Efficiency and safety considered tough road network repair time sequence multi-objective optimization method

Abstract

The invention provides a tough road network repair time sequence multi-objective optimization method giving consideration to efficiency and safety. The method comprises the steps that S1, road network system travel data and road network system accident data in the road network repair period are acquired, and a road network system efficiency evaluation index and a road network system safety evaluation index are established according to the road network system travel data and the road network system accident data; s2, respectively obtaining a first road network toughness index and a second road network toughness index according to dynamic changes of the road network system efficiency evaluation index and the road network system safety evaluation index, and constructing a toughness-oriented road network repair time sequence multi-objective optimization model based on the first road network toughness index and the second road network toughness index; s3, solving the toughness-oriented road network repair time sequence multi-objective optimization model by adopting a hybrid variable neighborhood non-dominated solution sorting genetic algorithm; and S4, compiling a hybrid variable neighborhood non-dominated solution sorting genetic algorithm through MATLAB, and obtaining the solution of the toughness-oriented road network repair time sequence multi-objective optimization model to obtain a Pareto optimal solution as an optimal road network repair time sequence multi-objective optimization scheme. The method has the beneficial effects that the influence of post-disaster road network restoration efficiency and restoration activities on road network traffic safety can be considered, and rapid restoration of the road network after a rainstorm event is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of road network repair methods, and in particular to a multi-objective optimization method for resilient road network repair timing that takes into account both efficiency and safety. Background Art

[0002] Formulating effective urban road repair strategies after a rainstorm event can reduce losses, lower accident risks, and enhance road network resilience.

[0003] At present, research on resilient road network recovery strategies under rainstorm disasters focuses on improving road network efficiency, ignoring the traffic safety issues that may be induced during the road network repair process. The waterlogging and reduced traffic capacity caused by rainstorms will cause traffic congestion and even paralysis, which can easily lead to frequent traffic accidents. In the existing disaster road network repair, there is no consideration of the resilience optimization index of road network safety during the repair period, which cannot reflect the road network safety issues during the repair period after the rainstorm event and cannot help the road network to be repaired quickly. Summary of the invention

[0004] The technical problem to be solved by the present invention is to take into account the efficiency of post-disaster road network repair and the impact of repair activities on road network traffic safety, and to ensure the rapid repair of the road network after a rainstorm event. In order to overcome the defects of the above-mentioned prior art (or related technology), the present invention provides a multi-objective optimization method for resilient road network repair timing that takes into account both efficiency and safety.

[0005] The present invention provides a multi-objective optimization method for resilient road network repair timing that takes into account both efficiency and safety, comprising the following steps:

[0006] Step S1, obtaining and establishing a road network system efficiency evaluation index and a road network system safety evaluation index based on the road network system travel data and road network system accident data during the road network repair period;

[0007] Step S2, obtaining a first road network resilience index and a second road network resilience index according to dynamic changes of the road network system efficiency evaluation index and the road network system safety evaluation index, and constructing a resilience-oriented road network repair time series multi-objective optimization model based on the first road network resilience index and the second road network resilience index;

[0008] Step S3, using a hybrid variable neighborhood non-dominated solution sorting genetic algorithm to solve the resilience-oriented road network repair time series multi-objective optimization model;

[0009] Step S4, compiling the hybrid variable neighborhood non-dominated solution sorting genetic algorithm through MATLAB, obtaining the solution of the toughness-oriented road network repair timing multi-objective optimization model to obtain the Pareto optimal solution as the optimal road network repair timing multi-objective optimization solution.

[0010] Compared with the prior art, the multi-objective optimization method for resilient road network repair timing that takes into account both efficiency and safety has the following advantages:

[0011] In the present application, step S1 is used to establish a road network system efficiency evaluation index and a road network system safety evaluation index, step S2 is used to establish a resilience-oriented road network repair timing multi-objective optimization model, step S3 is used to solve the resilience-oriented road network repair timing multi-objective optimization model, and step S4 is used to solve the Pareto optimal solution. According to the road network system travel data and the road network system accident data, the road network system efficiency evaluation index and the road network system safety evaluation index are obtained and the road network resilience index is obtained by changing them. Compared with the traditional method that considers a single goal of road network system efficiency, it takes into account both the system travel efficiency and the traffic network security during the road network repair period, and takes into account the post-disaster road network repair efficiency and the impact of repair activities on road network traffic safety. Moreover, using the Pareto optimal solution as the optimal road network repair timing multi-objective optimization scheme for road network repair can ensure the rapid repair of the road network after a rainstorm event.

