Multi-objective optimization method for site selection of pavement ice and snow condition monitoring station

Abstract

The present application belongs to the field of construction, and particularly relates to a multi-objective optimization site selection method for a pavement ice and snow condition monitoring station. The existing site selection method for the pavement ice and snow condition monitoring station has the problems of strong subjectivity and lack of scientific basis, thereby leading to an unreasonable site selection scheme. The present application is proposed, which comprises the following steps: Step 1: collecting road historical data; Step 2: establishing a set of road traffic safety risk points under ice and snow conditions according to the road historical data collected in Step 1; Step 3: constructing a multi-objective optimization site selection model according to the road historical data obtained in Step 1 and the set of road traffic safety risk points under ice and snow conditions obtained in Step 2; Step 4: constructing constraint conditions of the multi-objective optimization site selection model; Step 5: generating an initial population of an improved NSGAII algorithm according to the constraint conditions of the multi-objective optimization site selection model; and then solving the multi-objective optimization site selection model by using the NSGAII algorithm, and outputting a site selection result of the pavement ice and snow condition monitoring station.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of construction, and particularly relates to a multi-objective optimization site selection method for a pavement ice and snow condition monitoring station. BACKGROUND

[0002] In most areas of China, low temperature, precipitation and other meteorological factors often cause road snow and ice in winter, especially in special road sections such as shady places, tunnel entrances and other road sections where dark ice is not easy to detect, which seriously restricts the safe and efficient operation of road facilities. Therefore, many scholars have developed a pavement ice and snow condition monitoring system, which can collect meteorological data such as temperature, humidity, visibility and road surface water ice and snow thickness, video images and other road condition information, providing a data basis for pavement ice and snow condition evaluation, and having a wide application scenario in the field of roads where ice and snow disasters frequently occur.

[0003] In recent years, highway traffic meteorological monitoring station construction technical specifications have been introduced in Shandong, Hebei, Zhejiang and more than a dozen provinces and cities. These specifications put forward corresponding technical requirements for the equipment selection, monitoring station site selection, data collection and processing, operation and maintenance of the monitoring station, but for the site selection problem of the monitoring station, certain requirements and suggestions are put forward according to engineering experience, lacking scientific research support. This increases the blindness of the initial construction and the later operation to some extent, greatly affecting the actual application effect of the technology; at the same time, the research on the influencing factors of the site selection of the monitoring station is relatively scarce at the present stage, and the calculation of the construction cost and the monitoring effect lacks scientificity and systematicness, which also restricts the engineering practical application of the technology to some extent.

[0004] In summary, the existing pavement ice and snow condition monitoring station site selection method has strong subjectivity, lacks scientific basis, and thus leads to an unreasonable site selection scheme, resulting in a high traffic accident rate. SUMMARY

[0005] The purpose of the present application is to solve the problem of the existing pavement ice and snow condition monitoring station site selection method, which has strong subjectivity, lacks scientific basis, and thus leads to an unreasonable site selection scheme, resulting in a high traffic accident rate. We propose a multi-objective optimization site selection method for a pavement ice and snow condition monitoring station. It includes:

[0006] Step 1: Collecting road historical data;

[0007] Step 2: Establishing a set of road traffic safety risk points under ice and snow conditions according to the historical data collected in step 1;

[0008] Step 3: Building a multi-objective optimization site selection model according to the road historical data obtained in step 1 and the set of road traffic safety risk points under ice and snow conditions obtained in step 2;

[0009] Step 4: Building constraint conditions for the multi-objective optimization site selection model;

[0010] Step 5: generate the initial population of the improved NSGAII algorithm according to the constraint conditions of the multi-objective optimization site selection model; then solve the multi-objective optimization site selection model by using the NSGAII algorithm, and output the site selection result of the pavement ice and snow condition monitoring station.

[0011] The road historical data in step 1 include: road mileage post number data, road alignment data, road special terrain and object data, road bridge structure data, road traffic accident-prone point data under ice and snow conditions, road snow and ice-prone point data, road fog-prone point data, road existing meteorological observation station point data, and road power supply network point data;

[0012] The road historical data in step 1 are collected; the specific process is as follows:

[0013] Step 11: collect the design file of the road, and obtain the road mileage post number data, road alignment data, road special terrain and object point data, and road bridge structure data according to the design file of the road;

[0014] The road special terrain and object point data include: road sharp bend and steep slope point data, and road bridge and tunnel point data;

[0015] Step 12: collect the road infrastructure construction data provided by the traffic investment department; according to the road infrastructure construction data provided by the traffic investment department, determine the road existing meteorological observation station point data and the road power supply network point data;

[0016] Step 13: collect the road historical traffic accident investigation data; obtain the road traffic accident-prone point information under ice and snow conditions according to the road historical traffic accident investigation data in the past 10 years;

[0017] The road traffic accident-prone point under ice and snow conditions is a place where the number of road traffic accidents under ice and snow conditions exceeds 5 or more in the past 10 years;

