Heuristic road longitudinal section intelligent generation method and system under multi-constraint condition

Abstract

The invention relates to the field of road route longitudinal section design, and particularly discloses a heuristic road longitudinal section intelligent generation method and system under a multi-constraint condition. According to the method, multiple constraint conditions such as engineering safety, design specifications, terrain adaptability and cost control are comprehensively considered, wherein the cost control is combined with construction and installation cost and constraint damage cost. By adopting an intelligent agent strategy based on a heuristic algorithm, a series of key mark points can be quickly identified and optimized in a preset road route longitudinal design area. The algorithm starts from a preset starting point, performs random exploration by using a normal distribution principle, gradually advances to an end point, and dynamically adjusts the positions and the number of mark points according to constraint conditions and a cost-benefit principle. Finally, the optimized mark points are matched with a vertical curve, and a road longitudinal section design scheme which meets all constraint conditions and is optimal in cost is automatically generated. According to the method, the efficiency and precision of road longitudinal section design are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of vertical alignment design of road routes, and particularly to a heuristic method and system for intelligent generation of road vertical alignment under multiple constraints. Background Art

[0002] In road survey and design, the vertical alignment refers to a vertical section longitudinally and vertically cut along the center line of the road, which expresses the undulating changes along the road. The vertical alignment design mainly determines the size of the longitudinal slope and the elevation of each point according to the nature and grade of the road, the type and driving performance of vehicles, the conditions of the terrain and ground features along the line, the local climate, hydrology, soil conditions, and drainage requirements.

[0003] Currently, the vertical alignment design is mainly manually completed by designers. In the design process, in addition to meeting the requirements of the longitudinal slope size, length, front and rear longitudinal slope conditions, the radius of the vertical curve, and the combination relationship with the horizontal alignment specified in the specifications, it is also necessary to comprehensively consider factors such as the coordination of horizontal and vertical alignments, the layout and proportion of bridges and tunnels, the quantity of earthwork and stonework, and the project cost based on work experience, in order to design an ideal alignment with a reasonable longitudinal slope and a smooth and rounded alignment, so as to achieve a road vertical alignment scheme with safe and comfortable driving and reasonable cost. In addition, there are numerous basic data involved in the process of road vertical alignment design, and various data are interrelated and interact with each other. For example, after the vertical alignment design is modified, the data of structures such as large and medium-sized bridges, small bridges, separated interchanges, overpasses, tunnels, culverts, and channels associated with it need to change accordingly, and the data such as the cross-section slope, side ditch design, and earthwork and stonework quantity will also be adjusted accordingly. The entire vertical alignment design process requires repeated modification, which consumes a large amount of time. Summary of the Invention

[0004] The object of the present invention is to propose a heuristic method and system for intelligent generation of road vertical alignment under multiple constraints. The method comprehensively considers multiple constraints such as engineering safety, design specifications, terrain adaptability, and cost control, where cost control combines construction and installation costs and constraint violation costs. By adopting an intelligent agent strategy based on a heuristic algorithm, a series of key marker points can be quickly identified and optimized within a preset longitudinal design area of the road route. The algorithm starts from a preset starting point, conducts random exploration using the principle of normal distribution, and gradually advances towards the end point. At the same time, the positions and quantities of the marker points are dynamically adjusted according to the constraint conditions and the cost-benefit principle. Finally, these optimized marker points automatically generate a road vertical alignment design scheme that meets all constraint conditions and has the optimal cost through matching vertical curves. The implementation of the present invention significantly improves the efficiency and accuracy of road vertical alignment design, providing strong technical support for road engineering construction.

[0005] To further achieve the above object, the present invention adopts the following technical solutions: A heuristic method for intelligent generation of road vertical profiles under multiple constraints, including formulating the constraints that need to be satisfied for intelligent generation of vertical profiles, where the constraints include engineering safety constraints, specification requirement constraints, terrain condition constraints, and lower total cost constraints, among which:

[0006] The engineering safety constraints include: starting point longitudinal slope constraint S startslope , ending point longitudinal slope constraint S endslope , average slope length constraint L average and reverse slope constraint S reverseslopelimit ;

[0007] The specification requirement constraints include: maximum slope length constraint L max for different longitudinal slopes, minimum slope length constraint L min , maximum longitudinal slope constraint S maxslope , transition slope constraint S moderateslopelimit , bridge slope constraint S bridgelimit and tunnel slope constraint S tunnellimit ;

[0008] The terrain condition constraints include: the highest point of the control point ControlPt up and the lowest point of the control point ControlPt dwon constraining the boundary of the vertical profile; setting the minimum fill height H minbridge for bridge layout and the minimum excavation height H mintunnel for tunnel layout, and according to the actual ground elevation E ground and the designed elevation E of the vertical profile, the intelligent agent automatically arranges bridges and tunnels during the process of generating the vertical profile;

[0009] The lower total cost constraints include: construction and installation cost C ci and constraint violation cost C cb . The intelligent agent generates construction and installation cost C ci during exploration. The construction and installation cost C ci includes the sum of five items: earthwork and stonework construction and installation cost C se , masonry construction and installation cost C m , pavement construction and installation cost C pave , bridge construction and installation cost C bridge and tunnel construction and installation cost C tunnel . During the exploration process of the intelligent agent, the exploration of constraint violation generates corresponding constraint violation cost C cb . According to the principle of lower total cost, the algorithm converges and the generated result finally tends to be optimal.