[0012] In a possible implementation, the road network system travel data includes the traffic volume on section a before the disaster event occurs, the traffic volume on section a on the τth day during the road network repair period, the travel time on section a before the disaster event occurs, and the travel time on section a on the τth day during the road network repair period. Then, in step S1, the road network system efficiency evaluation index is obtained according to the traffic volume on section a before the disaster event occurs, the traffic volume on section a on the τth day during the road network repair period, the travel time on section a before the disaster event, and the travel time on section a on the τth day during the road network repair period.

[0013] In a possible implementation, in step S1, the road network system efficiency evaluation index is obtained by the following calculation formula:

[0014]

[0015] in,

[0016] Q τ represents the efficiency evaluation index of the road network system;

[0017] represents the traffic volume on the road section a before the disaster event occurs;

[0018] represents the traffic volume on the road section a on the τth day during the road network repair period;

[0019] represents the travel time on road section a before the disaster event occurs;

[0020] It represents the travel time of section a on the τth day during the road network repair period.

[0021] In one possible implementation, the road network system accident data includes the traffic volume on section a before the disaster event occurs, the traffic volume on section a on the τth day during the road network repair period, the capacity of section a before the disaster event occurs, the capacity of section a on the τth day during the road network repair period, the length of section a, the first model parameter, the second model parameter and the third model parameter. Then, in step S1, the road network system safety assessment index is obtained according to the traffic volume on section a before the disaster event occurs, the traffic volume on section a on the τth day during the road network repair period, the capacity of section a before the disaster event occurs, the capacity of section a on the τth day during the road network repair period, the length of section a, the first model parameter, the second model parameter and the third model parameter.

[0022] In a possible implementation, in step S1, the road network system safety assessment index is obtained by the following calculation formula:

[0023]

[0024] in,

[0025] S τ represents the safety assessment index of the road network system;

[0026] represents the traffic volume on the road section a before the disaster event occurs;

[0027] represents the traffic volume on the road section a on the τth day during the road network repair period;

[0028] It represents the traffic capacity of the road section a before the disaster event occurs;

[0029] represents the traffic capacity of the road section a on the τth day during the road network repair period;

[0030] L a Indicates the length of the road section a;

[0031] ρ1 represents the first model parameter;

[0032] ρ2 represents the second model parameter;

[0033] ρ3 represents the second model parameter.

[0034] In a possible implementation manner, in step S2, the first road network resilience index and the second road network resilience index are obtained by the following calculation formula:

[0035]

[0036] in

[0037] R1 represents the first road network resilience index;

[0038] R2 represents the second road network resilience index;

[0039] τ e Indicates the time when the road network repair is completed;

[0040] τ1 represents the start time of road network repair;

[0041] Q(τ) represents the efficiency evaluation index of the road network system on the τth day;

[0042] S(τ) represents the road network system safety evaluation index on the τth day.

[0043] In a possible implementation, in step S2, the toughness-oriented road network repair time sequence multi-objective optimization model is constructed by the following calculation formula:

[0044]

[0045]

[0046] in,

[0047] O represents the optimization objective of road network resilience;

[0048] represents the equilibrium traffic flow of section a on the τth day;

[0049] represents the road capacity of section a at the τth day;

[0050] R1 represents the first road network resilience index;

[0051] R2 represents the second road network resilience index;

[0052] γ a,τ Indicates the speed of road section repair;

[0053] Indicates the unit repair speed of the road section;

[0054] represents a 0-1 variable, which takes the value 1 if construction is carried out on section a on the τth day, and takes the value 0 otherwise;

[0055] represents a 0-1 variable, which takes the value 1 if the maintenance team h performs construction on the road section a on the τth day, and takes the value 0 otherwise.

