Intelligent line selection and evaluation decision-making method for railway engineering construction road in complex region

Abstract

The invention relates to the technical field of railway engineering construction, and discloses a complex region railway engineering construction road intelligent line selection and evaluation decision method, which comprises the following steps: S1, in combination with terrain environment conditions and construction technology standards of a construction road, constructing an improved HAO plane line shape mathematical model, designing plane line shape constraint conditions, and constructing a line shape model; and establishing a line comprehensive cost model as an optimization objective function, and searching an optimal plane linear position parameter in a target area by using a particle swarm optimization algorithm. Compared with other intelligent line selection methods, the method can quickly evaluate and decide the advantages and disadvantages of the route scheme while quickly generating the horizontal and vertical section line shape of the construction road, optimizes the design process, facilitates the shortening of the construction period of the construction road, can remarkably reduce the workload of site survey and manual design, and improves the construction efficiency. The working efficiency and benefits of line selection design are practically improved, the investment of line selection cost is reduced, use is convenient, and the degree of dependence on equipment performance is low.

Description

Technical Field

[0001] The invention belongs to the technical field of railway engineering construction, and relates to an intelligent line selection and evaluation decision-making method for railway engineering construction roads in complex areas. Background Art

[0002] With my country's emphasis on economic development in the west, the road construction industry in the west has received strong support. However, the geological terrain in the western region is extremely complex, and the scarce materials and weak transportation facilities make it difficult to support the needs of large-scale road construction. Therefore, a large number of construction roads need to be rebuilt and expanded to ensure the transportation and supply of construction materials. As the most critical part of road engineering, the line selection design of construction roads directly affects the economy, functionality and safety of construction roads. Traditional manual line selection design not only requires a lot of manpower investment, but also consumes a lot of money and time. At the same time, the decision-making of line plans also adopts subjective evaluation methods such as expert evaluation, which is greatly affected by the personal experience and cognitive level of the evaluator and has great subjective uncertainty. Therefore, studying an intelligent line design and evaluation decision-making method that can be applied to construction roads is a major innovative direction of line selection design research.

[0003] At present, there are relatively few studies that integrate intelligent route selection technology with route evaluation and decision-making, and there is not much attention paid to low-level highway construction such as construction roads. In addition, since route selection design is a complex process with multiple objectives and multiple constraints, the established intelligent route selection model and intelligent algorithm cannot fit well with the actual project, and further in-depth research is needed. Summary of the invention

[0004] The purpose of the present invention is to provide a method for realizing automatic design of route plans for new construction roads and route plan evaluation and optimization, which helps to automatically optimize the most economical and suitable route plan in complex terrain, and can effectively provide decision-making basis for designers in route plan comparison and selection, saving labor and cost investment.

[0005] To achieve the above object, the present invention provides the following technical solution: a method for intelligent line selection and evaluation decision-making of railway engineering construction roads in complex areas, the method comprising the following steps:

[0006] S1. Combined with the terrain and environmental conditions of the construction road and the construction technical standards, the plane alignment mathematical model of the improved HAO is constructed, the plane alignment constraint conditions are designed, the line comprehensive cost model is established as the optimization objective function, and the particle swarm optimization algorithm is used to search for the optimal plane alignment position parameters in the target area;

[0007] The specific steps include:

[0008] 1) Obtain geographic information of the design area, including topography and geology, land use, transportation network, hydrological distribution, and building facilities, as basic data for line selection and design;

[0009] 2) For tight slope sections, including those that require a return curve to reach the target elevation, the maximum slope section design is generally adopted. Combined with the minimum straight line length requirement, a distance l is agreed as the straight line distance of the next intersection point. Combined with the terrain elevation data, the point on the circle with a diameter of l whose height difference slope best meets the maximum longitudinal slope requirement is selected as the next intersection point. This cycle can be repeated to determine the tight slope plane line shape;

[0010] 3) Divide the starting and ending points of the gentle slope section into n perpendicular bisectors, whose coordinates are represented by O i (x Oi ,y Oi ), the intersection points of the lines can be selected by the plumb line on the equally divided points, and the plane linear position can be determined by the number of intersection points n, the distance d from the intersection point to the equally divided point i The radius R of the peace curve is determined,

[0011]

[0012] 4) Use the following formula to calculate the comprehensive cost of the route plan, including the cost of roadbed earthwork C T , roadbed construction costs C J , bridge, culvert engineering cost C Q , land acquisition compensation costs C Z , Demolition compensation costs C C and penalty fee C F ;

[0013] C=C T +C J +C Q +C Z +C C +C F

[0014] 5) Using the three plane linear position parameters as particle swarm variables and the line cost as the optimization objective function, randomly generate the initial population variable values, calculate the corresponding initial population objective function values, and record the initial population as the optimal population initial position gBest 0 , record the particle with the best fitness function value in the initial population as the initial best particle position pBest 0 ;

[0015] 6) Update the variable position and speed of the particle swarm to form a new route plan

[0016]

[0017] Where, d is the number of optimization iterations;

[0018] ω——weight coefficient

[0019] ——the velocity and position of particle k at the dth iteration;

[0020] c1, c2 – acceleration coefficients of particle position and velocity, also called learning factors;

[0021] ——Two random numbers in the interval [0, 1];

[0022] ——The best historical position of particle k during the optimization process;

[0023] gBest d ——The best historical position of the population during the optimization process;

[0024] When updating the particle speed or position, the value is obtained according to the following formula:

[0025]

[0026] 7) Recalculate the fitness function of each particle in the new population and compare it with the corresponding fitness value in the best population recorded. If a better particle position is found, update gBest; similarly, update the best particle pBest;

[0027] 8) Determine the termination condition. If the condition is not met, return to step e. When the condition is met, terminate the loop;

[0028] 9) Output the best particle pBest position;

[0029] S2, based on the roof truss function to smooth the ground line shape, use the intersection of the smooth curve to determine the initial longitudinal section line shape slope change point, set the longitudinal section constraint conditions, build the longitudinal section line shape optimization model optimized by genetic algorithm, and generate the best longitudinal section design scheme;

[0030] The specific steps are as follows:

[0031] 1) The fill and cut value on the center line of the longitudinal section, that is, the area formed by the elevation difference between the design elevation and the ground line, is used as the objective function F(X) for quickly generating the longitudinal section. ks ; Take the construction cost as the objective function of longitudinal section intelligent optimization F(X) zn , including roadbed earthwork cost f1(X), bridge cost f2(X), and tunnel cost f3(X);

[0032]

[0033] F(X) zn =f1(X)+f2(X)+f3(X)

[0034] Where, y(x)——longitudinal section design line;

[0035] h(x)——ground line of longitudinal section;

[0036] p(x)——weight function, when cross-sectional factors are not considered, p(x)=1;

[0037] 2) Determine the constraints according to the specifications, taking into account the slope length constraints, slope limit constraints, control point elevation constraints, and vertical curve constraints to complete the model construction;

[0038] 3) Establish the initial feasible area of ​​the longitudinal section with the slope constraint condition, and preliminarily trim the ground line; use the height weighted average method to process the trimmed ground line into a smooth curve, and generate multiple smooth curves by setting the size of the smooth radius. Their intersection is the inflection point, and the slope section is divided in combination with the ruggedness;

[0039] 4) The least square method is used to fit the slope line. The intersection of adjacent slope lines is the slope change point of the initial scheme. The resulting scheme is checked and processed for constraints. The vertical curves are configured and the bridges and tunnels are set to quickly generate the longitudinal section that meets the requirements.