[0010] Furthermore, the construction and installation cost C ci is calculated according to the following formula:

[0011] Cse = P se × V se ;

[0012] C m = P m × (L road × Q m );

[0013] C pave = P pave × (L road × W road );

[0014] C bridge = P bridge × (L bridge × W sub );

[0015] C tunnel = (P tunnel × C land / 2) × L tunnel ;

[0016] C ci = C se + C m + C pave + C bridge + C tunnel ;

[0017] Wherein, C total is the total cost, C ci is the construction and installation cost, C se is the earthwork and stonework construction and installation cost, C m is the masonry construction and installation cost, C pave is the pavement construction and installation cost, C bridge is the bridge construction and installation cost, C tunnel is the tunnel construction and installation cost, P se represents the price per cubic meter of earthwork and stonework, V se represents the volume of earthwork and stonework, Q m represents the quantity per kilometer of masonry, P m represents the unit price per cubic meter of masonry, P pave represents the price per square meter of pavement, P bridge represents the price per square meter of bridge, L bridge represents the bridge length, P tunnel represents the price per meter of tunnel, L road represents the route length, W road represents the pavement width, W sub represents the subgrade width, C land represents the number of tunnel lanes, L tunnel represents the tunnel length.

[0018] Furthermore, the constraint violation cost C cb is calculated according to the following formula:

[0019] C control = D control × V control

[0020] C reverseslope = S violate × V reverseslope

[0021] C moderateslope = S exceed × V moderateslope

[0022] C bridgetunnel = (S exceedbridge + S exceedtunnel + S lowtunnel ) × V bridgetunnel

[0023] C cb = C control + C reverseslope + C moderateslope + C bridgetunnel

[0024] S violate = |S reverseslope - S reverseslopelimit |

[0025] S exceed = |S moderateslope - S moderateslopelimit |

[0026] When S bridgeslope > S bridgemaxlimit , S exceedbridge = S bridgeslope - S bridgelimit ;

[0027] When S tunnelslope > S tunnelmaxlimit , S exceedtunnel = S tunnelslope - S tunnelmaxlimit ;

[0028] When S tunnelminlimit > S tunnelslope , S lowtunnel = S tunnelminlimit - S tunnelslope ;

[0029] Among them, C control is the control point constraint violation cost, D control is the distance from the control point, and V controlis the control point penalty value, C reverseslope is the reverse slope constraint failure cost, S violate is the reverse slope violation gradient, V reverseslope is the reverse slope penalty value, C moderateslope is the gentle slope constraint failure cost, S exceed is the value exceeding the gentle slope gradient, V moderateslope is the gentle slope penalty value, C bridgetunnel is the bridge-tunnel slope constraint failure cost, S exceedbridge is the value by which the bridge slope exceeds, S exceedtunnel is the value by which the tunnel slope exceeds, S lowtunnel is the value below which the tunnel slope is, V bridgetunnel is the penalty value for violating the bridge-tunnel constraint.

[0030] Furthermore, the total cost C total is the construction and installation cost C ci and the constraint failure cost C cb and is calculated according to the following formula: C total = C ci + C cb .

[0031] Optionally, the agent uses the point marking method to explore. In the point marking method, it is represented by two-dimensional points. The stake number Stake is represented along the X-axis direction, and the longitudinal section design elevation E is represented along the Y-axis direction;

[0032] Based on the set control point data, the agent quickly finds the upper and lower boundaries. The boundaries are composed of the starting point, the ending point, and the control points. The upper and lower boundaries form a closed longitudinal design area. The agent's exploration needs to be carried out within the closed area formed by the upper and lower boundaries and cannot exceed this area;

[0033] Adopting the heuristic algorithm idea, the specific method is:

[0034] According to the route length L road and the maximum slope length L maxslope , quickly find a set of marking points Pts{1...n} with a length of within the longitudinal design area, satisfying the following constraints:

[0035] The marking points in Pts satisfy that is, the exploration elevation range needs to be less than the maximum elevation value of the control points and greater than the minimum elevation value of the control points;

[0036] The calculation formula for the longitudinal slope corresponding to the kth marking point:

[0037] The calculation formula for the slope length corresponding to the kth marking point: L slopelength = Pts[k] stake - Pts[k - 1]stake ;

[0038] Vertical slope constraint between the first set of marked points Pts[1] and Pts[2] in Pts: S slope-1 = S startslope ;

[0039] Vertical slope constraint between the last set of marked points Pts[n - 1] and Pts[n] in Pts: S slope-n = S endslope ;

[0040] Vertical slope constraint between adjacent sets of marked points Pts[k - 1] and Pts[k] in Pts: S slope-k < S maxslope ;

[0041] Slope length constraint between adjacent sets of marked points Pts[k - 1] and Pts[k] in Pts: L slopelength > L minslope , L slopelength < L maxslope ;

[0042] Reverse slope constraint between adjacent sets of marked points Pts[k - 1], Pts[k] and Pts[k + 1] in Pts: S reverseslope < S reverseslopelimit ;

[0043] Transition slope constraint between adjacent sets of marked points Pts[k - 1], Pts[k] and Pts[k + 1] in Pts: S moderateslope < S moderateslopelimit ;

[0044] Difference calculation formula between the designed elevation Pts[k] and the ground elevation Pts[k] for adjacent sets of marked points Pts[k - 1], Pts[k] and Pts[k + 1] in Pts: ΔH = Pts[k] E - Pts[k] Eground ; Bridge layout rule: ΔH > H E- Pts[k] Eground , Tunnel layout rule: -ΔH > H minbridge ; mintunnel ;

[0045] Vertical slope constraint between the marked points Pts[k - 1] and Pts[k] within the bridge range: S bridgeslope < S bridgelimit ;

[0046] Vertical slope constraint between the marked points Pts[k - 1] and Pts[k] within the tunnel range: S tunnelslope < S tunnellimit .

