization scheme diagram of the intersection geometry design and signal control collaborative optimization method based on simulation optimization according to the present invention;

[0037] Figure 5 The original operation scheme diagram of the intersection according to the intersection geometry design based on simulation optimization and the signal control collaborative optimization method of the present invention;

[0038] Figure 6 It is a result diagram of the optimal lane configuration for minimum delay according to the intersection geometry design and signal control collaborative optimization method based on simulation optimization according to the present invention;

[0039] Figure 7 It is a result diagram of the optimal lane configuration for minimum carbon emission according to the intersection geometry design and signal control collaborative optimization method based on simulation optimization according to the present invention;

[0040] Figure 8 This is a result diagram of the optimal lane configuration for minimum average queue length according to the intersection geometry design and signal control collaborative optimization method based on simulation optimization according to the present invention. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] Reference Figure 1-8 , a method for collaborative optimization of intersection geometry design and signal control based on simulation optimization, comprising: building an intersection optimization model and a simulation model, wherein the intersection optimization model inputs traffic demand and geometry conditions of the intersection, and outputs decision variables related to intersection geometry design and signal control;

[0043] The input of the simulation model is the parameters of the signal control related decision variables, and the output is the operation evaluation index of the corresponding scheme;

[0044] Among them, the simulation model automatically adjusts the decision variables of the intersection design scheme to the geometric structure and signal timing parameters in the simulation model, runs the simulation to obtain the optimization target value, and uses this value as the fitness value of the particle swarm optimization algorithm to update the decision variables of the current scheme, and returns the updated variable values ​​to the intersection optimization model, thereby realizing scheme transfer and optimization.

[0045] Table 1. Relevant parameters in the model

[0046]

[0047]

[0048] Objective Function

[0049] The objective function value of this model is obtained through simulation, so the selection of the objective function is flexible, and any measurable operating indicators and their combinations can be used as the objective function. It can be delay, queue length, carbon emissions, etc. This section takes delay as an example, as shown in formula (1).

[0050] Min D(1) where D represents the average total vehicle delay (s) during the entire simulation period.

[0051] Constraints

[0052] For a conventional intersection (four-way intersection), its geometric design and signal control should meet the following constraints.

[0053] Geometric Design Constraints

[0054] (1) Constraint on the total number of lanes. The sum of the entrance and exit roads in each direction is equal to the total number of lanes, as shown in formula (2).

[0055]

[0056] (2) Minimum number of lanes constraint. Each direction should have at least one entrance lane and one exit lane, as shown in equations (3) and (4).

[0057]

[0058] (3) Constraints to prevent internal conflicts in import traffic. For two adjacent import lanes k (left side) and k+1 (right side) from direction i, if the k+1 lane is allowed to pass in the direction w, then the k lane should be prohibited from passing in the direction m, m∈(w+1,…,3), to eliminate internal cross conflicts in the import lanes. Figure 2 As shown, if Δ 1,2,2 =1, the adjacent lane on the left should be prohibited from moving in other directions, that is, Δ 1,2,1 =0, the above constraint can be expressed by formula (5).

[0059]

[0060] (4) Minimum flow direction constraint: To ensure that each lane plays a role, traffic should be allowed to flow in at least one direction, as shown in formula (6).

[0061]

[0062] (5) Constraint on the number of exit lanes. The number of exit lanes in each direction should not be less than the number of entrance lanes allowed to enter the direction from other directions, so as to prevent unreasonable traffic merging, as shown in formula (7).

[0063]

[0064] 2.2 Signal Timing Constraints

[0065] (1) Cycle duration constraint. The signal cycle length of the intersection should be controlled within a reasonable range, as shown in formula (8).

[0066] Cmin ≤C≤C max (8)

[0067] (2) Green light start time constraint. All green light start times should be set within the green light cycle duration range, as shown in formula (9).

[0068]

[0069] (3) Minimum green light duration constraint. The green light duration of each flow direction must meet the minimum green light duration requirement, as shown in formula (10).

[0070]

[0071] (4) Lane signal setting constraints. In actual signal control strategies, if a shared lane marking is set to allow an import lane to perform two or more turning actions, then in order to avoid ambiguity, the related turning actions are usually controlled by the same signal light, as shown in equations (11) and (12).