[0018] Step 14: determine the meteorological observation stations within a straight-line distance of 20 km along the road according to the road existing meteorological observation station point data;

[0019] Collect the historical meteorological data observed by the meteorological observation stations within a straight-line distance of 20 km along the road;

[0020] Obtain the road meteorological disaster-prone point data according to the historical meteorological data observed by the meteorological observation stations within a straight-line distance of 20 km along the road in the past 20 years; the road meteorological disaster-prone point data include: road snow and ice-prone point data, and road fog-prone point data,

[0021] The road weather disaster prone position is a position where the number of weather disasters in the past 20 years is more than 5 or more;

[0022] The road special terrain feature point and the road weather disaster prone position are taken as the road ice and snow disaster prone points.

[0023] The road traffic safety risk point set under the ice and snow condition in the step 2 comprises a road traffic safety 1-level risk point and a road traffic safety 2-level risk point.

[0024] The specific process for establishing the road traffic safety risk point set under the ice and snow condition according to the historical data collected in the step 1 in the step 2 comprises the following steps.

[0025] The road traffic accident-prone point under the ice and snow condition is taken as the road traffic safety 1-level risk point.

[0026] The road ice and snow disaster prone position is taken as the road traffic safety 2-level risk point.

[0027] The present application has the following advantages.

[0028] (1) The four types of site selection influencing factors of people, vehicles, roads and environment are introduced in the pavement ice and snow condition monitoring station layout problem, and the objectivity and representativeness of site selection are effectively improved.

[0029] The site selection and construction of the existing pavement ice and snow condition monitoring station are mostly determined by the experience of decision makers, and there are problems such as strong subjectivity and lack of scientific basis. The present application fully utilizes the road alignment, terrain, traffic, meteorological and infrastructure data, establishes a road traffic safety risk point set under the ice and snow condition, clearly defines important points that need to be monitored, effectively improves the objectivity and representativeness of site selection, and helps to clearly define the ice and snow condition information in the whole monitoring section and enhance the disaster response capability.

[0030] (2) The three major targets of total cost, traffic volume and monitoring effect are comprehensively considered, and the trade-off and optimal solution set is obtained, realizing the overall and maximization of target benefits.

[0031] The four types of site selection influencing factors of people, vehicles, roads and environment are introduced in the pavement ice and snow condition monitoring station layout problem, and the objectivity and representativeness of site selection are effectively improved.

[0032] (3) In the NSGAII algorithm, the initialization population method is improved, which greatly improves the site selection efficiency and rationality.

[0033] In the multi-objective optimization model of pavement ice and snow condition monitoring station site selection, the existing initialization population method has great randomness, slow solving speed and poor convergence; by limiting the minimum distance between adjacent stations, the diversity and rationality of the initialization population are improved, thereby effectively improving the algorithm efficiency.

[0034] (4) Based on the improved NSGAII algorithm, a variety of site selection scheme sets are provided to adapt to different decision makers' preferences, which significantly improves the application range of the method; through the number and distance constraints of the stations, the method ensures good results in different road application scenarios.

[0035] Based on the improved NSGAII algorithm, the randomly generated massive site selection schemes are selected and iterated, and an objective and diverse site selection scheme set is obtained; based on the coordinates, colors and point sizes, the objective function value of each solution in the Pareto solution set is displayed, and the multi-dimensional visualization output of the optimal site selection scheme is completed, and the decision maker can select the corresponding optimal scheme according to different preferences; the introduction of the minimum distance and the maximum number of two constraint conditions makes the method maintain good solving effect when dealing with different cases, has broad application range and good application prospect. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 The overall logic diagram of the multi-objective optimization model of pavement ice and snow condition monitoring station site selection;

[0037] Figure 2 The site selection diagram of the initialization population before the algorithm improvement;

[0038] Figure 3 The site selection diagram of the initialization population after the algorithm improvement;

[0039] Figure 4 Flowchart of NSGAII algorithm;

[0040] Figure 5 Visualization result diagram of the Pareto solution set;

[0041] Figure 6 Site selection scheme display diagram under the preference of decision maker X;

[0042] Figure 7 Site selection scheme display diagram under the preference of decision maker Y. DETAILED DESCRIPTION

[0043] Specific implementation one: combined Figure 1 The present application is described, including:

[0044] Step 1: Collect road historical data;

[0045] Step 2: Establish a set of road traffic safety risk points under ice and snow conditions according to the road historical data collected in Step 1;

[0046] Step 3: Construct a multi-objective optimization site selection model according to the road historical data obtained in Step 1 and the set of road traffic safety risk points under ice and snow conditions obtained in Step 2;

[0047] Step 4: Construct the constraint conditions of the multi-objective optimization site selection model;

[0048] Step 5: Generate an improved initial population of NSGAII algorithm according to the constraint conditions of the multi-objective optimization site selection model; then solve the multi-objective optimization site selection model using NSGAII algorithm, and output the site selection results of the road ice and snow condition monitoring station.