[0056] In one possible implementation, in step S3, the solution process of the hybrid variable neighborhood non-dominated solution sorting genetic algorithm includes A1 chromosome (repair timing) encoding, A2 random generation of the initial population, A3 non-dominated solution sorting, A4 selection operation and variable neighborhood search, A5 generation of a new population, A6 non-dominated solution sorting, A7 selection operation, A8 deep learning variable neighborhood search, and A9 judgment on whether the termination condition is met: if so, go to step S4, if not, repeat A5-A8 until the termination condition is met. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 is a flow chart of the steps of the present invention;

[0058] Figure 2 It is a flow chart of the hybrid variable neighborhood non-dominated solution sorting genetic algorithm of the present invention;

[0059] Figure 3 This is a schematic diagram of the variable neighborhood operator of the present invention. DETAILED DESCRIPTION

[0060] First, those skilled in the art should understand that these implementations are only used to explain the technical principles of the embodiments of the present application, and are not intended to limit the protection scope of the embodiments of the present application. Those skilled in the art can make adjustments to them as needed to adapt to specific application scenarios.

[0061] The present application is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0062] See also Figure 1 The present application embodiment discloses a multi-objective optimization method for resilient road network repair timing that takes into account both efficiency and safety, including:

[0063] Step S1, obtaining and establishing a road network system efficiency evaluation index and a road network system safety evaluation index based on the road network system travel data and road network system accident data during the road network repair period;

[0064] Step S2, obtaining a first road network resilience index and a second road network resilience index according to dynamic changes of a road network system efficiency evaluation index and a road network system safety evaluation index, and constructing a resilience-oriented road network repair time series multi-objective optimization model based on the first road network resilience index and the second road network resilience index;

[0065] Step S3, using a hybrid variable neighborhood non-dominated solution sorting genetic algorithm to solve a multi-objective optimization model for the time series of toughness-oriented road network repair;

[0066] Step S4, compile the hybrid variable neighborhood non-dominated solution sorting genetic algorithm through MATLAB, obtain the solution of the resilience-oriented road network repair timing multi-objective optimization model, and obtain the Pareto optimal solution as the optimal road network repair timing multi-objective optimization solution.

[0067] Continue to see Figure 1 In step S1, when calculating the evaluation index of the road network system efficiency, the travel time of the road section is first calculated, and then the travel time of all road sections in the road network is summed to obtain the system travel time. The efficiency of the road network system on the τth day during the repair period can be expressed as the ratio of the system travel time on the τth day to the system travel time when it is not damaged;

[0068] The calculation formula of the road network system efficiency evaluation index is as follows:

[0069]

[0070] Where: TTT o Indicates the system travel time when the road network is not damaged; TTT τ It represents the travel time of the road network system on the τth day during the repair period; represents the traffic volume on road section a before the disaster event occurs; represents the traffic volume on section a on the τth day during the road network repair period; It represents the travel time on road section a before the disaster event occurs; represents the travel time of section a on the τth day during the road network repair period, and

[0071]

[0072] In the formula, represents the travel time of the road section under the free flow state; α and β represent model parameters, α=0.15, β=4; represents the traffic capacity of the road section on the τth day, and

[0073]

[0074] In the formula, represents the traffic capacity of section a when it has not suffered any loss; k a Represents the remaining capacity ratio of section a after the rainstorm event; γ a,τ represents the repair speed on the road section a on the τth day; represents the unit repair speed of maintenance team h (i.e., the workload of a single maintenance team in one day); represents a 0-1 variable, which takes the value 1 if construction is carried out on section a on the τth day, and takes the value 0 otherwise; is a 0-1 variable, which takes the value of 1 if the maintenance team h performs construction on the road section a on the τth day, and 0 otherwise.