[0047] Further, fix the first and last marker points. Starting from the second marker point Pts[k] where k = 2, the agent makes constraint judgments at points k - 1 and k, and tries to make the k-th marker point follow a normal distribution in the X and Y directions respectively to randomly explore around. The exploration distance in the X direction is Δx * Z1 x , and the exploration distance in the Y direction is Δy * Z1 y ; where the value range of Δx is within the maximum slope length (0, L maxslope ), and the value range of Δy is within the maximum vertical height difference in the longitudinal design area When the X direction is satisfied, the horizontal exploration range of the subsequent marker point Pts[k] x shall not exceed that of the previous marker point Pts[k - 1] x , that is: Pts[k - 1] x < Pts[k] x , and when the Y direction is satisfied, the design height Pts[k] y exploration range of the marker point needs to be within the boundary of the constrained vertical section; Z1 is an independent random number following a standard normal distribution, distributed between (-2, 2); u1 and u2 are two independent uniformly distributed random numbers within the value range of (0, 1);

[0048] The agent explores according to the principle of the lowest total cost during exploration;

[0049] The cost C of the exploration point k' of the current k-th marker point total(k') is less than the current marker C total(k) , replace the position of the exploration point k' with the marker point k, continue the exploration, and loop the exploration process until the exploration cost is at the lowest. The current exploration ends, and the exploration advances towards the end point, that is, k = k + 1;

[0050] To avoid falling into a local optimal solution, the agent will continuously delete and add marker points. The change principle is that when the slope length L slopelength between Pts[n - 1] and Pts[n] is greater than the average slope length L average , insert a set of marker points in the middle and calculate the most appropriate number of marker points;

[0051] The agent continuously explores based on the heuristic algorithm until all marker points that meet the constraint conditions are explored. The set of these points is the variable slope points intelligently generated for the vertical section. Then, according to the route design specifications, vertical curves are matched to form the final intelligent generation result of the vertical section.

[0052] Further, during the exploration process, at least 3 groups of marker points are required to calculate the reverse slope and transition slope constraints.

[0053] A heuristic road vertical alignment intelligent generation system under multiple constraints, which is implemented based on the aforementioned heuristic road vertical alignment intelligent generation method under multiple constraints, includes:

[0054] A constraint module, which is used to establish parametric constraints and obtain constraint values during the generation process;

[0055] A cost module, which is used to calculate the cost during the agent generation process;

[0056] An agent heuristic exploration module, which is used to design exploration algorithms, cost calculation methods, and penalty values for constraint violations, and based on the lower cost principle and avoiding local optima, obtain the intelligent generation result of the road vertical alignment through heuristic exploration.

[0057] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0058] The present invention realizes the parameterization of multiple constraint conditions such as engineering safety, design specifications, terrain adaptability, and cost control, and introduces a heuristic road vertical alignment agent: the agent uses a heuristic algorithm to simplify the vertical alignment design process into the exploration of the neighborhood of adjacent grade change points. Starting from a set of initial vertical alignment schemes, several neighborhood solutions are generated through a neighborhood function under the control of the constraint parameters of the algorithm, and then the exploration parameters are adjusted according to the cost-optimal criterion; repeating the above exploration steps until the convergence criterion of the algorithm is met, and finally obtaining an optimized vertical alignment design result. The present invention can effectively avoid the disorderly exploration and exploration deadlock of the agent and improve the exploration efficiency. Through research and comparison, the present invention can perform overall automatic grade adjustment and local automatic optimization of the vertical alignment. The speed for a 50-km length is within 30 minutes, and the artificial intelligence maturity reaches L1 level, that is, the overall effect is better than that of an inexperienced person. Description of the Drawings

[0059] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0060] Figure 1 It is the algorithm flow block diagram of the present invention;

[0061] Figure 2 It is the agent exploration flow block diagram of the present invention;

[0062] Figure 3 It is the basic information of the vertical alignment design of the present invention;

[0063] Figure 4 It is the constraint setting of the present invention;

[0064] Figure 5 It is the control point constraint setting of the present invention;

[0065] Figure 6 This is the system operation diagram of the present invention. DETAILED DESCRIPTION

[0066] In order to better understand the present invention, various aspects of the present invention will be described in more detail with reference to the accompanying drawings.

[0067] A heuristic intelligent generation method for road longitudinal sections under multiple constraints includes the following steps:

[0068] S001: The intelligent generation process of longitudinal sections must meet certain constraints. The constraints of the present invention are of four types: engineering safety constraints, specification requirements constraints, terrain condition constraints, and lower total cost constraints.

[0069] S002: Engineering safety constraints: used to ensure that the longitudinal section generation results are safe enough to meet the requirements of road use. Specifically including: starting point longitudinal slope constraint S startslope , End point longitudinal slope constraint S endslope , average slope length constraint L average and reverse slope constraint S reverseslopelimit .

[0070] S003: Specification requirement constraints: refers to the requirement that the longitudinal section generation results must comply with relevant specifications and standards to ensure the legality and quality of the project. Specifically including: different longitudinal slope maximum slope length constraints L max , minimum slope length constraint L min , maximum longitudinal slope constraint S maxslope , gentle slope constraint S moderateslopelimit , bridge slope constraint S bridgelimit and tunnel slope constraint S tunnellimit .