[0049] Embodiment two: The difference between this embodiment and the specific embodiment one is that,

[0050] The road historical data in Step 1 includes: road mileage post number data, road alignment data, road special terrain and object data, road bridge structure data, road traffic accident-prone point data under ice and snow conditions, road snow and ice-prone point data, road fog-prone point data, road existing weather observation point data, and road power supply network point data;

[0051] The road historical data in Step 1 is collected; the specific process is as follows:

[0052] Step 11: Collect the road design file, and obtain road mileage post number data, road alignment data, road special terrain and object point data, and road bridge structure data according to the road design file;

[0053] The road special terrain and object point data includes: road sharp bend and steep slope point data, and road bridge and tunnel point data;

[0054] Step 12: Collect road infrastructure construction data provided by the transportation investment department; according to the road infrastructure construction data provided by the transportation investment department, determine the road existing weather observation point data and the road power supply network point data;

[0055] Step 13: Collect road historical traffic accident investigation data; obtain road traffic accident-prone point information under ice and snow conditions according to road historical traffic accident investigation data in the past 10 years;

[0056] The road traffic accident-prone point under ice and snow conditions is a place where the number of road traffic accidents under ice and snow conditions exceeds 5 or more in the past 10 years;

[0057] Step 14: determining the meteorological observation stations within a linear distance of 20 km along the road according to the existing meteorological observation station data of the road;

[0058] Collecting the historical meteorological data observed by the meteorological observation stations within a linear distance of 20 km along the road;

[0059] According to the historical meteorological data observed by the meteorological observation stations within a linear distance of 20 km along the road in the past 20 years, the road meteorological disaster prone point data is obtained, and the road meteorological disaster prone point data includes: road snow and ice prone point data, road fog prone point data,

[0060] The road meteorological disaster prone point is a place where the number of meteorological disasters in the past 20 years is more than 5 or more;

[0061] Step 15: taking the road special terrain feature point and the road meteorological disaster prone point as the road ice and snow disaster prone point; other steps and parameters are the same as in the first embodiment.

[0062] The third embodiment is different from the first embodiment in that,

[0063] The road traffic safety risk point set under the ice and snow conditions in step 2 includes: road traffic safety level 1 risk point and road traffic safety level 2 risk point;

[0064] The specific process of establishing the road traffic safety risk point set under the ice and snow conditions according to the historical data collected in step 1 in step 2 is:

[0065] The road traffic accident-prone point under the ice and snow conditions is taken as the road traffic safety level 1 risk point under the ice and snow conditions;

[0066] The road ice and snow disaster prone point is taken as the road traffic safety level 2 risk point under the ice and snow conditions;

[0067] The present application introduces four types of site selection influencing factors of people, vehicles, roads and environment in the problem of pavement ice and snow condition monitoring station layout, which effectively improves the objectivity and representativeness of site selection.

[0068] The site selection and construction of the existing pavement ice and snow condition monitoring station are mostly determined by the experience of decision makers, and there are problems such as strong subjectivity and lack of scientific basis. The present application makes full use of road alignment, terrain, traffic, meteorological and infrastructure data, establishes a road traffic safety risk point set under the ice and snow conditions, clearly defines important points that need to be monitored, effectively improves the objectivity and representativeness of site selection, and helps to clearly define the ice and snow condition information in the entire monitoring section, and enhances the disaster response ability. Other steps and parameters are the same as in the second embodiment.

[0069] Specific implementation four: the difference between this embodiment and specific implementation one is that,

[0070] The step 3 is to construct a multi-objective optimization site selection model according to the road history data obtained in step 1 and the set of road traffic safety risk points under ice and snow conditions obtained in step 2; the specific process is:

[0071] The set of road traffic safety risk points under ice and snow conditions is taken as a node of the multi-objective optimization site selection model;

[0072] The total cost, the monitoring coverage index and the traffic volume are taken as three objective functions of the multi-objective optimization site selection model;

[0073] The objective function of the total cost in the multi-objective optimization site selection model is expressed by a formula as follows:

[0074]

[0075] Wherein, min represents minimization, Cost represents the total cost of constructing the pavement ice and snow condition monitoring station, including the fixed construction cost of the monitoring station (including the equipment cost and the construction cost) and the cost of erecting the cable and construction between the station and the power supply point; c i is the fixed construction cost of constructing the pavement ice and snow condition monitoring station at node i; x i is a 0-1 variable, which represents that if node i is selected to be constructed as the pavement ice and snow condition monitoring station, then x i = 1, otherwise 0, i = 1, 2, …, I; I represents the total number of nodes; G n represents the milepost number of the node n selected to be constructed as the pavement ice and snow condition monitoring station, n = 1, 2, …, N; N represents the total number of nodes selected to be constructed as the pavement ice and snow condition monitoring station, B j represents the milepost number of the power supply point j, j = 1, 2, …, J; J represents the total number of power supply points;