[0072]

[0073] Consider lane k from entrance lane i, if a turn w is allowed in this lane, where M is an arbitrarily large positive number. If lane k from i allows a turn w, then Δ i,w,k =1, so both sides of the above two inequalities become zero. This means that the constraint requires Θ i,k =θ i,w and That is, for all turns w allowed on lane k, the same signal setting is used. However, if it is a single lane, except for the one turn w, no other turns are allowed, then the Δ i,w′,k = 0, the constraint is invalid. The lane only uses the same signal setting as the w turn allowed in the lane.

[0074] (5) Optional signal phase constraints. In actual operation, the relative time of the start and end of different phases needs to be set to ensure the practicality of intersection signal timing. They are shown in equations (13) and (14) respectively.

[0075]

[0076] With the above constraints, when setting or When the signal timing is used, it can be ensured that the two phases start or end at the same time. For example, in the subsequent cases, in order to ensure the practicality of the signal timing and compare with the previous optimization cases, the signal timing adopts a double-loop structure, setting the straight and right turn phases respectively. and Make the start and end times of straights and right turns equal.

[0077] (6) Green light display order constraint. For any set of conflicting flow directions, the signal phase sequence is represented by the successor function, as shown in formula (15).

[0078]

[0079] (7) Clearing time constraint. For safety reasons, the interval between the start and end time of the green light for any set of conflicting flows should at least meet the clearing time requirement, as shown in formula (16).

[0080]

[0081] Parameter passing

[0082] The SUMO simulation modeling file is directly called by programming language to modify its statements. The text characters of the simulation modeling file are directly operated by Python language to realize the parameter assignment of the algorithm update. Therefore, under the condition that the simulation plan is clear, the automatic assignment of simulation model parameters can be realized according to the specific plan. At the same time, the SUMO simulation platform is equipped with a COM interface, which can communicate with the Python programming language through protocol. After the parameter values ​​of the simulation model are assigned, the COM interface program can be loaded to realize the automatic operation of the simulation program.

[0083] The model decision variables include the number of import lanes α i , variable Δ representing lane function division i,w,k , signal cycle length C, green light start time for each flow direction θ i,w and duration These decision variables are written into the SUMO simulation model in real time through Python. The optimization target obtained by simulation calculation is passed into the algorithm as the fitness value. The algorithm updates the variables to realize the interaction between the intersection optimization model and the simulation model. The specific details are as follows Figure 1 In addition, the specific method of writing variables into the simulation is as follows:

[0084] (1) Geometric design parameters. These parameters need to be input when configuring the road network. The road network file of an intersection in SUMO includes nodes, edges, and connections (connections from the entry lane to the exit lane). This paper considers a typical four-way intersection, so the node settings are fixed. The edge and connection settings are related to the parameters of the model. The edges include the number of entry lanes in each direction, α i and the number of exit lanes ε i , it needs to be written into the SUMO file; the connection is with the variable Δ i,w,k If the model outputs Δ i,w,k = 1, then write that there is a connection from lane k in the i direction to lane w, if Δi,w,k =0, the road network file will not be written.

[0085] (2) Signal control parameters. The signal timing parameter input in the simulation only requires the green light start time θ of each lane. i,k and duration Φ i,k In each round of updating, the geometric design variables Δ i,w,k Update, conflict flow collection Ψ s Update, conflicting flow signal phase sequence P iw,i′w′ Formula (15) must be satisfied, and the green light start time θ for all directions i,w and duration The remaining constraints must be satisfied, and the green light start time of the lane is Θ i,k and duration Φ i,k Constraints (11) and (12) must be satisfied. Each iteration will update the i,k and Φ i,k , signal cycle length C, and the yellow light g automatically assigned to each phase y Time and all red time g r , and write them into the signal control file.

[0086] (3) Traffic flow parameter configuration. Traffic flow parameters are configured according to the traffic flow of the vehicle turning from direction i to direction w.