[0071] S004: Terrain constraints: refers to the ability of the longitudinal section generation process to adapt to terrain changes and restrictions, specifically by setting the highest control point ControlPt up and the lowest point ControlPt dwon Constrain the boundaries of the longitudinal section; set the minimum fill height H for bridge layout minbridge and the minimum excavation height H for tunnel layout mintunnel , according to the actual ground elevation E ground and the longitudinal section design elevation E. The intelligent agent automatically lays out bridges and tunnels during the longitudinal section generation process.

[0072] S005: Lower total cost Constraint: Total cost C total Including construction cost C ci and constraint violation cost Ccb , C total = C ci + C cb . When the agent explores and generates the vertical section, it calculates the construction and installation cost, and for the exploration that violates the constraints, it generates the corresponding constraint violation cost. It continuously explores according to the principle of lower total cost, making the algorithm converge, and finally the generated result tends to be optimal.

[0073] S006: Construction and installation cost C ci : Includes earthwork and stonework construction and installation cost C se , masonry construction and installation cost C m , pavement construction and installation cost C pave , bridge construction and installation cost C bridge and tunnel construction and installation cost C tunnel , the sum of the five items.

[0074] During the cost calculation: P se represents the price per cubic meter of earthwork and stonework, with the unit of yuan; V se represents the volume of earthwork and stonework, with the unit of cubic meters; Q m represents the quantity of masonry per kilometer, with the unit of cubic meters; P m represents the unit price per cubic meter of masonry, with the unit of yuan; P pave represents the price per square meter of pavement, with the unit of yuan; P bridge represents the price per square meter of bridge, with the unit of yuan; L bridge represents the length of the bridge, with the unit of meters; P tunnel represents the price per meter of tunnel, with the unit of yuan; L road represents the route length, with the unit of meters; W road represents the pavement width, with the unit of meters; W sub represents the subgrade width, with the unit of meters; C land represents the number of tunnel lanes, which is a constant; L tunnel represents the tunnel length, with the unit of meters.

[0075] The specific method is as follows:

[0076] C se = P se × V se ;

[0077] C m = P m × (L road × Q m );

[0078] C pave = P pave × (L road × W road );

[0079] C bridge=P bridge ×(L bridge ×W sub );

[0080] C tunnel =(P tunnel ×C land / 2)×L tunnel ;

[0081] C ci =C se +C m +C pave +Cb r id ge +C tunnel ;

[0082] S007: Total cost of constraint violation C cb : Includes the cost of constraint violation C of the control point control , the cost of constraint violation C of the reverse slope reverseslope , the cost of constraint violation C of the transition slope moderateslope , the cost of constraint violation C of the bridge-tunnel slope bridgetunnel : The sum is C cb =C control +C reverseslope +C moderateslope +C bridgetunnel , in yuan.

[0083] Explanation of penalty value: The starting value of each type of penalty value is the constant 1. During the exploration process, whenever a corresponding constraint is violated, the penalty value base of the corresponding type becomes 10 times the current penalty value until the optimal solution is found in the current exploration. Then the penalty value returns to 1 and the next exploration begins;

[0084] S008: Cost of constraint violation C of the control point control

[0085] When the agent encounters a control point and violates the control point constraint during exploration, the agent tries to satisfy the control point constraint and moves in the direction of reducing the degree of constraint violation, so as to finally satisfy the control point constraint. In order to make the agent move in the direction of reducing the degree of constraint violation, the degree of constraint violation needs to be quantified. Therefore, the degree of constraint violation is expressed by cost. When the control point constraint is violated, the distance D between the current position and the control point is calculated control , and this distance is multiplied by a penalty value V control [[ID=BO]]as the cost and added to the total cost. In this way, the agent explores in the direction of less constraint violation according to the principle of cost reduction during exploration.

[0086] C control =D control [[ID=7B]]×V control .

[0087] S009: Reverse slope constraint violation cost C reverseslope

[0088] The reverse slope constraint is the same as the control point constraint. When the two slope directions before and after the vertical curve grade change point are inconsistent, determine which slope S reverseslope violates the reverse slope constraint S reverseslopelimit , and it will violate the slope S violate multiplied by the penalty value V reverseslope and added to the total cost.

[0089] C reverseslope = S violate ×V reverseslope ;

[0090] S violate = |S reverseslope - S reverseslopelimit |.

[0091] S010: Transition slope constraint violation cost C moderateslope

[0092] The transition slope constraint is that when two consecutive slopes exceed a certain value, a transition slope must be added in the middle. That is, when two consecutive slopes S moderateslope exceed the transition slope constraint S moderateslopelimit , the exceeded value S exceed is multiplied by the penalty value V moderateslope and added to the total cost.

[0093] C moderateslope = S exceed ×V moderateslope ;

[0094] S exceed = |S moderateslope - S moderateslopelimit |.