[0040]

[0041] 5) The slope change points are randomly distributed by combining the uniform distribution method and the normal distribution method, and the initial population of the longitudinal section scheme is generated after relevant constraints are imposed;

[0042] 8) Setting genetic algorithm optimization control parameters, using real number coding;

[0043] 9) The quality of the longitudinal section scheme is evaluated and judged based on the established construction project cost objective function, thereby constructing the longitudinal section scheme fitness function E(i) and the slope section fitness function E(i) k ;

[0044] E(i)=G max -F i (X)

[0045] Where, E(i) is the fitness value of the i-th longitudinal section scheme;

[0046] F i (X)——the objective function value of each longitudinal section scheme;

[0047] G max ——The maximum value of the objective function in the longitudinal section scheme;

[0048]

[0049] Where, E(i) k ——Fitness value of the kth slope section in scheme i;

[0050] C cub ——the cost of filling and excavation per cubic meter of earth and stone;

[0051] V (j、cub) ——the fill and cut volume of the jth road section in the kth slope section;

[0052] L bir , L tun ——The length of the bridge and tunnel in the kth slope section;

[0053] C bir , C tun ——Cost per linear meter of bridges and tunnels;

[0054] ΔL k ——the length of the kth slope section;

[0055] 8) Design genetic operators; use roulette method to select operators; combine arithmetic crossover and heuristic crossover to form a crossover operator, and introduce slope fitness function to improve it, so that the optimization operation is more suitable for longitudinal section design; use adaptive mutation operator to determine the mutation range by fitness; in summary, complete the design and improvement of genetic operators;

[0056] 9) Set the termination criteria of the genetic algorithm to terminate after reaching the preset maximum number of iterations or evolutionary generations or the cumulative running time reaches 420 minutes;

[0057] S3. Construct a multi-objective route direction scheme evaluation system, evaluate the designed route scheme, construct a TOPSIS route decision model based on grey correlation, select multiple schemes, and decide on the best scheme; the specific steps are as follows:

[0058] 1) Data acquisition: obtaining the basic data contained in the construction road route plan;

[0059] 2) Clarify the indicators and determine the weights of the corresponding evaluation indicators based on the importance comparison judgment matrix of the preset evaluation and optimization decision indicators;

[0060] 3) Determine the weights. The weights of the corresponding evaluation and optimization indicators can be determined by the least square method of weights, which specifically includes: according to the pairwise judgment matrix between the preset indicators or the pairwise judgment matrix between the indicators determined by the decision maker, further calculate the weight vector of the formed judgment matrix, and use the value of the weight vector as the weight of the corresponding indicator;

[0061] The specific steps to determine the weights are:

[0062] (a) Determine the hierarchy of evaluation options

[0063] This step divides the evaluation and optimization index system into the target layer, the criterion layer and the index layer;

[0064] (b) Construct a preliminary judgment matrix A = (a ij ) n×n

[0065] According to the emphasis and cognition of decision makers on the importance of evaluation and optimization indicators, the evaluation and optimization indicators are compared pairwise to obtain a preliminary judgment matrix:

[0066]

[0067] (c) Solve matrix B based on the judgment matrix

[0068]

[0069] (d) Calculate the sorting vector W by combining the identity matrix e

[0070] The calculation formula is as follows:

[0071] e=(1,1,1,...,1) T

[0072]

[0073] The final ranking vector W is the weight vector value of the evaluation index of the corresponding solution;

[0074] (e) Scheme evaluation, according to the fuzzy membership relationship between the determined weight of the evaluation index and the preset construction road route scheme evaluation index, different optional route design schemes are calculated, and the advantages and disadvantages of the route scheme are determined by the evaluation score; in a specific evaluation embodiment, the method includes: according to the evaluation results of the construction road route scheme, selecting valuable information for display; displaying the overall evaluation score of the selected scheme, and also displaying the intermediate process data such as the basic data acquisition results and the evaluation index weight determination results in the above steps;

[0075] The specific processing steps are as follows:

[0076] a) Constructing the evaluation factor set

[0077] Each evaluation index in the evaluation index system is used as an evaluation factor, with u i To represent it, a set of evaluation factors is constructed based on it, and U is used to represent the set of evaluation factors; that is, U = {U1, U2, ..., U t}, where U irepresents the i-th factor class, corresponding to the criterion layer in the evaluation index system, and each U i It also contains a certain number of factors, namely U i = {u i1 ,u i2 ,...,u is}, corresponding to the specific evaluation indicators obtained by refining and decomposing the criterion layer within the evaluation indicator system;

[0078] b) Determine the review set

[0079] By analyzing the pros and cons of the evaluation indicators, the evaluation indicator levels are divided into five types, namely, very good, very good, good, average and poor; different evaluation levels correspond to a comprehensive evaluation score, and when the corresponding evaluation indicator meets the range of a corresponding evaluation level, it is allowed to take the comprehensive evaluation value of the corresponding level, and the rest of the situations are the same as above; according to the actual situation of the evaluation indicator system involved, the comment set is set to V, where V = {V1, V2, V3, V4, V5} = {100, 90, 80, 70, 60};

[0080] c) Determine the quantitative indicator evaluation level membership

[0081] The value range of the evaluation index membership is specified to be between 0 and 1. When the membership value of the corresponding evaluation index is larger, it is considered that the index is closer to a certain evaluation level. The evaluation indexes selected here are all quantitative indicators. When the value of a certain evaluation index falls within the corresponding evaluation index level interval, it is considered that its membership value at the corresponding level is 1, and the membership values ​​at other levels are all 0. For example, when the value of the i-th evaluation index satisfies the evaluation level interval of "very good", the membership value of the index at the "very good" level is 1, and the membership values ​​of the remaining very good, good, average and poor levels are all 0.

[0082] After determining the corresponding membership degree of each evaluation indicator in the evaluation indicator system according to the evaluation indicator level, the membership degree of the evaluation indicator is expressed as R i =(r i1 ,r i2 ,...,r im ),i=1,2,...,n;R i Where n is the number of alternative evaluation indicators, m is the number of membership levels corresponding to the alternative evaluation indicators, and r i j represents the membership value of the i-th evaluation indicator at the j-th membership level; based on this, the corresponding membership matrix R of the scheme to be evaluated can be constructed, which is expressed as follows:

[0083]

[0084] d) Determine the weight of evaluation indicators

[0085] In the evaluation index system, different evaluation indicators have a certain degree of importance, and the importance of the evaluation indicators is usually reflected in the form of weights; the weight of the evaluation indicator reflects its importance in the route plan evaluation index system, and the sum of the weights of all indicators in the evaluation index system should be 1; that is, the weight set W = {w1, w2, ..., w n}middle And any w i ≥0;

[0086] e) Multi-level fuzzy comprehensive evaluation

[0087] The evaluation index system can be decomposed into the criterion layer and the indicator layer. The indicator layer is composed of specific evaluation indicators after the criterion layer is refined. Therefore, the fuzzy comprehensive evaluation of the route plan needs to be carried out in two steps;

[0088] First-level fuzzy comprehensive evaluation:

[0089] By assigning weights to the specific evaluation indicators of the indicator layer obtained after the detailed decomposition of each criterion layer, the weight set W of the route plan under different indicator layers is obtained. i , and the specific evaluation indicators under each indicator layer can be calculated to obtain the corresponding evaluation indicator membership matrix R i Through the calculation of the weight set of each indicator layer and its membership matrix, the fuzzy comprehensive evaluation results B of each criterion layer can be obtained. i ; The calculation formula is:

[0090]

[0091] Secondary fuzzy comprehensive evaluation:

[0092] The calculation result of the first-level fuzzy comprehensive evaluation is B i Arrange it and make it as the secondary fuzzy membership evaluation matrix R i ', based on which the criterion layer membership matrix R = (R1', R'2, ..., R t ') T , according to the weight set W of the route plan under the criterion layer, the final fuzzy comprehensive evaluation result B of the route plan is further calculated, and the calculation formula is:

[0093]

[0094] After obtaining the secondary fuzzy comprehensive evaluation results, the results are further processed with the comment set to obtain the final fuzzy comprehensive evaluation score. The calculation formula is as follows:

[0095]

[0096] f) Scheme optimization, according to the weight of the determined evaluation index and the data information of the preset construction road route scheme optimization index, the gray correlation TOPSIS comprehensive method is used to perform optimization decision calculation on different optional route design schemes, and the optimal route scheme is determined by the relative closeness; in a specific optimization decision embodiment, the method further includes: according to the optimization result of the construction road route scheme, selecting valuable information for display; displaying the relative closeness of the selected scheme, and also displaying the intermediate process data of the basic data acquisition result and the optimization index weight determination result in the above steps;

[0097] The specific processing steps are as follows:

[0098] ① Establish an initial decision matrix

[0099] According to m alternatives and the evaluation indicators of each alternative, an initial decision matrix is ​​constructed:

[0100]

[0101] ② Normalize the initial decision matrix

[0102] The initial decision information matrix is ​​normalized to obtain the normalized matrix V = (V ij ) m×n :

[0103]

[0104] ③Determine positive and negative ideal solutions

[0105] Based on the normalized standard matrix elements and the concept of ideal solutions, positive and negative ideal solutions that meet the conditions are selected:

[0106] Determination of the ideal solution: in:

[0107]

[0108] Determination of negative ideal solution: in:

[0109]

[0110] ④ Calculate the Euclidean distance between the alternative solution and the positive and negative ideal solutions

[0111]

[0112] ⑤ Calculate the grey correlation coefficient and grey correlation degree between the alternative scheme and the positive and negative ideal schemes

[0113]

[0114] In the formula, ρ is the resolution coefficient. Studies have shown that a relatively good resolution can be obtained when its value is 0.5;

[0115] The grey correlation degree of different schemes is solved by the following formula, corresponding to the i-th scheme:

[0116]

[0117] ⑥Optimal ranking of construction road route plans

[0118] Since the Euclidean distance obtained by TOPSIS method and the grey correlation degree obtained by grey correlation method have different dimensions, they are dimensionless for comprehensive consideration:

[0119]

[0120] Let the relative closeness obtained by TOPSIS method be

[0121]

[0122] Let the relative closeness obtained by the grey correlation method be

[0123]

[0124] By organically combining the relative closeness obtained by TOPSIS method and the relative closeness obtained by grey correlation method, the comprehensive closeness can be obtained:

[0125]

[0126] In the formula, α and β are both taken as 0.5, ω i It represents the combined weight vector obtained above. When the comprehensive closeness Q value is larger, the corresponding route plan is better.

[0127] Preferably, in step S3, the basic data refers to a plurality of optional construction road route plans in the early preliminary design plan, and these initial input data constituting the construction road route plan will affect the route evaluation and optimization decision results.

[0128] Preferably, in step S3, the evaluation and optimization decision indicators of the construction road route plan are selected according to the design route data generated by the system, and specifically include the engineering economic indicators, route operation indicators, environmental impact indicators and route plan, and the technical indicators and engineering economic indicators involved in the optimization; wherein the engineering economic indicators in the aforementioned route plan evaluation include route length, slope utilization rate, bridge-tunnel ratio, earthwork quantity per kilometer, and construction cost per kilometer; route operation indicators include route length, number of intersections per kilometer, and number of slope changes per kilometer; environmental impact indicators include bridge-tunnel ratio and earthwork quantity per kilometer; and the road The technical indicators in the route scheme optimization decision include route length, maximum slope, minimum radius of horizontal curve, minimum radius of vertical curve, number of intersections of horizontal curve and number of intersections of vertical curve; the engineering economic indicators include bridge-tunnel ratio, amount of roadbed earthwork, number of culverts, land acquisition area, demolition area and project cost; in order to avoid the disadvantage that the process of determining the advantages and disadvantages of qualitative indicators is interfered by human factors and the evaluation and optimization results are highly subjective, the indicators used in this method are all quantitative indicators, and the route schemes are no longer evaluated and optimized with comments such as "more", "less", "good" and "bad", and the best route scheme is evaluated and optimized based on this.

[0129] Compared with the prior art, the present invention has the following beneficial effects:

[0130] 1. Compared with other intelligent route selection methods, the present invention can quickly generate the horizontal and vertical section alignment of the construction road, and can quickly evaluate and decide on the pros and cons of the route plan, thereby optimizing the design process and helping to shorten the construction period of the construction road;

[0131] 2. The present invention can significantly reduce the workload of on-site investigation and manual design, effectively improve the work efficiency and benefits of line selection design, and reduce the investment in line selection costs;

[0132] 3. The present invention is easy to use, has low dependence on equipment performance, and has a short design process, making it easy to use on the construction site;

[0133] 4. The present invention can be applied to most similar projects, and the results are highly reliable and have good universal applicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0134] Figure 1 It is a schematic diagram of the improved HAO model of the present invention;

[0135] Figure 2 A diagram of a tight slope alignment method of the present invention;

[0136] Figure 3 is a plane line diagram of the present invention;

[0137] Figure 4It is a setting plane optimization parameter diagram of the present invention;

[0138] Figure 5 It is the plane alignment result diagram of the tight slope section of the present invention;

[0139] Figure 6 It is the plane linear optimization result diagram of the gentle slope section of the present invention;

[0140] Figure 7 It is a longitudinal section constraint parameter setting diagram of the present invention;

[0141] Figure 8 Generating a result graph for the initial solution of the longitudinal section solution of the present invention;

[0142] Fig. 9 It is the optimization parameter setting diagram of the present invention;

[0143] Fig.10 It is a longitudinal section optimization result diagram of the present invention;

[0144] Fig.11 The main index data diagram of the above-mentioned linear generation scheme of the present invention;

[0145] Fig.12 An interface diagram for scoring the technical indicator layer of the circuit scheme of the present invention;

[0146] Fig.13 An interface diagram for scoring the engineering economic layer of the line scheme of the present invention;

[0147] Fig.14 It is a diagram of the decision result of the scheme evaluation of the present invention;

[0148] Fig.15 The linear diagrams a, b and c of the longitudinal section initial scheme of the present invention are shown;

[0149] Fig.16 This is the linear diagram of the longitudinal section artificial slope pulling scheme of the present invention.

[0150] Fig.17 The linear diagrams a, b and c of the longitudinal section optimization scheme of the present invention are shown;

[0151] Fig.18 A weight map of the evaluation criteria layer of the solution of the present invention;

[0152] Fig.19 The engineering economic index weight diagram of the present invention;

[0153] Fig. 20 The route operation index weight diagram of the present invention;

[0154] Fig.21 is the environmental impact index weight map of the present invention;

[0155] Fig. 22 This is a diagram showing the evaluation results of Scheme 1 of the present invention;

[0156] Fig.23 This is a diagram showing the evaluation results of Scheme 2 of the present invention;

[0157] Fig.24 It is a weight map of the optimization criteria layer of the solution of the present invention;

[0158] Fig.25 It is the technical indicator weight diagram of the present invention;

[0159] Fig.26 The engineering economic index weight diagram of the present invention;

[0160] Fig. 27 This is a graph showing the optimization results of the grey correlation TOPSIS comprehensive method solution of the present invention. DETAILED DESCRIPTION

[0161] 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.