[0076] The monitoring coverage index in the multi-objective optimization site selection model includes the monitoring coverage index of the 1st risk point and the monitoring coverage index of the 2nd risk point;

[0077] The objective function of the monitoring coverage index in the multi-objective optimization site selection model is expressed by a formula as follows:

[0078] The monitoring coverage index of the 1st risk point:

[0079]

[0080] The monitoring coverage index of the 2nd risk point:

[0081]

[0082] Wherein, MCI1 represents the monitoring coverage ratio of all pavement ice and snow condition monitoring stations for road traffic safety 1 level risk point, MCI2 represents the monitoring coverage ratio of all pavement ice and snow condition monitoring stations for road traffic safety 2 level risk point; y k is a 0-1 variable, indicating that if at least one pavement ice and snow condition monitoring station monitors the traffic safety risk point k, then y k =1, otherwise 0,

[0083] When there is a pavement ice and snow condition monitoring station within 2 kilometers of the road traffic safety risk point under ice and snow conditions, it is considered that the point is monitored and covered; k=1, 2, …, K; K represents the total number of road traffic safety risk points; S is the total number of ice and snow conditions under which road traffic accidents occur frequently; P is the total number of road ice and snow disaster prone points;

[0084] The target function of traffic volume in the multi-objective optimization site selection model is represented by the formula:

[0085]

[0086] Wherein, WMADT is the average winter monthly average daily traffic volume of the road section monitored by all pavement ice and snow condition monitoring stations; WMADT i represents the winter monthly average daily traffic volume of the road section where node i is located; N is the total number of nodes selected to be constructed as pavement ice and snow condition monitoring stations; other steps and parameters are the same as in the third embodiment.

[0087] The fifth embodiment is different from the first embodiment in that,

[0088] The constraint conditions of the multi-objective optimization site selection model in step 4 include site quantity constraint and site distance constraint,

[0089] The constraint conditions of the multi-objective optimization site selection model are constructed; the specific process is:

[0090] The site quantity constraint is represented by the formula:

[0091]

[0092] Wherein, x i is a 0-1 variable, indicating that if node i is selected as a site, then x i =1, otherwise 0, i=1, 2, …, I; Q is the maximum number of pavement ice and snow condition monitoring stations allowed to be constructed in this construction period;

[0093] The site distance constraint is represented by the formula:

[0094]

[0095] wherein, represents arbitrary selection one, d n represents the distance between the pavement ice and snow condition monitoring station n1 and the pavement ice and snow condition monitoring station n2, n1 1,2,..., N-1; n2 1,2,..., N-1; D represents the allowable minimum distance between two adjacent pavement ice and snow condition monitoring stations.

[0096] The application comprehensively considers three targets of total cost, traffic volume and monitoring effect, obtains a trade-off and optimal solution set, and realizes comprehensive and maximized target benefits.

[0097] The four types of site selection influencing factors of people, vehicles, roads and environment are quantified to establish a road traffic safety risk point set under ice and snow conditions; three target functions of minimum total cost, highest monitoring coverage index and maximum traffic volume are comprehensively considered, a trade-off and optimal solution set under the multi-objective function is obtained through the non-dominated sorting and crowding degree comparison mechanism of the NSGAII algorithm, comprehensive and maximized target benefits are realized, and the three types of targets are closely related to engineering practice, which provides multi-aspect evaluation for the site selection scheme and has good application effect in actual engineering.

[0098] The other steps and parameters are the same as those in the fourth embodiment.

[0099] The sixth embodiment is different from the first embodiment in that,

[0100] The initial population of the improved NSGAII algorithm is generated according to the constraint conditions of the multi-objective optimization site selection model in step 5; then the multi-objective optimization site selection model is solved by using the NSGAII algorithm, and the site selection result of the pavement ice and snow condition monitoring station is output; the specific process is as follows:

[0101] Step 51, the initial population P of the improved NSGAII algorithm is generated according to the constraint conditions of the multi-objective optimization site selection model;

[0102] Step 52, the individuals of the initial population P of the improved NSGAII algorithm are subjected to non-dominated sorting, and the individuals of the initial population P of the improved NSGAII algorithm are divided into Z frontiers according to the results of the non-dominated sorting, Z is a positive integer; the individuals in each frontier are assigned a level according to the rank of the frontier sorting;

[0103] All individuals in the first frontier are assigned a level of 1, individuals in the second frontier are assigned a level of 2, and individuals in the Zth frontier are assigned a level of Z;

[0104] Then the crowding degree in each frontier is calculated;

[0105] The frontier is similar to a classification group, which is an academic term in a multi-objective optimization algorithm and is a meaning known to those skilled in the art;

[0106] Step 53: Use the initial population P obtained from the improved NSGAII algorithm as the parent population, and perform the three basic operations of selection, crossover, and mutation on the parent population in sequence to obtain the offspring population.