[0095] S011: Bridge and tunnel slope constraint violation cost C bridgetunnel

[0096] During the process of the agent generating the vertical alignment, bridges and tunnels are automatically arranged. The limit values of the bridge and tunnel slopes S bridgemaxlimit are 4%, S tunnelminlimit is 0.3%, S tunnelmaxlimit is 3% (the relevant specification values need to be obtained (stipulated in 8.2.4 and 8.2.5 of "Highway Route Design Specification" (JTGD20 - 2017)), that is: the longitudinal slope on large and medium bridges should not be greater than 4%, and the longitudinal slope in tunnels should be greater than 0.3% and less than 3%. When the longitudinal slope is on a bridge, the longitudinal slope S slope is the bridge longitudinal slope S bridgeslopeand the longitudinal slope is greater than 4%, multiply the difference exceeding 4% by the penalty value V bridgetunnel and add it to the total cost. When the longitudinal slope is on the tunnel, the longitudinal slope S slope is the longitudinal slope S of the tunnel tunnelslope and the longitudinal slope is less than 0.3% or greater than 3%, multiply the difference less than 0.3% or greater than 3% by the penalty value V bridgetunnel and add it to the total cost.

[0097] C bridgetunnel =(S exceedbridge +S exceedtunnel +S lowtunnel )×V bridgetunnel ;

[0098] When S bridgeslope >S bridgemaxlimit , S exceedbridge =S bridgeslope -S bridgemaxlimit ;

[0099] When S tunnelslope >S tunnelmaxlimit , S exceedtunnel =S tunnelslope -S tunnelmaxlimit ;

[0100] When S tunnelminlimit >S tunnelslope , S lowtunnel =S tunnelminlimit -S tunnelslope .

[0101] Among them, C bridgetunnel is the cost of bridge-tunnel slope constraint violation, S exceedbridge is the excess value of the bridge slope, S exceedtunnel is the excess value of the tunnel slope, S lowtunnel is the value below the tunnel slope, V bridgetunnel is the penalty value for violating the bridge-tunnel constraint.

[0102] S012: The agent of the present invention mainly uses the point marking method for exploration; in the point marking method, it is represented by two-dimensional points. The stake number Stake is represented along the X-axis direction, and the designed elevation E of the longitudinal section is represented along the Y-axis direction;

[0103] S013: According to the control point data set in step S004, the agent quickly finds the upper and lower boundaries, that is, the upper boundary ControlPts up and the lower boundary ControlPts down . The boundaries are composed of the starting point, the ending point and the control points. The upper and lower boundaries form a closed longitudinal design area, and the agent exploration needs to be carried out within the closed area formed by the upper and lower boundaries;

[0104] The intersection point of adjacent different gradient lines on the vertical section is called the grade change point, and the longitudinal curve set at the grade change is the vertical curve. S slope is the longitudinal gradient, and L slopelength is the slope length. According to the route design line specification, the slope length has the design requirements of the maximum slope length L maxslope and the minimum slope length L minslope .

[0105] S014: The intelligent agent uses the point marking method to quickly find a set of marking points Pts{1...n} with a length of road within the route vertical section design range according to the route length L maxslope and the maximum slope length L , satisfying the following constraints:

[0106] S015: The marking points in Pts satisfy that is, the explored elevation range needs to be less than the maximum elevation value of the control points and greater than the minimum elevation value of the control points;

[0107] The calculation formula for the longitudinal gradient corresponding to the kth marking point:

[0108] The calculation formula for the slope length corresponding to the kth marking point: L slopelength = Pts[k] stake - Pts[k - 1] stake ;

[0109] S016: The longitudinal gradient constraint between the first group of marking points Pts[1] and Pts[2] in Pts: S slope-1 = S startslope ;

[0110] S017: The longitudinal gradient constraint between the last group of marking points Pts[n - 1] and Pts[n] in Pts: S slope-n = S endslope ;

[0111] S018: The longitudinal gradient constraint between adjacent groups of marking points Pts[k - 1] and Pts[k] in Pts: S slope-k < S maxslope ;

[0112] S019: The slope length constraint between adjacent groups of marking points Pts[k - 1] and Pts[k] in Pts: L slopelength > L minslope , L slopelength < L maxslope ;

[0113] S020: The reverse slope constraint between adjacent groups of marking points Pts[k - 1], Pts[k] and Pts[k + 1] in Pts: S reverseslope<S reverseslopelimit ;

[0114] S021: Transition slope constraint between a group of adjacent marked points Pts[k - 1], Pts[k] and Pts[k + 1] in Pts: S moderateslope <S moderateslopelimit ;

[0115] S022: Designed height Pts[k] between a group of adjacent marked points Pts[k - 1], Pts[k] and Pts[k + 1] in Pts E and ground height Pts[k] Eground Difference calculation formula: ΔH = Pts[k] E- Pts[k] Eground , Bridge layout rule: ΔH > H minbridge ; Tunnel layout rule: -ΔH > H mintunnel ;

[0116] Minimum height H of the bridge to be laid minbridge and minimum height H of the tunnel mintunnel are set by the user through the interactive operation interface;

[0117] S023: Longitudinal slope constraint between marked points Pts[k - 1] and Pts[k] within the bridge range: S bridgeslope <S[[ID=�7]] bridgelimit ;

[0118] S024: Longitudinal slope constraint between marked points Pts[k - 1] and Pts[k] within the tunnel range: S tunnelslope <S tunnellimit ;