[0075] Continue to see Figure 1 In step S1, when calculating the safety evaluation index of the road network system, the number of traffic accidents per unit time of the road section is first calculated, and then the number of traffic accidents of all road sections in the road network is summed to obtain the number of traffic accidents of the road network system. The road network system safety index on the τth day during the repair period can be expressed as the ratio of the number of system traffic accidents on the τth day to the number of system traffic accidents when the system is not damaged;

[0076] The calculation formula for road network system safety assessment index is as follows:

[0077]

[0078] Where: P o represents the number of traffic accidents when the road network is not damaged; P τ represents the number of traffic accidents on the τth day during the road network repair period; represents the traffic capacity of section a before the disaster occurs; L a represents the length of section a; ρ1, ρ2 and ρ3 represent model parameters, ρ1=358.6, ρ2=-407.7, ρ3=175.3.

[0079] Continue to see Figure 1 In step S2, establishing a resilient road network optimization target means taking the road network system efficiency resilience and the road network system safety resilience as the optimization targets. When calculating the road network system efficiency resilience and the road network system safety resilience, they can be obtained by integrating the road network efficiency index and the road network system safety index in the repair stage over time;

[0080] The multi-objective calculation formula of road network resilience is as follows:

[0081]

[0082] Where, τ1 represents the start time of road network repair; τ e Indicates the time when the road network repair is completed.

[0083] Continue to see Figure 1 , in step S2, establishing a multi-objective optimization model for resilient road network repair timing that takes into account both efficiency and safety refers to establishing a multi-objective optimization model that integrates the efficiency resilience of the road network system and the safety resilience of the road network system;

[0084] The calculation formula of the multi-objective optimization model for the time series of toughness-oriented road network repair is as follows:

[0085]

[0086] Among them, O represents the optimization objective of the resilient road network; represents the equilibrium traffic flow on day τ during the repair period, and

[0087]

[0088] Where: It represents a 0-1 variable, indicating the relationship between the road segment and the path. If the OD pair (trip origin and destination) w path flow passes through the road segment a, it takes 1, otherwise it takes 0; represents the flow of OD on path r of w on day τ during the repair period; It represents the total flow of OD on path w on day τ during the repair period.

[0089] Continue to see Figure 1 In step S3, compared with single-objective optimization, the characteristic of multi-objective optimization is to pursue the optimization of each objective at the same time to achieve a comprehensive optimal state. The challenge faced by multi-objective optimization is that there are often conflicts between the objectives, that is, when one objective reaches the optimal, the performance of other objectives may decrease significantly. One strategy to solve the multi-objective optimization problem is to convert it into a single-objective problem, but this method is prone to introduce errors in the conversion process. In order to avoid these errors and directly obtain the optimal solution, a hybrid variable neighborhood non-dominated solution sorting genetic algorithm can be used to solve the multi-objective optimization model;

[0090] Hybrid variable neighborhood non-dominated solution sorting genetic algorithm solution, refer to Figure 2 and Figure 3 , the specific process is as follows:

[0091] (1) Chromosome encoding

[0092] Integer encoding is used. A single gene (integer) in the chromosome represents the road section number. The order of chromosome genes from left to right represents the order of road section repair. The chromosome encoding can be expressed as:

[0093] X=(l1,l2,……,l N )

[0094] Where: X represents chromosome, each chromosome represents a set of road section repair timing plans; N represents the number of damaged road sections;

[0095] (2) Generate the first generation of subgroups

[0096] Randomly generate chromosomes of size P size The initial population is as follows:

[0097]

[0098] Where: X s represents the sth chromosome; unit represents the initial individual;

[0099] (3) Calculate the fitness value and sort the non-dominated solutions

[0100] The fitness value can be obtained by calculating the objective value of the optimization function. By sorting non-dominated solutions, the genetic algorithm can effectively distinguish different good and bad solutions in multi-objective optimization, ensuring that the population converges to the Pareto frontier and maintains diversity. First, for each solution p in the population, calculate the number of solutions φ(p) dominated by solution p and the number of solutions φ(p) that dominate solution p. Secondly, determine the first frontier and classify all solutions with φ(p) = 0 into the first frontier F1, that is, these solutions are not dominated by any other solutions. Finally, recursively calculate other frontiers. For each solution p∈F1, update the number of solutions q dominated by it. if Then put q into the next frontier F2; repeat this process until all solutions are assigned to the frontier;