[0107] The parent population in the t-th iteration is represented by P. t The offspring population in the t-th iteration is denoted as Qt;

[0108] Step 54: Merge the parent and offspring populations to generate a merged population.

[0109] The parent population for the next iteration is generated based on the merged population;

[0110] Step 55: Repeat steps 52 to 54 until the maximum number of iterations is reached or a satisfactory population is obtained, then stop iterating and obtain the final population.

[0111] Step 56: Use the individuals in the final population as the site selection results for pavement ice and snow condition monitoring stations. The initialization method of the population in the NSGAII algorithm has been improved, which greatly improves the site selection efficiency and the rationality of the site selection.

[0112] In the multi-objective optimization model for the selection of pavement ice and snow condition monitoring stations, the existing initialization population method has extremely high randomness, resulting in slow solution speed and poor convergence. By limiting the minimum distance between adjacent stations, the diversity and rationality of the initialization population are improved, thereby effectively improving the algorithm efficiency.

[0113] The other steps and parameters are the same as in Specific Implementation Method 5.

[0114] Specific Implementation Method Seven: The difference between this implementation method and Specific Implementation Method One is that...

[0115] The specific process of generating the initial population P improved by the NSGAII algorithm in step 51 based on the constraints of the multi-objective optimization location model is as follows:

[0116] Step 511: Set the initial population size N * Crossover probability nMu, mutation probability, and maximum number of iterations M;

[0117] Step 512: Set initial population generation constraints according to the constraints of the multi-objective optimization location model;

[0118] Step 513: Based on the initial population generation constraints in Step 512, randomly generate N. * The population consists of individuals, resulting in an initial population P improved by the NSGAII algorithm;

[0119] N *For 200, the cross probability nMu is 65%, the maximum number of iterations is 150, and N * individuals are randomly generated as the initial population P

[0120] In the multi-objective optimization model for site selection of pavement ice and snow condition monitoring stations, any node can be selected as a station, and the initialization population is prone to randomness, slow solving speed, and poor convergence, as shown in Figure 2 ; At the same time, such a population is eliminated only after non-dominated sorting and calculation of crowding degree because it does not meet the model constraint conditions, which greatly slows down the algorithm efficiency. Therefore, when generating the initialization population, two limiting conditions of “minimum distance between stations” and “maximum number of stations per hundred kilometers” are introduced to improve the diversity and rationality of the initialization population, thereby effectively improving the algorithm efficiency, as shown in Figure 3 ; The improvement principle and step 4 are consistent with the station number constraint and station distance constraint. Other steps and parameters are the same as in embodiment six.

[0121] Embodiment eight: The difference between this embodiment and embodiment one is that,

[0122] The individuals in the initial population P improved by the NSGAII algorithm in step 52 are non-dominated sorted, and the individuals in the initial population P improved by the NSGAII algorithm are divided into Z fronts according to the non-dominated sorting results, Z being a positive integer; the individuals in each front are assigned a level according to the ranking of the front sorting; and then the crowding degree in each front is calculated; the specific process is as follows:

[0123] Step 521: Calculate the objective function value of each individual in the initial population P; the i-th individual in the initial population P is denoted as pi

[0124] Step 522: According to the objective function values of all individuals obtained and the dominance relationship in the NSGAII algorithm, all individuals are divided into Z front layers,

[0125] The individuals in the first front layer are individuals that are not dominated by any individual, and the individuals in the next front layer are individuals dominated by the individuals in the previous front layer;

[0126] For example, the individuals in the second front are individuals dominated by the individuals in the first front, and so on, to finally obtain Z fronts; the dominance relationship between the individuals is determined according to the dominance relationship in the NSGAII algorithm;

[0127] Step 523: According to the objective function values of the individuals in each front and the crowding degree calculation method in the NSGAII algorithm, the crowding degree of the individuals in each front layer is calculated, the individuals on each objective are sorted, and the distance between adjacent individuals is calculated, to obtain the crowding degree value of each individual in each front layer.

[0128] Other steps and parameters are the same as Embodiment Seven.

[0129] Embodiment Nine: The difference between this embodiment and Embodiment One is that,

[0130] The step 53 sequentially performs selection, crossover and mutation on the parent population to obtain the offspring population; the specific process is as follows:

[0131] Step 531: According to the non-dominated level of the parent population individuals and the crowding degree in each front, two parent individuals are selected from the parent population,

[0132] The selection method of the parent individuals in the step 531 according to the non-dominated level and the crowding degree includes any one of the following methods: tournament selection, roulette wheel selection, ranking selection and the like;

[0133] Step 532: According to the set crossover probability nMu, the selected parent individuals are subjected to crossover operation to generate offspring individuals,

[0134] The crossover method includes any one of the following methods: simulated binary crossover, uniform crossover, multi-point crossover and the like;

[0135] Step 533: According to the set mutation probability, all the generated offspring individuals are subjected to mutation operation to obtain the offspring population;

[0136] The mutation method can use any one of the following methods: polynomial mutation, displacement mutation, non-uniform mutation and the like.