[0087] Each time the algorithm is updated, a new decision variable is generated. According to the above method, the Python file reading and writing function is used to write the decision into the SUMO simulation, run the simulation to obtain the simulation result, and calculate the fitness value based on it and pass it into the optimization algorithm. The optimal solution is obtained by continuous iteration. The acquisition of simulation results is specifically realized through the data acquisition tool TraCI (Traffic Control Interface). TraCI is connected to the running SUMO simulation, and query commands are sent to obtain real-time vehicle position and speed information. Taking the calculation of the average vehicle delay in formula (1) as an example, the average vehicle delay of the simulation step can be calculated by obtaining the vehicle speed of each simulation step through formula (17), and then the average vehicle delay of the required simulation period can be obtained through formula (18). Simulation steps s1 and s2 are flexibly set. Since there will be a preparatory stage at the beginning of the simulation, the length of the preparatory stage of different road networks is different. Therefore, after the simulation runs stably, the start simulation step s1 and the end simulation step s2 of the set evaluation index to be collected are set according to the needs. If the evaluation index of the entire simulation period needs to be collected, the simulation steps s1 and s2 can be set to the entire simulation period.

[0088]

[0089] Simulation model construction

[0090] The micro-simulation model of the corresponding scheme can adopt any simulation model that can modify the relevant parameters of the intersection in real time. The micro-simulation system is composed of multiple sub-models, including the following model, the lane change model, the lateral motion model, etc. The parameters of each sub-model jointly affect the interaction between vehicles, control the operation behavior of individual vehicles, and deduce the vehicle action process in complex traffic systems. Since the SUMO platform can encode the geometric design and signal control parameters of the intersection, it has strong scalability and can continuously iterate and optimize the intersection design scheme by modifying these parameter codes in real time. Therefore, this paper uses the SUMO simulation platform to simulate traffic operation under given geometric and signal control conditions and calculate evaluation indicators. The steps to establish a basic intersection simulation model in SUMO are as follows:

[0091] (1) Build the road geometry design of the intersection in SUMO and input the number of entrance and exit lanes and the flow direction of the entrance lanes.

[0092] (2) Input the traffic flow in each direction of the intersection and the signal control strategy, including the phase sequence, green light start time, duration, and signal cycle duration.

[0093] (3) Set the relevant parameters of SUMO simulation operation, including saturation flow rate, lane width, headway, and simulation time step.

[0094] Each element of the above simulation basic model is written into a code file, which can be read and written by Python, and the simulation file can be modified in real time. After the simulation basic model is built, it is connected with the optimization algorithm. The specific steps are as follows:

[0095] (1) Run the above intersection simulation and calculate the optimization index value.

[0096] (2) The index value obtained by simulation is connected with the optimization algorithm and fed back to the optimization algorithm as the adaptation value.

[0097] (3) The optimization algorithm updates the corresponding geometric and signal timing parameters, which are read and written into the SUMO simulation file using the Python tool, that is, transmitted to the SUMO simulation model.

[0098] (4) Repeat (1) to (3) until the algorithm converges or the maximum number of iterations is reached.

[0099] 5. Solution method

[0100] The particle swarm optimization algorithm (PSO) is used to solve the above model, and the evaluation index value obtained by simulation is used as the fitness value of the particle swarm optimization algorithm. The best solution is obtained through iterative optimization. Therefore, the optimization process is a process of continuously adjusting the geometric design and signal timing to achieve the convergence of the optimization target value through simulation combined with the optimization algorithm. The specific process is as follows:

[0101] Step 1: Generate the initial solution. Determine the initial geometric design and initial signal timing of the intersection. The initial geometric design divides the lane functions according to the given number of import lanes based on the traffic flow ratio. The initial signal timing uses the traditional Webster timing method to perform phase design and signal timing calculations on the intersection. Set the number of particles m and the initial position of each particle. and initial velocity Initialize the iteration index T = 0, set the individual's historical optimal value pBest as the current position Select the best individual in the group as the current global optimal value gBest.

[0102] Step 2: Calculate the fitness value. Input the time-space design plan of the intersection into the simulation software SUMO to obtain the specific optimization index value. In each iteration, calculate the fitness value

[0103] Step 3: Solution update. Based on the particle swarm optimization algorithm, the solution is updated. In each iteration, the corresponding variables are updated by comparing the fitness values, and the solution is taken as the current optimal solution. If the individual's current fitness value is better than the optimal value, Update the individual optimal value If the current fitness value of the group is better than the optimal value of the previous iteration Then update the global optimal value The velocity and position of each particle c are updated by equations (19) and (20), respectively.