[0101] (4) Select an operation

[0102] Through tournament simulation competition, the individual with the highest fitness is selected from a number of randomly selected individuals as the winner and added to the next generation; first, random sampling is performed, that is, a number of individuals are randomly selected from the current population, called the tournament group; second, fitness is compared, that is, the fitness value of each individual is calculated, and the higher the fitness value, the better the individual; then the winner is selected, that is, according to the fitness value, the individual with the highest fitness is selected from the tournament group as the winner; finally, the above process is repeated until 1 / 2 of the individuals of the required population size are selected;

[0103] (5) Deep Learning Variable Neighborhood Search

[0104] Using the long short-term memory network model principle in deep learning, the relationship between chromosomes and neighborhood structures under different update generations is analyzed to match the chromosomes with the best domain operation operator (see Figure 3 ), the learning process mainly includes: forgetting process, input process, and output process. The calculation formula is as follows:

[0105] Forgetting process: Assume that the number of generations executed by the current algorithm is Gen and the input is The hidden state at the previous moment is h Gen-1 , the state of the memory cell at the previous moment is C Gen-1 , then the forgetting process calculation formula is expressed as:

[0106] f Gen =σ(W f ·[h Gen-1 ,x Gen ])+b f )

[0107] Where: Chro represents chromosome; FitVal represents chromosome fitness value; f Genrepresents the output of the forgetting process, with a value range of [0,1]. The closer to 1, the more retained. σ represents the Sigmoid activation function, which is used to generate values ​​in the range of [0,1]. W f Represents the weight matrix of the forgetting process; b f represents the bias of the forgetting process;

[0108] The input process calculation formula can be expressed as:

[0109]

[0110] i Gen =σ(W i ·[h Gen-1 ,x Gen ])+b i )

[0111]

[0112] Where: Represents candidate information; tanh represents the hyperbolic tangent activation function, which is used to generate values ​​in the range of [-1,1]; i Gen Indicates input process output; W C , W i Represents the weight matrix of the input process; b C , b i Represents the bias of the input process; C Gen Represents the updated memory unit;

[0113] The output process calculation formula can be expressed as:

[0114] o Gen =σ(W o ·[h Gen-1 ,x Gen ])+b o )

[0115] h Gen-1 =o Gen tanh(C Gen )

[0116] In the formula: o Gen Represents output information, that is W o Represents the weight matrix of the output process; b o Represents the bias of the output process; h Gen-1 Represents the output hidden state; Neibo represents the neighborhood action, that is, the chromosome neighborhood transformation operation;

[0117] (6) Update individual units and generate new populations;

[0118] (7) Determine whether the termination condition is met and repeat (3) to (6) until the termination condition is met.

[0119] In the description of the present application, the description with reference to the terms "one embodiment", "some embodiments", "in the present embodiment", "specific example", or "some examples" etc. means that the specific features, mechanisms, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, mechanisms, materials or characteristics described may be combined in a suitable manner in any one or more embodiments or examples. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0120] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A multi-objective optimization method for resilient road network repair timing that takes into account both efficiency and safety, characterized in that: The following steps are involved: Step S1, obtaining and establishing a road network system efficiency evaluation index and a road network system safety evaluation index based on the road network system travel data and road network system accident data during the road network repair period; Step S2, obtaining a first road network resilience index and a second road network resilience index according to dynamic changes of the road network system efficiency evaluation index and the road network system safety evaluation index, and constructing a resilience-oriented road network repair time series multi-objective optimization model based on the first road network resilience index and the second road network resilience index; Step S3, using a hybrid variable neighborhood non-dominated solution sorting genetic algorithm to solve the resilience-oriented road network repair time series multi-objective optimization model; Step S4, compiling the hybrid variable neighborhood non-dominated solution sorting genetic algorithm through MATLAB, obtaining the solution of the toughness-oriented road network repair timing multi-objective optimization model to obtain the Pareto optimal solution as the optimal road network repair timing multi-objective optimization solution.