[0119] S025: Fix the first and last marked points. The agent starts from the second marked point (i.e., k = 2 in Pts[k]) and makes constraint judgments at points k - 1 and k. Try to make the k-th marked point randomly explore around in the X and Y directions according to the normal distribution The exploration distance in the X direction is Δx * Z1 x , and the exploration distance in the Y direction is Δy * Z1 y ; Among them, the value range of Δx is within the maximum slope length range (0, L maxslope ), and the value range of Δy is within the maximum vertical design area height difference range In the X direction, the horizontal exploration range of the subsequent marked point Pts[k] x shall not exceed that of the previous marked point Pts[k - 1] x , that is: Pts[k - 1] x < Pts[k] x , and in the Y direction, the designed height Pts[k] that satisfies the marked point explorationy The range needs to be within the boundary of the constrained vertical profile; Z1 is an independent random number that follows a standard normal distribution, mainly distributed between (-2, 2); u1 and u2 are two independent uniformly distributed random numbers within the value range of (0, 1); calculate the construction and installation cost according to the method of step S006;

[0120] S026: Exploration process, calculate the construction and installation cost according to step S006, and calculate the constraint violation cost for the exploration of constraint violation according to steps S007, S008, S009, S010, and S011;

[0121] S027: The cost C of the exploration point k' of the current k-th marked point total(k') is less than the current mark C total(k) , replace the position of the exploration point k' with the marked point k, continue the exploration, loop the exploration process until the exploration cost is the lowest, the current exploration ends, and the exploration advances towards the end point, that is, k = k + 1;

[0122] S028: In order to avoid falling into local optimal solutions, the agent will continuously delete and add marked points. The change principle is that the slope length L between Pts[n - 1] and Pts[n] slopelength is greater than the average slope length L average , insert a set of marked points in the middle and calculate the most appropriate number of marked points;

[0123] S029: At least three groups of marked points are required to calculate the reverse slope and transition slope constraints;

[0124] S030: When the agent can no longer optimize the existing plan, convert the marked points into grade change points, obtain the vertical curve according to the route design specifications, and combine them into the final intelligent generation result of the vertical profile;

[0125] S031: For the grade change points of the vertical profile automatically generated by the agent, they can be evaluated manually. For the unreasonably evaluated intervals, the constraint conditions can be adjusted according to steps S002, S003, S004, and S005, and then the corresponding algorithm is executed. If adding or reducing grade change points is executed, the algorithm will intercept the interval and then execute the above steps S013 - S030.

[0126] Example 1:

[0127] Taking the vertical profile design of a certain first-class highway in the plain area as an example, the total designed mileage of this project is about 26 km, the designed speed is 80 km / h, the roadbed width is 24.5 m, and the intelligent generation process of the vertical profile refers to Figure 1 :

[0128] (1) Input the engineering safety constraint parameters, including the starting longitudinal slope constraint, the ending longitudinal slope constraint, the average slope length constraint, and the reverse slope constraint. For specific parameter settings, refer to Figure 3 and Figure 4 ;

[0129] (2) Obtain the constraint parameters required by the specification according to the design speed, including the maximum slope length constraint, the minimum slope length constraint, the maximum longitudinal slope constraint, the transition slope constraint, and the bridge-tunnel slope constraint for different longitudinal slopes obtained according to the design speed. For specific parameter settings, refer to Tables 1 - 3;

[0130] (3) Obtain the terrain condition constraint parameters of the design document and parameterize them into control points to control the elevation range of the generated results;

[0131] (4) Input the prices and calculation rules of each construction and installation cost, including the construction and installation costs of earthwork, masonry, pavement, bridges, and tunnels. The calculation rules refer to step S006. For specific parameter settings, refer to Figure 4 ;

[0132] (5) Formulate the calculation rule for the cost of violating the control point constraint. For specific parameter settings, refer to Figure 5 , and refer to step S008;

[0133] (6) Formulate the calculation rule for the cost of violating the reverse slope constraint, referring to step S009;

[0134] (7) Formulate the calculation rule for the cost of violating the transition slope constraint, referring to step S010;

[0135] (8) Formulate the calculation rule for the cost of violating the bridge-tunnel slope constraint, referring to step S011;

[0136] (9) The agent explores using the point marking method. In the point marking method, it is represented by two-dimensional points. The abscissa represents the station number along the X-axis, and the ordinate represents the longitudinal section design elevation along the Y-axis;

[0137] (10) Use the set parameters in steps (1) and (3) to quickly find the longitudinal design area: the upper and lower boundaries from the starting point to the ending point, and this boundary only meets the maximum longitudinal slope constraint and the control point constraint;

[0138] (11) The agent quickly finds a set of points that meet the boundary constraints between the starting point and the ending point using the point marking method;

[0139] (12) Starting from the second marked point set, randomly explore around. During the exploration process, continuously calculate the cost. The calculation process refers to Figure 2 ;

[0140] (13) When the cost of the current exploration point is at a lower position, the current exploration ends; the exploration advances towards the ending point;

[0141] (14) To avoid falling into local optima, the agent continuously deletes and adds marked points to calculate an optimal number of marked points.

[0142] (15) When the agent can no longer optimize the existing solution, it converts the marked points into grade break points and then matches the corresponding vertical curve radius according to the vertical curve specification corresponding to the design speed. Referring to Table 4, the combination of grade break points and vertical curves is the final generated result of the vertical profile.

[0143] (16) For the vertical profile results automatically generated by the agent, as Figure 6 shown, they can be evaluated manually. For unreasonably evaluated intervals, the constraint conditions can be adjusted according to steps (2), (3), and (4), and then the corresponding algorithm can be executed.