[0162] A method for intelligent route selection and evaluation decision-making for railway engineering construction roads in complex areas, the method comprising the following steps:

[0163] S1. Combined with the terrain and environmental conditions of the construction road and the construction technical standards, the plane alignment mathematical model of the improved HAO is constructed, the plane alignment constraint conditions are designed, the line comprehensive cost model is established as the optimization objective function, and the particle swarm optimization algorithm is used to search for the optimal plane alignment position parameters in the target area;

[0164] The specific steps include:

[0165] 1) Obtain geographic information of the design area, including topography and geology, land use, transportation network, hydrological distribution, and building facilities, as basic data for line selection and design;

[0166] 2) For tight slope sections, including those that require a return curve to reach the target elevation, the maximum slope section design is generally adopted. Combined with the minimum straight line length requirement, a distance l is agreed as the straight line distance of the next intersection point. Combined with the terrain elevation data, the point on the circle with a diameter of l and the slope that best meets the maximum longitudinal slope requirement is selected as the next intersection point. The plane line shape of the tight slope can be determined by continuous cycles, such as Figure 2 As shown;

[0167] 3) Divide the starting and ending points of the gentle slope section into n perpendicular bisectors, whose coordinates are represented by O i (x Oi ,y Oi ), the intersection points of the lines can be selected by the plumb line on the equally divided points, and the plane linear position can be determined by the number of intersection points n, the distance d from the intersection point to the equally divided point i The radius R of the peace curve is determined as follows: Figure 1 As shown,

[0168]

[0169] 4) Use the following formula to calculate the comprehensive cost of the route plan, including the cost of roadbed earthwork C T , roadbed construction costs C J , bridge, culvert engineering cost C Q , land acquisition compensation costs C Z , Demolition compensation costs C C and penalty fee C F ;

[0170] C=C T +C J +C Q +C Z +C C +C F

[0171] 5) Using the three plane linear position parameters as particle swarm variables and the line cost as the optimization objective function, randomly generate the initial population variable values, calculate the corresponding initial population objective function values, and record the initial population as the optimal population initial position gBest 0 , record the particle with the best fitness function value in the initial population as the initial best particle position pBest 0 ;

[0172] 6) Update the variable position and speed of the particle swarm to form a new route plan

[0173]

[0174] Where, d is the number of optimization iterations;

[0175] ω——weight coefficient

[0176] ——the velocity and position of particle k at the dth iteration;

[0177] c1, c2 – acceleration coefficients of particle position and velocity, also called learning factors;

[0178] ——Two random numbers in the interval [0, 1];

[0179] ——The best historical position of particle k during the optimization process;

[0180] gBest d ——The best historical position of the population during the optimization process;

[0181] When updating the particle speed or position, the value is obtained according to the following formula:

[0182]

[0183] 7) Recalculate the fitness function of each particle in the new population and compare it with the corresponding fitness value in the best population recorded. If a better particle position is found, update gBest; similarly, update the best particle pBest;

[0184] 8) Determine the termination condition. If the condition is not met, return to step e. When the condition is met, terminate the loop;

[0185] 9) Output the best particle pBest position;

[0186] S2, based on the roof truss function to smooth the ground line shape, use the intersection of the smooth curve to determine the initial longitudinal section line shape slope change point, set the longitudinal section constraint conditions, build the longitudinal section line shape optimization model optimized by genetic algorithm, and generate the best longitudinal section design scheme;

[0187] The specific steps are as follows:

[0188] 1) The fill and cut value on the center line of the longitudinal section, that is, the area formed by the elevation difference between the design elevation and the ground line, is used as the objective function F(X) for quickly generating the longitudinal section. ks ; Take the construction cost as the objective function of longitudinal section intelligent optimization F(X) zn , including roadbed earthwork cost f1(X), bridge cost f2(X), and tunnel cost f3(X);

[0189]

[0190] F(X) zn =f1(X)+f2(X)+f3(X)

[0191] Where, y(x)——longitudinal section design line;

[0192] h(x)——ground line of longitudinal section;

[0193] p(x)——weight function, when cross-sectional factors are not considered, p(x)=1;

[0194] 2) Determine the constraints according to the specifications, taking into account the slope length constraints, slope limit constraints, control point elevation constraints, and vertical curve constraints to complete the model construction;

[0195] 3) Establish the initial feasible area of ​​the longitudinal section with the slope constraint condition, and preliminarily trim the ground line; use the height weighted average method to process the trimmed ground line into a smooth curve, and generate multiple smooth curves by setting the size of the smooth radius. Their intersection is the inflection point, and the slope section is divided in combination with the ruggedness;

[0196] 4) The least square method is used to fit the slope line. The intersection of adjacent slope lines is the slope change point of the initial scheme. The resulting scheme is checked and processed for constraints. The vertical curves are configured and the bridges and tunnels are set to quickly generate the longitudinal section that meets the requirements.

[0197]

[0198] 5) The slope change points are randomly distributed by combining the uniform distribution method and the normal distribution method, and the initial population of the longitudinal section scheme is generated after relevant constraints are imposed;

[0199] 10) Setting genetic algorithm optimization control parameters, using real number coding;

[0200] 11) The quality of the longitudinal section scheme is evaluated and judged based on the established construction project cost objective function, thereby constructing the longitudinal section scheme fitness function E(i) and the slope section fitness function E(i) k ;

[0201] E(i)=G max -F i (X)

[0202] Where, E(i) is the fitness value of the i-th longitudinal section scheme;

[0203] F i (X)——the objective function value of each longitudinal section scheme;

[0204] G max ——The maximum value of the objective function in the longitudinal section scheme;

[0205]

[0206] Where, E(i) k ——Fitness value of the kth slope section in scheme i;

[0207] C cub ——the cost of filling and excavation per cubic meter of earth and stone;

[0208] V (j、cub) ——the fill and cut volume of the jth road section in the kth slope section;

[0209] L bir , L tun ——The length of the bridge and tunnel in the kth slope section;

[0210] C bir , C tun ——Cost per linear meter of bridges and tunnels;

[0211] ΔL k ——the length of the kth slope section;

[0212] 8) Design genetic operators; use roulette method to select operators; combine arithmetic crossover and heuristic crossover to form a crossover operator, and introduce slope fitness function to improve it, so that the optimization operation is more suitable for longitudinal section design; use adaptive mutation operator to determine the mutation range by fitness; in summary, complete the design and improvement of genetic operators;

[0213] 9) Set the termination criteria of the genetic algorithm to terminate after reaching the preset maximum number of iterations or evolutionary generations or the cumulative running time reaches 420 minutes;

[0214] S3. Construct a multi-objective route direction scheme evaluation system, evaluate the designed route scheme, construct a TOPSIS route decision model based on grey correlation, select multiple schemes, and decide on the best scheme; the specific steps are as follows:

[0215] 1) Data acquisition: obtaining the basic data contained in the construction road route plan;

[0216] 2) Clarify the indicators and determine the weights of the corresponding evaluation indicators based on the importance comparison judgment matrix of the preset evaluation and optimization decision indicators;

[0217] 3) Determine the weights. The weights of the corresponding evaluation and optimization indicators can be determined by the least square method of weights, which specifically includes: according to the pairwise judgment matrix between the preset indicators or the pairwise judgment matrix between the indicators determined by the decision maker, further calculate the weight vector of the formed judgment matrix, and use the value of the weight vector as the weight of the corresponding indicator;

[0218] The specific steps to determine the weights are:

[0219] (a) Determine the hierarchy of evaluation options

[0220] This step divides the evaluation and optimization index system into the target layer, the criterion layer and the index layer;