2. The multi-objective optimization method for resilient road network repair timing according to claim 1 is characterized in that: The travel data of the road network system includes the traffic volume on the road section a before the disaster event, the traffic volume on the road section a on the τth day during the road network repair period, the travel time on the road section a before the disaster event, and the travel time on the road section a on the τth day during the road network repair period. Then, in step S1, the road network system efficiency evaluation index is obtained according to the traffic volume on the road section a before the disaster event, the traffic volume on the road section a on the τth day during the road network repair period, the travel time on the road section a before the disaster event, and the travel time on the road section a on the τth day during the road network repair period.

3. The multi-objective optimization method for resilient road network repair timing according to claim 2 is characterized in that: In step S1, the road network system efficiency evaluation index is obtained by the following calculation formula: in, Q τ represents the efficiency evaluation index of the road network system; represents the traffic volume on the road section a before the disaster event occurs; represents the traffic volume on the road section a on the τth day during the road network repair period; represents the travel time on road section a before the disaster event occurs; It represents the travel time of section a on the τth day during the road network repair period.

4. The multi-objective optimization method for resilient road network repair timing according to claim 1 is characterized in that: The road network system accident data includes the traffic volume on section a before the disaster event occurs, the traffic volume on section a on the τth day during the road network repair period, the capacity of section a before the disaster event occurs, the capacity of section a on the τth day during the road network repair period, the length of section a, the first model parameter, the second model parameter and the third model parameter. In step S1, the road network system safety assessment index is obtained according to the traffic volume on section a before the disaster event occurs, the traffic volume on section a on the τth day during the road network repair period, the capacity of section a before the disaster event occurs, the capacity of section a on the τth day during the road network repair period, the length of section a, the first model parameter, the second model parameter and the third model parameter.

5. The multi-objective optimization method for resilient road network repair timing according to claim 4 is characterized in that: In step S1, the road network system safety assessment index is obtained by the following calculation formula: in, S τ represents the safety assessment index of the road network system; represents the traffic volume on the road section a before the disaster event occurs; represents the traffic volume on the road section a on the τth day during the road network repair period; It represents the traffic capacity of the road section a before the disaster event occurs; represents the traffic capacity of the road section a on the τth day during the road network repair period; L a Indicates the length of the road section a; ρ1 represents the first model parameter; ρ2 represents the second model parameter; ρ3 represents the second model parameter.

6. The multi-objective optimization method for resilient road network repair timing according to claim 1 is characterized in that: In step S2, the first road network resilience index and the second road network resilience index are obtained by the following calculation formula: in R1 represents the first road network resilience index; R2 represents the second road network resilience index; τ e Indicates the time when the road network repair is completed; τ1 represents the start time of road network repair; Q(τ) represents the efficiency evaluation index of the road network system on the τth day; S(τ) represents the road network system safety evaluation index on the τth day.

7. The multi-objective optimization method for resilient road network repair timing according to claim 1 is characterized in that: In step S2, the multi-objective optimization model for the time sequence of toughness-oriented road network repair is constructed by the following calculation formula: in, O represents the optimization objective of road network resilience; represents the equilibrium traffic flow of section a on the τth day; represents the road capacity of section a at the τth day; R1 represents the first road network resilience index; R2 represents the second road network resilience index; γ a,τ Indicates the speed of road section repair; Indicates the unit repair speed of the road section; represents a 0-1 variable, which takes the value 1 if construction is carried out on section a on the τth day, and takes the value 0 otherwise; represents a 0-1 variable, which takes the value 1 if the maintenance team h performs construction on the road section a on the τth day, and takes the value 0 otherwise.

8. The multi-objective optimization method for resilient road network repair timing according to claim 1 is characterized in that: In the step S3, the solution process of the hybrid variable neighborhood non-dominated solution sorting genetic algorithm includes A1 chromosome (repair timing) encoding, A2 random generation of initial population, A3 non-dominated solution sorting, A4 selection operation and variable neighborhood search, A5 generation of new population, A6 non-dominated solution sorting, A7 selection operation, A8 deep learning variable neighborhood search, A9 judgment on whether the termination condition is met: if so, go to step S4, if not, repeat A5-A8 until the termination condition is met.

Images