[0144] Table 1 Maximum longitudinal slope

[0145] Design speed (km / h) 120 100 80 60 40 30 20 Maximum longitudinal slope (%) 3 4 5 6 7 8 9

[0146] Table 2 Minimum slope length

[0147] Design speed (km / h) 120 100 80 60 40 30 20 Minimum slope length (m) 300 250 200 150 120 100 60

[0148] Table 3 Maximum slope length (m) for different longitudinal slopes

[0149]

[0150]

[0151] Table 4 Minimum vertical curve radius and vertical curve length

[0152]

[0153] It should be understood that the above description of the preferred embodiment is relatively detailed and should not be considered as a limitation on the protection scope of the invention patent. Under the inspiration of the present invention, those of ordinary skill in the art can also make substitutions or deformations without departing from the protection scope defined by the claims of the present invention, and all fall within the protection scope of the present invention. The scope of protection claimed by the present invention shall be subject to the appended claims.

Claims

1. A heuristic method for intelligent generation of road vertical profiles under multiple constraint conditions, including formulating the constraint conditions that need to be satisfied for intelligent generation of vertical profiles, characterized in that, The constraint conditions include engineering safety constraints, specification requirement constraints, terrain condition constraints, and lower total cost constraints, where: The engineering safety constraints include: starting point longitudinal slope constraint S startslope , ending point longitudinal slope constraint S endslope , average slope length constraint L average and reverse slope constraint S reverseslopelimit ; The specification requirements and constraints include: the maximum slope length constraint L for different longitudinal slopes max , the minimum slope length constraint L min , the maximum longitudinal slope constraint S maxslope , the transition slope constraint S moderateslopelimit , the bridge slope constraint S bridgelimit and the tunnel slope constraint S tunnellimit ; The terrain condition constraints include: the highest point of the control point ControlPt up and the lowest point of the control point ControlPt dwon constraining the boundaries of the vertical section; setting the minimum fill height H minbridge for bridge layout and the minimum excavation height H mintunnel for tunnel layout, and automatically laying out bridges and tunnels during the process of the intelligent agent generating the vertical section according to the actual ground elevation E ground and the designed elevation E of the vertical section; The total cost lower constraint includes: construction and installation cost C ci and constraint violation cost C cb . When the agent explores, it generates construction and installation cost C ci . The construction and installation cost C ci includes the earthwork and stonework construction and installation cost C se , the dock construction and installation cost C m , the road surface construction and installation cost C pave , the bridge construction and installation cost C bridge and the tunnel construction and installation cost C tunnel , which is the sum of the five items. During the exploration process of the agent, the exploration of constraint violation generates the corresponding constraint violation cost C cb . According to the principle of lower total cost, the algorithm converges, and the generated result finally tends to be optimal.

2. The heuristic method for intelligent generation of road vertical profiles under multiple constraints according to claim 1, characterized in that The construction cost C ci is calculated according to the following formula: C se = P se × V se ; C m = P m × (L road × Q m ); C pave = P pave × (L road × W road ); C bridge = P bridge × (L bridge × W sub ); C tunnel = (P tunnel × C land / 2) × L tunnel ; C ci = C se + C m + C pave + C bridge + C tunnel ; Among them, C total is the total cost, C ci is the construction and installation cost, C se is the earthwork and stonework construction and installation cost, C m is the masonry construction and installation cost, C pave is the pavement construction and installation cost, C bridge is the bridge construction and installation cost, C tunnel is the tunnel construction and installation cost, P se is the price per cubic meter of earthwork and stonework, V se is the volume of earthwork and stonework, Q m is the quantity per kilometer of masonry, P m is the unit price per cubic meter of masonry, P pave is the price per square meter of pavement, P bridge is the price per square meter of bridge, L bridge is the length of the bridge, P tunnel is the price per meter of tunnel, L road is the route length, W road is the pavement width, W sub is the subgrade width, C land is the number of lanes in the tunnel, L tunnel is the tunnel length.

3. The heuristic method for intelligent generation of road vertical profiles under multiple constraints according to claim 2, characterized in that The constraint breaking cost C cb is calculated according to the following formula: C control = D control × V control C reverseslope = S violate × V reverseslope C moderateslope = S exceed × V moderateslope C bridgetunnel = (S exceedbridge + S exceedtunnel + S lowtunnel ) × V bridgetunnel C cb = C control + C reverseslope + C moderateslope + Cb r id getunne l S violate = |S reverseslope - S reverseslopelimit | S exceed = |S moderateslope - S moderateslopelimit | When S bridgeslope > S bridgemaxlimit , S exceedbridge = S bridgeslope - S bridgemaxlimit ; When S tunnelslope > S tunnelmaxlimit , S exceedtunnel = S tunnelslope - S tunnelmaxlimit ; When S tunnelminlimit > S tunnelslope then S lowtunnel = S tunnelminlimit - S tunnelslope ; Among them, C control is the cost of control point constraint damage, D control is the distance from the control point, V control is the control point penalty value, C reverseslope is the cost of reverse slope constraint damage, S violate is the reverse slope violation gradient, V reverseslope is the reverse slope penalty value, C moderateslope is the cost of transition slope constraint damage, S exceed is the value exceeding the transition slope gradient, V moderateslope is the transition slope penalty value, C bridgetunnel is the cost of bridge and tunnel slope constraint damage, S exceedbridge is the value by which the bridge slope exceeds, S exceedtunnel is the value by which the tunnel slope exceeds, S lowtunnel is the value by which the tunnel slope is lower, V bridgetunnel is the penalty value for damaging the bridge and tunnel constraints.

4. The heuristic method for intelligent generation of road vertical profiles under multiple constraints according to claim 3, wherein The total cost C total is the construction and installation cost C ci and the constraint failure cost C cb The sum is calculated according to the following formula: C total = C ci + C cb .