[0221] (b) Construct a preliminary judgment matrix A = (a ij ) n×n

[0222] According to the emphasis and cognition of decision makers on the importance of evaluation and optimization indicators, the evaluation and optimization indicators are compared pairwise to obtain a preliminary judgment matrix:

[0223]

[0224] (c) Solve matrix B based on the judgment matrix

[0225]

[0226] (d) Calculate the sorting vector W by combining the identity matrix e

[0227] The calculation formula is as follows:

[0228] e=(1,1,1,...,1) T

[0229]

[0230] The final ranking vector W is the weight vector value of the evaluation index of the corresponding solution;

[0231] (e) Scheme evaluation, according to the fuzzy membership relationship between the determined weight of the evaluation index and the preset construction road route scheme evaluation index, different optional route design schemes are calculated, and the advantages and disadvantages of the route scheme are determined by the evaluation score; in a specific evaluation embodiment, the method includes: according to the evaluation results of the construction road route scheme, selecting valuable information for display; displaying the overall evaluation score of the selected scheme, and also displaying the intermediate process data such as the basic data acquisition results and the evaluation index weight determination results in the above steps;

[0232] The specific processing steps are as follows:

[0233] a) Constructing the evaluation factor set

[0234] Each evaluation index in the evaluation index system is used as an evaluation factor, with u i To represent it, a set of evaluation factors is constructed based on it, and U is used to represent the set of evaluation factors; that is, U = {U1, U2, ..., U t}, where U i represents the i-th factor class, corresponding to the criterion layer in the evaluation index system, and each U i It also contains a certain number of factors, namely U i = {u i1 ,u i2 ,...,u is}, corresponding to the specific evaluation indicators obtained by refining and decomposing the criterion layer within the evaluation indicator system;

[0235] b) Determine the review set

[0236] By analyzing the pros and cons of the evaluation indicators, the evaluation indicator levels are divided into five types, namely, very good, very good, good, average and poor; different evaluation levels correspond to a comprehensive evaluation score, and when the corresponding evaluation indicator meets the range of a corresponding evaluation level, it is allowed to take the comprehensive evaluation value of the corresponding level, and the rest of the situations are the same as above; according to the actual situation of the evaluation indicator system involved, the comment set is set to V, where V = {V1, V2, V3, V4, V5} = {100, 90, 80, 70, 60};

[0237] c) Determine the quantitative indicator evaluation level membership

[0238] The value range of the evaluation index membership is specified to be between 0 and 1. When the membership value of the corresponding evaluation index is larger, it is considered that the index is closer to a certain evaluation level. The evaluation indexes selected here are all quantitative indicators. When the value of a certain evaluation index falls within the corresponding evaluation index level interval, it is considered that its membership value at the corresponding level is 1, and the membership values ​​at other levels are all 0. For example, when the value of the i-th evaluation index satisfies the evaluation level interval of "very good", the membership value of the index at the "very good" level is 1, and the membership values ​​of the remaining very good, good, average and poor levels are all 0.

[0239] After determining the corresponding membership degree of each evaluation indicator in the evaluation indicator system according to the evaluation indicator level, the membership degree of the evaluation indicator is expressed as R i =(r i1 ,r i2 ,...,r im ),i=1,2,...,n;R i Where n is the number of alternative evaluation indicators, m is the number of membership levels corresponding to the alternative evaluation indicators, and r i j represents the membership value of the i-th evaluation indicator at the j-th membership level; based on this, the corresponding membership matrix R of the scheme to be evaluated can be constructed, which is expressed as follows:

[0240]

[0241] d) Determine the weight of evaluation indicators

[0242] In the evaluation index system, different evaluation indicators have a certain degree of importance, and the importance of the evaluation indicators is usually reflected in the form of weights; the weight of the evaluation indicator reflects its importance in the route plan evaluation index system, and the sum of the weights of all indicators in the evaluation index system should be 1; that is, the weight set W = {w1, w2, ..., w n}middle And any w i ≥0;

[0243] e) Multi-level fuzzy comprehensive evaluation

[0244] The evaluation index system can be decomposed into the criterion layer and the indicator layer. The indicator layer is composed of specific evaluation indicators after the criterion layer is refined. Therefore, the fuzzy comprehensive evaluation of the route plan needs to be carried out in two steps;

[0245] First-level fuzzy comprehensive evaluation:

[0246] By assigning weights to the specific evaluation indicators of the indicator layer obtained after the detailed decomposition of each criterion layer, the weight set W of the route plan under different indicator layers is obtained. i , and the specific evaluation indicators under each indicator layer can be calculated to obtain the corresponding evaluation indicator membership matrix R i Through the calculation of the weight set of each indicator layer and its membership matrix, the fuzzy comprehensive evaluation results B of each criterion layer can be obtained. i ; The calculation formula is:

[0247]

[0248] Secondary fuzzy comprehensive evaluation:

[0249] The calculation result of the first-level fuzzy comprehensive evaluation is B i Arrange it and make it as the secondary fuzzy membership evaluation matrix R i ', based on which the criterion layer membership matrix R = (R1', R'2, ..., R t ') T , according to the weight set W of the route plan under the criterion layer, the final fuzzy comprehensive evaluation result B of the route plan is further calculated, and the calculation formula is:

[0250]

[0251] After obtaining the secondary fuzzy comprehensive evaluation results, the results are further processed with the comment set to obtain the final fuzzy comprehensive evaluation score. The calculation formula is as follows:

[0252]

[0253] f) Scheme optimization, according to the weight of the determined evaluation index and the data information of the preset construction road route scheme optimization index, the gray correlation TOPSIS comprehensive method is used to perform optimization decision calculation on different optional route design schemes, and the optimal route scheme is determined by the relative closeness. In a specific optimization decision embodiment, the method also includes: according to the optimization result of the construction road route scheme, selecting valuable information for display; displaying the relative closeness of the selected scheme, and also displaying the intermediate process data of the basic data acquisition result and the optimization index weight determination result in the above steps;

[0254] The specific processing steps are as follows:

[0255] ① Establish an initial decision matrix

[0256] According to m alternatives and the evaluation indicators of each alternative, an initial decision matrix is ​​constructed:

[0257]

[0258] ② Normalize the initial decision matrix

[0259] The initial decision information matrix is ​​normalized to obtain the normalized matrix V = (V ij ) m×n :

[0260]

[0261] ③Determine positive and negative ideal solutions

[0262] Based on the normalized standard matrix elements and the concept of ideal solutions, positive and negative ideal solutions that meet the conditions are selected:

[0263] Determination of the ideal solution: in:

[0264]

[0265] Determination of negative ideal solution: in:

[0266]

[0267] ④ Calculate the Euclidean distance between the alternative solution and the positive and negative ideal solutions

[0268]

[0269] ⑤ Calculate the grey correlation coefficient and grey correlation degree between the alternative scheme and the positive and negative ideal schemes

[0270]

[0271] In the formula, ρ is the resolution coefficient. Studies have shown that a relatively good resolution can be obtained when its value is 0.5;

[0272] The grey correlation degree of different schemes is solved by the following formula, corresponding to the i-th scheme:

[0273]

[0274] ⑥Optimal ranking of construction road route plans

[0275] Since the Euclidean distance obtained by TOPSIS method and the grey correlation degree obtained by grey correlation method have different dimensions, they are dimensionless for comprehensive consideration:

[0276]

[0277] Let the relative closeness obtained by TOPSIS method be

[0278]

[0279] Let the relative closeness obtained by the grey correlation method be

[0280]

[0281] By organically combining the relative closeness obtained by TOPSIS method and the relative closeness obtained by grey correlation method, the comprehensive closeness can be obtained:

[0282]

[0283] In the formula, α and β are both taken as 0.5, ω i It represents the combined weight vector obtained above. When the comprehensive closeness Q value is larger, the corresponding route plan is better.