[0104]

[0105] Step 4: Terminate the iteration. If the iteration index of particle c reaches the maximum value T max (i.e., T ≥ T max ), then the particle completes the search. Otherwise, update the iteration index of the particle T = T + 1, and go to step 2 to update its position in the next iteration. If all particles complete the search or the fitness value results converge, the solution corresponding to the global best known position of the particle swarm is taken as the optimal solution. The following is the pseudo code of the above algorithm process.

[0106]

[0107]

[0108] Embodiment 1:

[0109] 1. Input Data

[0110] In order to further verify the optimization performance and applicability of this method, field data is used to further illustrate the work of the present invention. The intersection of Gaoke West Road and Qi'ai Road in Shanghai is selected. The traffic volume data and geometric data are shown in Table 2.

[0111] Table 2 Traffic volume and geometry data of the intersection of Gaokexi Road and Qiai Road in Shanghai

[0112]

[0113] 2. Optimization steps

[0114] Step 1: Initialize the particle swarm. Each particle represents a possible design of an intersection, including geometric design (such as the number of lanes, lane channelization, etc.) and signal timing scheme (such as the green light start time and green light duration of each phase). Set the size of the particle swarm, that is, how many particles are required to participate in the optimization process. The initial position of each particle represents an intersection design and signal timing scheme, while the initial velocity is randomly generated, indicating the direction and step size of the particle's search in the solution space.

[0115] Step 2: Particle swarm iterative optimization. According to the rules of the particle swarm optimization algorithm, update the speed and position of each particle. The speed and position of each particle are updated by equations (19) and (20) respectively.

[0116] Step 3: Fitness evaluation. The fitness function is used to evaluate the pros and cons of the intersection design scheme represented by each particle. The fitness evaluation index in this case is delay. In the subsequent comparison of different optimization target results, the fitness function value is carbon emissions and average queue length. When a new geometric design and signal timing scheme are generated after the particle update, it is automatically input into the simulation system to calculate the fitness value of each particle at the current position. If the fitness value of a particle's new position is better than its historical optimal value, then update the particle's historical optimal value. If it is better than the global optimal value, then update the global optimal solution.

[0117] Step 4: Constraint check. Each time the particle position is updated, it is necessary to check whether the solution meets the design constraints in the actual project, such as lane width, signal phase duration, etc. Only solutions that meet these constraints are considered feasible solutions. If the position of a particle (i.e., the intersection design solution) does not meet the constraints, the particle may be repositioned.

[0118] Step 5: Update the optimal solution. In each round of iteration, the system records the historical optimal solution of each particle (local optimal solution) and the best solution among all particles (global optimal solution). As the number of iterations increases, the particle swarm gradually converges to the global optimal solution. When the maximum number of iterations (T max ) or when , the algorithm believes that the optimal solution has been found and the iteration ends.

[0119] Step 6: Optimal result output: Based on the fitness evaluation of the simulation feedback, the optimal solution for geometric design and signal timing is determined.

[0120] 3. Optimization results and effect comparison

[0121] Since the traffic volume in the east-west direction of this intersection is low, the left-turn permission phase is adopted in the east-west direction. The left-turn and through-traffic volume in the north-south direction is large, so the left-turn protection phase is adopted in the north-south direction. At the same time, the right-turn traffic volume at the east entrance is much higher than the through-traffic and left-turn traffic, so right-turning vehicles in this direction can pass in advance with left-turning vehicles in the north-south direction. Figure 3 Design a phase plan for this intersection.

[0122] The geometric design and signal timing scheme of the intersection are optimized by simulation-based methods. Figure 4 As shown in the figure, the current geometric design and signal timing scheme of the intersection obtained through traffic survey is as follows Figure 5 The delay obtained by simulation-based optimization is 27.59s, while the result of the original design of the intersection input into the simulation is 29.16s, which is a 5.7% reduction in delay.

[0123] 4. Comparison of optimization results for different objectives

[0124] The model proposed by the method can be easily expanded to an optimization model with other indicators (such as carbon emissions and average queue length) as the target, and there is no need to establish an analytical calculation model for the evaluation indicators. The following will expand the above case and use different optimization objectives to optimize intersections with different geometric conditions.