5. The heuristic road vertical profile intelligent generation method under multiple constraint conditions according to claim 1, wherein, The agent uses the point marking method for exploration. In the point marking method, it is represented by two-dimensional points. The stake number Stake is represented along the X-axis direction, and the longitudinal section design elevation E is represented along the Y-axis direction. Based on the set control point data, the agent quickly finds the upper and lower boundaries. The boundaries are composed of a starting point, an ending point, and control points. The upper and lower boundaries form a closed longitudinal design area. The agent's exploration needs to be carried out within the closed area formed by the upper and lower boundaries and cannot exceed this area. Adopt the heuristic algorithm idea, and the specific method is as follows: According to the route length L road and the maximum slope length L maxslope , quickly find a set of marker points Pts{1...n} with a length of in the longitudinal design area, satisfying the following constraints: The marked points in Pts satisfy that is, the elevation range to be explored needs to be less than the maximum elevation value of the control points and greater than the minimum elevation value of the control points; Longitudinal slope calculation formula corresponding to the k-th marked point: Calculation formula for slope length corresponding to the k-th marked point: L slopelength = Pts[k] stake - Pts[k - 1] stake ; Longitudinal slope constraint between the first set of marked points Pts[1] and Pts[2] in Pts: S slope-1 = S startslope ; Longitudinal slope constraint between the last set of marked points Pts[n - 1] and Pts[n] in Pts: S slope-n = S endslope ; Longitudinal slope constraint between a set of adjacent marked points Pts[k - 1] and Pts[k] in Pts: S slope-k <S maxslope ; Slope length constraint between a group of adjacent marker points Pts[k - 1] and Pts[k] in Pts: L slopelength > L minslope , L slopelength <L maxslope ; Reverse slope constraint between a set of adjacent marked points Pts[k - 1], Pts[k], and Pts[k + 1] in Pts: S reverseslope <S reverseslopelimit ; Transition slope constraint between a set of adjacent marked points Pts[k-1], Pts[k] and Pts[k+1] in Pts: S moderateslope <S moderateslopelimit ; Design a high Pts[k] among a group of adjacent marked points Pts[k-1], Pts[k] and Pts[k+1] in Pts E With the ground high Pts[k] Eground Difference calculation formula: ΔH = Pts[k] E- Pts[k] Eground , Bridge layout rule: ΔH > H minbridge ; Tunnel layout rule: -ΔH > H mintunnel ; Longitudinal slope constraint between the marked points Pts[k-1] and Pts[k] within the bridge range: Sb r id ges l ope <Sb r id ge li m i t ; Longitudinal slope constraint between the marked points Pts[k-1] and Pts[k] within the tunnel range: S tunnelslope <S tunnellimit 。 6. The heuristic method for intelligent generation of road vertical profile under multiple constraints according to claim 5, characterized in that Fix the first and last marker points. Starting from the second marker point Pts[k] where k = 2, the agent makes constraint judgments at points k - 1 and k, and attempts to make the k-th marker point follow a normal distribution in the X and Y directions respectively to randomly explore around. The exploration distance in the X direction is Δx * Z1 x , and the exploration distance in the Y direction is Δy * Z1 y ; where the value range of Δx is within the maximum slope length (0, L maxslope ), and the value range of Δy is within the maximum vertical height difference in the longitudinal design area (0, ); in the X direction, the horizontal exploration range of the subsequent marker point Pts[k] x should not exceed that of the previous marker point Pts[k - 1] x , that is: Pts[k - 1] x < Pts[k] x , and in the Y direction, the design height Pts[k] y exploration range of the marker point needs to be within the boundary of the constrained vertical section; Z1 is an independent random number following the standard normal distribution, distributed between (-2, 2); u1 and u2 are two independent uniformly distributed random numbers within the value range of (0, 1); When exploring, the agent explores in accordance with the principle of the lowest total cost. The exploration point k' cost C of the current k-th marked point total(k') Less than the current mark C total(k) , replace the position of the exploration point k' with the marked point k, continue the exploration, and loop the exploration process until the exploration cost is at the lowest. The current exploration ends, and the exploration advances towards the end point, i.e., k = k + 1; In order to avoid falling into local optimal solutions, the agent continuously deletes and adds marked points. The change principle is the slope length L between Pts[n - 1] and Pts[n]. slopelength Greater than the average slope length L average When this occurs, a set of marked points is inserted in the middle, and the most appropriate number of marked points is calculated. Based on the heuristic algorithm, the agent continuously explores until all marked points that meet the constraint conditions are explored. The set of these points is the variable slope points generated by the intelligent longitudinal section. Then, according to the route design specifications, vertical curves are matched to form the final intelligent generation result of the longitudinal section.

7. The heuristic road vertical profile intelligent generation method under multiple constraint conditions according to claim 6, characterized in that, During the exploration process, at least 3 groups of marked points are required to calculate the reverse slope and transition slope constraints.

8. A heuristic road vertical profile intelligent generation system under multiple constraints, which is implemented based on the heuristic road vertical profile intelligent generation method under multiple constraints described in any one of claims 1-7, and is characterized in that, Including: A constraint module for establishing parametric constraints and obtaining constraint values during the generation process. A cost module for calculating cost expenses during the agent generation process. An agent heuristic exploration module for designing exploration algorithms, cost calculation methods, and punishment values for constraint violations, and combining the lower cost principle and avoiding local optimality to obtain the intelligent generation result of the road longitudinal section based on heuristic exploration.

Images