[0137] Other steps and parameters are the same as Embodiment Eight.

[0138] Embodiment Ten: The difference between this embodiment and Embodiment Nine is that

[0139] The step 54 combines the parent population P t and the offspring population Q t of the tth iteration to generate the combined population R t of the tth iteration;

[0140] The parent population R t of the t+1th iteration is generated according to the combined population R t+1 of the tth iteration; the specific process is as follows:

[0141] Step 541: The parent population P t of the tth iteration and the offspring population Q t of the tth iteration are combined to generate the combined population R t of the tth iteration; which is expressed by the formula as follows:

[0142] Rt=Pt∪Qt, denotes ∪ union

[0143] The parent has N * individuals, the offspring has N * individuals, and the combined population contains 2N * individuals.

[0144] Step 542: Non-dominated sorting is performed on the individuals of the tth iteration of the combined population R t , and r frontiers are obtained, and the individuals in each frontier are assigned a rank according to the rank of the frontier sorting;

[0145] All individuals in the first frontier are assigned a rank of 1, individuals in the second frontier are assigned a rank of 2, and individuals in the rth frontier are assigned a rank of r;

[0146] The individuals will be divided into multiple non-dominated layers (frontiers). The individuals in the first frontier are the best, the individuals in the second frontier are slightly worse, and so on;

[0147] Step 543: According to the rank of the individual, N * individuals are selected to form the population P t+1 of the t+1th iteration.

[0148] The present application preferentially selects individuals with lower ranks, that is, all individuals in the first frontier are selected first, then individuals in the second frontier are selected, and so on, until N * individuals are selected to form the population P t+1 of the next generation.

[0149] The present application is based on the improved NSGAII algorithm, and provides a diverse set of site selection schemes to adapt to the preferences of different decision makers, significantly improving the application range of the method; through the number and distance constraints of the stations, it is ensured that the method has good effect in different road application scenarios.

[0150] Based on the improved NSGAII algorithm, the randomly generated massive site selection schemes are selected and iterated to obtain an objective and diverse set of site selection schemes; based on coordinates, colors and point sizes, the objective function value of each solution in the Pareto solution set is displayed, and multi-dimensional visualization output of the optimal site selection scheme is completed, and the decision maker can select the corresponding optimal scheme according to different preferences; the introduction of the minimum distance and the maximum number of two constraint conditions makes the method maintain good solving effect when dealing with different cases, has a wide application range and good application prospect

[0151] See Figure 5The Pareto solution set of the optimal site selection scheme is shown, the target function value of each solution in the Pareto solution set is displayed based on coordinates, colors and point sizes, the multi-dimensional visual output of the optimal site selection scheme is completed, and the optimal site selection scheme is selected by the decision maker according to different target preferences, for example, in the Suiman highway Harbin-Yabuli section in the embodiment:

[0152] If the decision maker X requires that the first-level monitoring coverage index is greater than or equal to 90%, the second-level monitoring coverage index is greater than or equal to 80%, the total cost is minimum, and the traffic volume is maximum, the optimal scheme is shown in Figure 6 The total cost of the solution scheme is 104.1 million yuan, and the traffic volume is 295,000 vehicles / day.

[0153] If the decision maker Y requires that the first-level monitoring coverage index is greater than or equal to 80%, the second-level monitoring coverage index is greater than or equal to 65%, the total cost is minimum, and the traffic volume is maximum, the optimal scheme is shown in Figure 7 The total cost of the solution scheme is 84.1 million yuan, and the traffic volume is 306,000 vehicles / day. Other steps and parameters are the same as those in the ninth embodiment.

[0154] The above is only the preferred embodiment of the present application, and it should be understood that the present application is not limited to the above specific embodiments. Although the present application has been disclosed as above with reference to the preferred embodiments, it is not intended to limit the present application, and any person skilled in the art can make some changes or modifications to the above disclosed technical content without departing from the scope of the technical solution of the present application, and any simple modification, equivalent replacement and improvement of the above embodiments within the scope of the technical solution of the present application, according to the technical essence of the present application, within the spirit and principles of the present application, are still within the protection scope of the present application.