[0284] In this embodiment: in step S3, the basic data refers to a plurality of optional construction road route plans in the early preliminary design and planning. These initial input data constituting the construction road route plans will affect the route evaluation and optimization decision results.

[0285] In this embodiment: in step S3, the evaluation and optimization decision-making indicators of the construction road route plan are selected according to the design route data generated by the system, and specifically include the engineering economic indicators, route operation indicators, environmental impact indicators and route plan, and the technical indicators and engineering economic indicators involved in the optimization; wherein the engineering economic indicators in the aforementioned route plan evaluation include route length, slope utilization rate, bridge-tunnel ratio, earthwork quantity per kilometer, and construction cost per kilometer; route operation indicators include route length, number of intersections per kilometer, and number of slope changes per kilometer; environmental impact indicators include bridge-tunnel ratio and earthwork quantity per kilometer; and The technical indicators in the route scheme optimization decision include route length, maximum slope, minimum radius of horizontal curve, minimum radius of vertical curve, number of intersections of horizontal curve and number of intersections of vertical curve; engineering economic indicators include bridge-tunnel ratio, amount of roadbed earthwork, number of culverts, land acquisition area, demolition area and project cost; in order to avoid the disadvantage that the process of determining the advantages and disadvantages of qualitative indicators is interfered by human factors and the evaluation and optimization results are highly subjective, the indicators used in this method are all quantitative indicators, and the route scheme is no longer evaluated and optimized with comments such as "more", "less", "good" and "bad", and the best route scheme is evaluated and optimized based on this.

[0286] An application plug-in was developed based on CAD to demonstrate the application of this method

[0287] 1) Automatic plane generation and optimization

[0288] (I) Linearity determination, such as Figure 3 Shown

[0289] Combined with the topographic map, set the plane optimization parameters and start searching for the best plane alignment in the area, such as Figure 4 shown.

[0290] Plane linear results are displayed, such as Figure 5 and Figure 6 Shown

[0291] (II) Longitudinal section generation and optimization, such as Figure 7-10 Shown

[0292] 3. Program Evaluation and Decision-making

[0293] Compared with the existing schemes: Scheme 1 is generated by this method, and Scheme 2 is the existing scheme data. It can be seen that this method has advantages in designing better lines more quickly, such as Figure 11-Figure 14 shown.

[0294] 2) Automatic generation and optimization of longitudinal sections

[0295] Take a construction road from Qamdo to Linzhi as an example for design optimization. The total length of the line is 11.351km, the road grade is a transportation trunk line, the design speed is 20km / h, and the longitudinal and cross-sectional data can be obtained by interpolating the digital ground model according to the designed plane line shape. It adopts a double lane, the cross-sectional roadbed width is 6.5m, the cross slope of the lane is 2%, the slope of the earth shoulder is 4%, the fill slope is 1:1.5, and the cut slope is 1:0.75; and the longitudinal section design technical indicators can be obtained from the highway grade:

[0296] Table 1 Design technical indicators

[0297]

[0298] The starting point of the longitudinal section design is (0, 4482.840), and the end point is (11351, 4480.510). The start and end point elevation control constraints are as follows:

[0299] Table 2 Start and end elevation constraints

[0300]

[0301] Quick generation of initial longitudinal section plan

[0302] Running the quick generation program for the initial longitudinal section plan can generate the line shape of the initial longitudinal section plan as follows: Fig.15 As shown:

[0303] The relevant data of this line shape are shown in Table 3:

[0304] Table 4 Data of initial longitudinal section plan

[0305]

[0306]

[0307]

[0308] Artificial slope plan

[0309] The initial longitudinal section line is obtained by artificial slope pulling. Fig.16 As shown:

[0310] The relevant data of this line is shown in the following table:

[0311] Table 4 Data of artificial slope pulling scheme for longitudinal section

[0312]

[0313]

[0314]

[0315] Improved genetic algorithm to optimize longitudinal section scheme

[0316] Run the improved genetic algorithm longitudinal section optimization program to obtain the longitudinal section

[0317] The line shape of the initial solution of the existing surface is as follows Fig.17 As shown:

[0318] The relevant data of this line shape is:

[0319] Table 5 Longitudinal section optimization scheme data

[0320]

[0321]

[0322]

[0323] Comparison between manual solution and optimized solution

[0324] The scheme obtained by using the improved genetic algorithm optimization has 34 slope change points (including the start and end points), no bridges or tunnels, and a volume of earthwork of 217,921.723 m 3 The project cost was RMB 435,854.460. The artificial slope plan had 37 slope change points (including the starting and ending points), no bridges or tunnels, and a volume of earthwork of 226,572.87 m 3 The engineering cost is 4531457.40 yuan. It can be seen from the above results that the calculation results of the two methods are relatively close. Therefore, it is feasible to use the improved genetic algorithm to generate and optimize the scheme to provide a reference for the longitudinal section design.

[0325] 3) Route plan evaluation

[0326] In order to verify the practicability of fuzzy comprehensive evaluation method in route plan evaluation and decision-making, two route plans for the newly built construction road were evaluated and decided. The specific contents are as follows:

[0327] a) Import of solution data information

[0328] Import the two route plan data information of the new construction road into the system.

[0329] b) Determine the weight of evaluation indicators

[0330] The importance of the route plan evaluation index criterion layer and the specific evaluation index under the criterion layer is compared pairwise, and the corresponding criterion layer weight and evaluation index weight are given through system calculation, such as Figures 18 to 21 As shown:

[0331] c) Program evaluation and decision-making

[0332] After determining the weights of the route plan evaluation indicators, the system calculates the comprehensive evaluation scores corresponding to the two different route plans based on the fuzzy comprehensive evaluation method, and judges that both plans are within the evaluation range of good and very good based on the comprehensive evaluation scores. Fig. 22 and Fig.23 As shown:

[0333] 4) Route optimization

[0334] In order to verify the practicability of the grey correlation TOPSIS comprehensive method for route plan optimization decision-making, two plans for the new construction road are optimized. The specific contents are as follows:

[0335] a) Import of solution data information

[0336] Import the two route plan data information of the new construction road into the system.

[0337] b) Determine the weight of the preferred indicators

[0338] The importance of the route plan optimization index criterion layer and the specific optimization index under the criterion layer is compared pairwise, and the corresponding criterion layer weight and optimization index weight are given by the system, such as Figure 24 to Figure 26 As shown:

[0339] c) Decision making on optimal solution

[0340] After determining the weights of the route plan optimization indicators, the relative closeness of the two different route plans is obtained through system calculation according to the grey correlation TOPSIS comprehensive method optimization model, and the plan 2 is judged to be better than the plan 1 according to the size of the relative closeness. Fig. 27 As shown:

[0341] Taking into account the technical, economic, operational and environmental factors affecting the route, Option 2, which has a better comprehensive evaluation score and relative closeness, was selected as the recommended option, and the final evaluation and optimization effect was good.