[0125] To better verify the applicability of the model, considering the road red line, there are a maximum of 6 lanes in the north-south direction and a maximum of 4 lanes in the east-west direction; therefore, the number of lanes in the north-south direction is set to 4 to 6, as cases 5 to 7, respectively, and the number of lanes in the east-west direction remains unchanged. The signal control method is consistent with Section 4.2.2. The intersection geometry design and signal control optimization results are shown in Figure 2. Figure 6 , 7 , 8 and Table 3. For comparison, Table 3 lists the other indicator values ​​corresponding to each case when a certain indicator is taken as the optimal target.

[0126] Table 3 Intersection geometry design and signal control optimization results

[0127]

[0128]

[0129] The number of devices and processing scales described here are used to simplify the description of the present invention, and the application, modification and variation of the present invention will be obvious to those skilled in the art.

[0130] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and implementation modes. They can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and the illustrations shown and described herein.

Claims

1. A method for collaborative optimization of intersection geometry design and signal control based on simulation optimization, characterized in that: include: Building an intersection optimization model and a simulation model, wherein the intersection optimization model inputs the intersection traffic demand and geometric conditions, and outputs the intersection geometry design and signal control related decision variables; The input of the simulation model is the parameters of the signal control related decision variables, and the output is the operation evaluation index of the corresponding scheme; Among them, the simulation model automatically adjusts the decision variables of the intersection design scheme to the geometric structure and signal timing parameters in the simulation model, runs the simulation to obtain the optimization target value, and uses this value as the fitness value of the particle swarm optimization algorithm to update the decision variables of the current scheme, and returns the updated variable values ​​to the intersection optimization model, thereby realizing scheme transfer and optimization.

2. The method for collaborative optimization of intersection geometry design and signal control based on simulation optimization as claimed in claim 1, characterized in that: The intersection optimization model includes an objective function, which is obtained through simulation, and any measurable operating indicators and their combinations can be used as the objective function, including queue length, carbon emissions and delays.

3. The method for collaborative optimization of intersection geometry design and signal control based on simulation optimization as claimed in claim 1, characterized in that: The intersection optimization model includes constraints, which include geometric design constraints and signal timing constraints; The geometric design constraints include total number of lanes constraint, minimum number of lanes constraint, prevention of internal conflict of import traffic flow, minimum number of flow directions constraint and exit lane number constraint; The signal timing constraints include cycle duration constraints, green light start time constraints, minimum green light duration constraints, lane signal setting constraints, optional signal phase constraints, green light display order constraints and clearing time constraints.

4. The method for collaborative optimization of intersection geometry design and signal control based on simulation optimization as claimed in claim 2, characterized in that: The decision variables related to signal control include the number of entrance lanes, variables representing the functional division of lanes, signal cycle duration, the start time and duration of green lights for each direction; the decision variables related to signal control are written into the simulation model in real time through Python, and the optimization target obtained by simulation calculation is passed into the algorithm as the fitness value. The algorithm updates the variables to realize the interaction between the intersection optimization model and the simulation model.

5. The method for collaborative optimization of intersection geometry design and signal control based on simulation optimization as claimed in claim 1, characterized in that: The simulation model construction includes the construction of a simulation basic model, and the specific process is as follows: Build the road geometry design of the intersection in SUMO, input the number of entrance and exit lanes and the entrance lane flow; Enter the traffic flow in each direction of the intersection and the signal control strategy, including phase sequence, green light start time, duration and signal cycle length; Set the relevant parameters of SUMO simulation operation, including saturation flow rate, lane width, headway and simulation time step.

6. The method for collaborative optimization of intersection geometry design and signal control based on simulation optimization as claimed in claim 5, characterized in that: Write each element of the simulation basic model into a code file, which can be read and written by Python, and the simulation file can be modified in real time. After the simulation basic model is built, it is connected to the optimization algorithm, as follows: (1) Run the simulation basic model and calculate the optimization index value; (2) Connect the index value obtained by simulation with the optimization algorithm and feed it back to the optimization algorithm as the fitness value; (3) The optimization algorithm updates the corresponding geometric and signal timing parameters, which are read and written into the SUMO simulation file using Python tools, that is, transmitted to the SUMO simulation model; (4) Repeat (1) to (3) until the algorithm converges or the maximum number of iterations is reached.

Images