Claims

1. A multi-objective optimization method for site selection of pavement ice and snow condition monitoring stations, characterized in that, The method comprises: Step 1: collecting road historical data; Step 2: establishing a set of road traffic safety risk points under ice and snow conditions according to the historical data collected in step 1; including: road traffic safety 1-level risk points and road traffic safety 2-level risk points; Step 3: constructing a multi-objective optimization site selection model according to the road historical data obtained in step 1 and the set of road traffic safety risk points under ice and snow conditions obtained in step 2; the specific process is: regarding the set of road traffic safety risk points under ice and snow conditions as nodes of the multi-objective optimization site selection model; regarding total cost, monitoring coverage index and traffic volume as three objective functions of the multi-objective optimization site selection model; the objective function of the total cost in the multi-objective optimization site selection model is expressed by a formula as follows: wherein min denotes minimization, Cost denotes the total cost of constructing the pavement ice and snow condition monitoring station, is the fixed construction cost of constructing the pavement ice and snow condition monitoring station at node i; is a 0-1 variable, indicating that if node i is selected to be constructed as the pavement ice and snow condition monitoring station, = 1, otherwise 0, i = 1, 2, …, I; I denotes the total number of nodes; denotes the milepost number of the node n selected to be constructed as the pavement ice and snow condition monitoring station, n = 1, 2, …, N; N denotes the total number of nodes selected to be constructed as the pavement ice and snow condition monitoring station, denotes the milepost number of the power supply point j, j = 1, 2, …, J; J denotes the total number of power supply points; the monitoring coverage index in the multi-objective optimization site selection model includes: the monitoring coverage index of 1-level risk points and the monitoring coverage index of 2-level risk points; the objective function of the monitoring coverage index in the multi-objective optimization site selection model is expressed by a formula as follows: the monitoring coverage index of 1-level risk points: the monitoring coverage index of 2-level risk points: wherein, represents the monitoring coverage ratio of all pavement ice and snow condition monitoring stations for the road traffic safety 1st level risk point, represents the monitoring coverage ratio of all pavement ice and snow condition monitoring stations for the road traffic safety 2nd level risk point; is a 0-1 variable, representing that if at least one pavement ice and snow condition monitoring station monitors the traffic safety risk point k, then = 1, otherwise 0, when there is a road surface ice and snow condition monitoring station within 2 kilometers of the road traffic safety risk point under ice and snow conditions, it is considered that the point is monitored and covered; k = 1, 2, …, K; K represents the total number of road traffic safety risk points; S is the total number of road traffic accident-prone points under ice and snow conditions; P is the total number of road ice and snow disaster-prone points; the objective function of the traffic volume in the multi-objective optimization site selection model is expressed by a formula as follows: wherein, is the average winter monthly average daily traffic volume of the covered road segment monitored by the entire pavement ice and snow condition monitoring station; is the winter monthly average daily traffic volume of the road segment where node i is located; and N is the total number of nodes selected to be built as pavement ice and snow condition monitoring stations. Step 4: constructing constraint conditions of the multi-objective optimization site selection model; including: site quantity constraint and site distance constraint, the constraint conditions of the multi-objective optimization site selection model are constructed as follows: the site quantity constraint is constructed by a formula as follows: wherein, is a 0-1 variable indicating if node i is selected as a site, = 1 otherwise, i = 1, 2,..., I; Q is the maximum number of ice and snow condition monitoring stations allowed to be built. the site distance constraint is constructed by a formula as follows: wherein, represents arbitrarily selecting one, represents a pavement ice and snow condition monitoring station and a pavement ice and snow condition monitoring station between the two, n1∈1,2,…,N-1; n2∈1,2,…,N-1; D represents the allowable minimum distance between the two adjacent pavement ice and snow condition monitoring stations. Step 5: generating an initial population of an improved NSGAII algorithm according to the constraint conditions of the multi-objective optimization site selection model; then solving the multi-objective optimization site selection model by using the NSGAII algorithm, and outputting a site selection result of the road surface ice and snow condition monitoring station; the specific process is: Step 51, generating the initial population P of the improved NSGAII algorithm according to the constraint conditions of the multi-objective optimization site selection model; Step 52, performing non-dominated sorting on individuals of the initial population P of the improved NSGAII algorithm, and dividing the individuals of the initial population P of the improved NSGAII algorithm into Z frontiers according to the non-dominated sorting result, Z being a positive integer; the individuals in each frontier are assigned a level according to the rank of the frontier sorting; then, the crowding degree in each frontier is calculated; Step 53, taking the initial population P of the improved NSGAII algorithm as a parent population, and performing selection, crossover and mutation three basic operations on the parent population in turn to obtain a child population; wherein the parent population of the tth iteration is denoted by P t and the offspring population of the tth iteration is denoted by Qt Step 54, merging the parent population and the child population to generate a merged population, generating a parent population of the next iteration according to the merged population; Step 55, repeating steps 52 to 54 until a maximum iteration number is reached or a satisfactory population is obtained, and then stopping iteration to obtain a final population; Step 56, the individuals in the final population are selected as the site selection results of the road ice and snow condition monitoring station.