[0342] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0343] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent route selection and evaluation decision-making method for railway engineering construction roads in complex areas, characterized by: The method comprises the following steps: S1. Combined with the terrain and environmental conditions of the construction road and the construction technical standards, the plane alignment mathematical model of the improved HAO is constructed, the plane alignment constraint conditions are designed, the line comprehensive cost model is established as the optimization objective function, and the particle swarm optimization algorithm is used to search for the optimal plane alignment position parameters in the target area; The specific steps include: 1) Obtain geographic information of the design area, including topography and geology, land use, transportation network, hydrological distribution, and building facilities, as basic data for line selection and design; 2) For tight slope sections, including those that require a return curve to reach the target elevation, the maximum slope section design is generally adopted. Combined with the minimum straight line length requirement, a distance l is agreed as the straight line distance of the next intersection point. Combined with the terrain elevation data, the point on the circle with a diameter of l whose height difference slope best meets the maximum longitudinal slope requirement is selected as the next intersection point. This cycle can be repeated to determine the tight slope plane line shape; 3) Divide the starting and ending points of the gentle slope section into n perpendicular bisectors, whose coordinates are represented by O i (x Oi ,y Oi ), the intersection points of the lines can be selected by the plumb line on the equally divided points, and the plane linear position can be determined by the number of intersection points n and the distance d from the intersection point to the equally divided point i The radius R of the peace curve is determined, 4) Use the following formula to calculate the comprehensive cost of the route plan, including the cost of roadbed earthwork C T , roadbed construction costs C J , bridge, culvert engineering cost C Q , land acquisition compensation costs C Z , Demolition compensation costs C C and penalty fee C F ; C=C T +C J +C Q +C Z +C C +C F 5) Using the three plane linear position parameters as particle swarm variables and the line cost as the optimization objective function, randomly generate the initial population variable values, calculate the corresponding initial population objective function values, and record the initial population as the optimal population initial position gBest 0 , record the particle with the best fitness function value in the initial population as the initial best particle position pBest 0 ; 6) Update the variable position and speed of the particle swarm to form a new route plan Where, d is the number of optimization iterations; ω——weight coefficient ——the velocity and position of particle k at the dth iteration; c1, c2 – acceleration coefficients of particle position and velocity, also called learning factors; ——Two random numbers in the interval [0, 1]; ——The best historical position of particle k during the optimization process; gBest d ——The best historical position of the population during the optimization process; When updating the particle speed or position, the value is obtained according to the following formula: 7) Recalculate the fitness function of each particle in the new population and compare it with the corresponding fitness value in the best population recorded. If a better particle position is found, update gBest; similarly, update the best particle pBest; 8) Determine the termination condition. If the condition is not met, return to step e. When the condition is met, terminate the loop; 9) Output the best particle pBest position; S2, based on the roof truss function to smooth the ground line shape, use the intersection of the smooth curve to determine the initial longitudinal section line shape slope change point, set the longitudinal section constraint conditions, build the longitudinal section line shape optimization model optimized by genetic algorithm, and generate the best longitudinal section design scheme; The specific steps are as follows: 1) The fill and cut value on the center line of the longitudinal section, that is, the area formed by the elevation difference between the design elevation and the ground line, is used as the objective function F(X) for quickly generating the longitudinal section. ks ; Take the construction cost as the objective function of longitudinal section intelligent optimization F(X) zn , including roadbed earthwork cost f1(X), bridge cost f2(X), and tunnel cost f3(X); F(X) zn =f1(X)+f2(X)+f3(X) Where, y(x)——longitudinal section design line; h(x)——ground line of longitudinal section; p(x)——weight function, when cross-sectional factors are not considered, p(x)=1; 2) Determine the constraints according to the specifications, taking into account the slope length constraints, slope limit constraints, control point elevation constraints, and vertical curve constraints to complete the model construction; 3) Establish the initial feasible area of ​​the longitudinal section with the slope constraint condition, and preliminarily trim the ground line; use the height weighted average method to process the trimmed ground line into a smooth curve, and generate multiple smooth curves by setting the size of the smooth radius. Their intersection is the inflection point, and the slope section is divided in combination with the ruggedness; 4) The least square method is used to fit the slope line. The intersection of adjacent slope lines is the slope change point of the initial scheme. The resulting scheme is checked and processed for constraints. The vertical curves are configured and the bridges and tunnels are set up to quickly generate the longitudinal section that meets the requirements. 5) The slope change points are randomly distributed by combining the uniform distribution method and the normal distribution method, and the initial population of the longitudinal section scheme is generated after relevant constraints are imposed; 6) Setting the genetic algorithm to optimize the control parameters, using real number coding; 7) The quality of the longitudinal section scheme is evaluated and judged based on the established construction project cost objective function, thereby constructing the longitudinal section scheme fitness function E(i) and the slope section fitness function E(i) k ; E(i)=G max -F i (X) Where, E(i) is the fitness value of the i-th longitudinal section scheme; F i (X)——the objective function value of each longitudinal section scheme; G max ——The maximum value of the objective function in the longitudinal section scheme; Where, E(i) k ——Fitness value of the kth slope section in scheme i; C cub ——the cost of filling and excavation per cubic meter of earth and stone; V (j、cub) ——the fill and cut volume of the jth road section in the kth slope section; L bir , L tun ——The length of the bridge and tunnel in the kth slope section; C bir , C tun ——Cost per linear meter of bridges and tunnels; ΔL k ——the length of the kth slope section; 8) Design genetic operators; use roulette method to select operators; combine arithmetic crossover and heuristic crossover to form a crossover operator, and introduce slope fitness function to improve it, so that the optimization operation is more suitable for longitudinal section design; use adaptive mutation operator to determine the mutation range by fitness; in summary, complete the design and improvement of genetic operators; 9) Set the termination criteria of the genetic algorithm to terminate after reaching the preset maximum number of iterations or evolutionary generations or the cumulative running time reaches 420 minutes; S3. Construct a multi-objective route direction scheme evaluation system, evaluate the designed route scheme, construct a TOPSIS route decision model based on grey correlation, select multiple schemes, and decide on the best scheme; the specific steps are as follows: 1) Data acquisition: obtaining the basic data contained in the construction road route plan; 2) Clarify the indicators and determine the weights of the corresponding evaluation indicators based on the importance comparison judgment matrix of the preset evaluation and optimization decision indicators; 3) Determine the weights. The weights of the corresponding evaluation and optimization indicators can be determined by the least square method of weights, which specifically includes: according to the pairwise judgment matrix between the preset indicators or the pairwise judgment matrix between the indicators determined by the decision maker, further calculate the weight vector of the formed judgment matrix, and use the value of the weight vector as the weight of the corresponding indicator; The specific steps to determine the weights are: (a) Determine the hierarchy of evaluation options This step divides the evaluation and optimization index system into the target layer, the criterion layer and the index layer; (b) Construct a preliminary judgment matrix A = (a ij ) n×n According to the emphasis and cognition of decision makers on the importance of evaluation and optimization indicators, the evaluation and optimization indicators are compared pairwise to obtain a preliminary judgment matrix: (c) Solve matrix B based on the judgment matrix (d) Calculate the sorting vector W by combining the identity matrix e The calculation formula is as follows: e＝(1,1,1,...,1) T The final ranking vector W is the weight vector value of the evaluation index of the corresponding solution; (e) Scheme evaluation, according to the fuzzy membership relationship between the determined weight of the evaluation index and the preset construction road route scheme evaluation index, different optional route design schemes are calculated, and the advantages and disadvantages of the route scheme are determined by the evaluation score; in a specific evaluation embodiment, the method includes: according to the evaluation results of the construction road route scheme, selecting valuable information for display; displaying the overall evaluation score of the selected scheme, and also displaying the intermediate process data such as the basic data acquisition results and the evaluation index weight determination results in the above steps; The specific processing steps are as follows: a) Constructing the evaluation factor set Each evaluation index in the evaluation index system is used as an evaluation factor, with u i To represent it, a set of evaluation factors is constructed based on it, and U is used to represent the set of evaluation factors; that is, U = {U1, U2, ..., U t }, where U i represents the i-th factor class, corresponding to the criterion layer in the evaluation index system, and each U i It also contains a certain number of factors, namely U i = {u i1 ,u i2 ,...,u is }, corresponding to the specific evaluation indicators obtained by refining and decomposing the criterion layer within the evaluation indicator system; b) Determine the review set By analyzing the pros and cons of the evaluation indicators, the evaluation indicator levels are divided into five types, namely, very good, very good, good, average and poor; different evaluation levels correspond to a comprehensive evaluation score, and when the corresponding evaluation indicator meets the range of a corresponding evaluation level, it is allowed to take the comprehensive evaluation value of the corresponding level, and the rest of the situations are the same as above; according to the actual situation of the evaluation indicator system involved, the comment set is set to V, where V = {V1, V2, V3, V4, V5} = {100, 90, 80, 70, 60}; c) Determine the quantitative indicator evaluation level membership The value range of the evaluation index membership is specified to be between 0 and 1. When the membership value of the corresponding evaluation index is larger, it is considered that the index is closer to a certain evaluation level. The evaluation indexes selected here are all quantitative indicators. When the value of a certain evaluation index falls within the corresponding evaluation index level interval, it is considered that its membership value at the corresponding level is 1, and the membership values ​​at other levels are all 0. For example, when the value of the i-th evaluation index satisfies the evaluation level interval of "very good", the membership value of the index at the level of "very good" is 1, and the membership values ​​of the remaining corresponding levels of very good, good, average and poor are all 0. After determining the corresponding membership degree of each evaluation indicator in the evaluation indicator system according to the evaluation indicator level, the membership degree of the evaluation indicator is expressed as R i =(r i1 ,r i2 ,...,r im ),i=1,2,...,n;R i Where n is the number of evaluation indicators for alternative solutions, m is the number of membership levels corresponding to the evaluation indicators for alternative solutions, and r i j represents the membership value of the i-th evaluation index at the j-th membership level; based on this, the corresponding membership matrix R of the scheme to be evaluated can be constructed, which is expressed as follows: d) Determine the weight of evaluation indicators In the evaluation index system, different evaluation indicators have a certain degree of importance, and the importance of the evaluation indicators is usually reflected in the form of weights; the weight of the evaluation indicator reflects its importance in the route plan evaluation index system, and the sum of the weights of all indicators in the evaluation index system should be 1; that is, the weight set W = {w1, w2, ..., w n }middle And any w i ≥0; e) Multi-level fuzzy comprehensive evaluation The evaluation index system can be decomposed into the criterion layer and the indicator layer. The indicator layer is composed of specific evaluation indicators after the criterion layer is refined. Therefore, the fuzzy comprehensive evaluation of the route plan needs to be carried out in two steps; First-level fuzzy comprehensive evaluation: By weighting the specific evaluation indicators of the indicator layer obtained after the detailed decomposition of each criterion layer, the weight set W of the route plan under different indicator layers is obtained. i , and the specific evaluation indicators under each indicator layer can be calculated to obtain the corresponding evaluation indicator membership matrix R i , the fuzzy comprehensive evaluation results B of each criterion layer can be obtained by calculating the weight set of each indicator layer and its membership matrix i ; The calculation formula is: Secondary fuzzy comprehensive evaluation: The calculation result of the first-level fuzzy comprehensive evaluation is B i Arrange it and make it as the secondary fuzzy membership evaluation matrix R′ i , based on which the criterion layer membership matrix R = (R′1, R′2, ..., R′ t ) T , according to the weight set W of the route plan under the criterion layer, the final fuzzy comprehensive evaluation result B of the route plan is further calculated, and the calculation formula is: After obtaining the secondary fuzzy comprehensive evaluation results, the results are further processed with the comment set to obtain the final fuzzy comprehensive evaluation score. The calculation formula is as follows: f) Scheme optimization, according to the weights of the determined evaluation indicators and the data information of the preset construction road route scheme optimization indicators, the grey correlation TOPSIS comprehensive method is used to perform optimization decision calculations on different optional route design schemes, and the optimal route scheme is determined by the relative closeness. In a specific optimization decision embodiment, the method further includes: according to the optimization results of the construction road route scheme, selecting valuable information for display; displaying the relative closeness of the selected scheme, and also displaying the intermediate process data of the basic data acquisition results and the optimization indicator weight determination results in the above steps; The specific processing steps are as follows: ① Establish an initial decision matrix According to m alternatives and the evaluation indicators of each alternative, an initial decision matrix is ​​constructed: ② Normalize the initial decision matrix The initial decision information matrix is ​​normalized to obtain the normalized matrix V = (V ij ) m×n : ③Determine positive and negative ideal solutions Based on the normalized standard matrix elements and the concept of ideal solutions, positive and negative ideal solutions that meet the conditions are selected: Determination of the ideal solution: in: Determination of negative ideal solution: in: ④ Calculate the Euclidean distance between the alternative solution and the positive and negative ideal solutions ⑤ Calculate the grey correlation coefficient and grey correlation degree between the alternative scheme and the positive and negative ideal schemes In the formula, ρ is the resolution coefficient. Studies have shown that a relatively good resolution can be obtained when its value is 0.5; The grey correlation degree of different schemes is solved by the following formula, corresponding to the i-th scheme: ⑥Optimal ranking of construction road route plans Since the Euclidean distance obtained by TOPSIS method and the grey correlation degree obtained by grey correlation method have different dimensions, they are dimensionless for comprehensive consideration: Let the relative closeness obtained by TOPSIS method be Let the relative closeness obtained by the grey correlation method be By organically combining the relative closeness obtained by TOPSIS method and the relative closeness obtained by grey correlation method, the comprehensive closeness can be obtained: In the formula, α and β are both taken as 0.5, ω i It represents the combined weight vector obtained above. When the comprehensive closeness Q value is larger, the corresponding route plan is better.

2. According to claim 1, the intelligent route selection and evaluation decision-making method for railway engineering construction roads in complex areas is characterized by: In step S3, the basic data refers to a plurality of optional construction road route plans in the early preliminary design plan. These initial input data constituting the construction road route plan will affect the route evaluation and optimization decision results.

3. According to claim 1, the intelligent route selection and evaluation decision-making method for railway engineering construction roads in complex areas is characterized by: In step S3, the evaluation and optimization decision indicators of the construction road route plan are selected according to the design route data generated by the system, and specifically include the engineering economic indicators, route operation indicators, environmental impact indicators and route plan, and the technical indicators and engineering economic indicators involved in the optimization; wherein, the engineering economic indicators in the aforementioned route plan evaluation include route length, slope utilization rate, bridge-tunnel ratio, earthwork quantity per kilometer, and construction cost per kilometer; route operation indicators include route length, number of intersections per kilometer, and number of slope changes per kilometer; environmental impact indicators include bridge-tunnel ratio and earthwork quantity per kilometer; and route plan The technical indicators in the optimal decision-making of the scheme include route length, maximum slope, minimum radius of horizontal curve, minimum radius of vertical curve, number of intersections of horizontal curve and number of intersections of vertical curve; the engineering economic indicators include bridge-tunnel ratio, amount of roadbed earthwork, number of culverts, land acquisition area, demolition area and project cost; in order to avoid the disadvantage that the process of determining the advantages and disadvantages of qualitative indicators is interfered by human factors and the evaluation and optimization results are highly subjective, the indicators used in this method are all quantitative indicators, and the route schemes are no longer evaluated and optimized with comments such as "more", "less", "good" and "bad", and the best route scheme is evaluated and optimized based on this.

Images