2. The multi-objective optimization site selection method of the road ice and snow condition monitoring station according to claim 1, wherein, The road historical data in step 1 includes: road milepost data, road alignment data, road special terrain and feature data, road bridge structure data, road traffic accident-prone point data under ice and snow conditions, road snow and ice-prone point data, road fog-prone point data, road existing weather observation station data, and road power supply network point data; The road historical data is collected in step 1, and the specific process is as follows: Step 11: Collect the design documents of the road, and obtain the road milepost data, road alignment data, road special terrain and feature point data, and road bridge structure data according to the design documents of the road; The road special terrain and feature point data includes: road sharp bend and steep slope point data, and road bridge and tunnel point data; Step 12: Collect the road infrastructure construction data provided by the traffic investment department, and determine the road existing weather observation station data and the road power supply network point data according to the road infrastructure construction data provided by the traffic investment department; Step 13: Collect the road historical traffic accident investigation data, and obtain the road traffic accident-prone point information under ice and snow conditions according to the road historical traffic accident investigation data in the past 10 years; The road traffic accident-prone point under ice and snow conditions is a place where the number of road traffic accidents under ice and snow conditions exceeds 5 or more in the past 10 years; Step 14: Determine the weather observation stations within a straight-line distance of 20 km along the road according to the road existing weather observation station data; Collect historical weather data observed by the weather observation stations within a straight-line distance of 20 km along the road; Obtain the road weather disaster-prone point data according to the historical weather data observed by the weather observation stations within a straight-line distance of 20 km along the road in the past 20 years; the road weather disaster-prone point data includes: road snow and ice-prone point data and road fog-prone point data, The road weather disaster-prone point is a place where the number of weather disasters exceeds 5 or more in the past 20 years; Step 15: Take the road special terrain and feature point and the road weather disaster-prone point as the road ice and snow disaster-prone point.

3. The multi-objective optimization site selection method of the road ice and snow condition monitoring station according to claim 2, wherein, The specific process of establishing the road traffic safety risk point set under ice and snow conditions according to the historical data collected in step 1 in step 2 is as follows: Take the road traffic accident-prone point under ice and snow conditions as the road traffic safety 1st-level risk point under ice and snow conditions; Take the road ice and snow disaster-prone point as the road traffic safety 2nd-level risk point under ice and snow conditions.

4. The multi-objective optimization site selection method of the road ice and snow condition monitoring station according to claim 3, wherein, The specific process of generating the initial population P of the improved NSGAII algorithm according to the constraint conditions of the multi-objective optimization site selection model in step 51 is as follows: Step 511: Set the size of the initial population , a crossover probability nMu, a mutation probability, and a maximum number of iterations M. Step 512: generate constraints according to the constraint conditions of the multi-objective optimization site selection model; Step 513: Based on the initial population generation constraints in Step 512, randomly generate... The population consists of individuals, resulting in an initial population P improved by the NSGAII algorithm.

5. The multi-objective optimization siting method of a pavement ice and snow condition monitoring station according to claim 4, characterized in that, In step 52, the individuals of the improved initial population P of the NSGAII algorithm are non-dominantly sorted, and the individuals of the improved initial population P of the NSGAII algorithm are divided into Z frontiers according to the results of the non-dominant sorting, Z being a positive integer; the individuals in each frontier are assigned a level according to the ranking of the frontier sorting; then the crowding degree in each frontier is calculated; the specific process is: Step 521: calculate the objective function value of each individual in the initial population P; the i-th individual in the initial population P is denoted as pi Step 522: according to the objective function value of all individuals obtained and the dominance relation in the NSGAII algorithm, all individuals are divided into Z frontiers, The individuals in the first frontier are individuals that are not dominated by any individual, and the individuals in the subsequent frontiers are individuals dominated by the individuals in the previous frontiers; Step 523: according to the objective function value of the individuals in each frontier and the crowding degree calculation method in the NSGAII algorithm, the crowding degree of the individuals in each frontier is calculated to obtain the crowding degree value of each individual in each frontier.

6. The multi-objective optimization siting method of a pavement ice and snow condition monitoring station according to claim 5, characterized in that, In step 53, the parent population is sequentially selected, crossed, and mutated to obtain the offspring population; the specific process is: Step 531: according to the non-dominant level of the parent population individuals and the crowding degree in each frontier, two parent individuals are selected from the parent population, Step 532: according to the set crossover probability nMu, the selected parent individuals are crossed to generate offspring individuals, Step 533: according to the set mutation probability, all generated offspring individuals are mutated to obtain the offspring population.

7. The multi-objective optimization site selection method of the pavement ice and snow condition monitoring station according to claim 6, characterized in that, The parent population P of the tth iteration is generated in the step 54 t and the offspring population Q of the tth iteration is generated in the step 53 t are combined to generate the combined population R of the tth iteration t ; Merge the population R according to the tth iteration t Generate the parent population R of the (t+1)th iteration t+1 The specific process is as follows: Step 541 : merge the parent population P of the tth iteration t and the offspring population Q of the tth iteration t to generate a merged population R of the tth iteration t ; Step 542: Non-dominated sorting of individuals in the tth iteration merged population R t ; r frontiers are obtained, and each individual in each frontier is assigned a rank according to the rank of the individual in the frontier; Step 543: Selecting according to the individual's rank The individual forms the population of the t+1 iteration .

Images