CN113554467A - Railway three-dimensional linear intelligent design method based on co-evolution - Google Patents

Railway three-dimensional linear intelligent design method based on co-evolution Download PDF

Info

Publication number
CN113554467A
CN113554467A CN202110848678.5A CN202110848678A CN113554467A CN 113554467 A CN113554467 A CN 113554467A CN 202110848678 A CN202110848678 A CN 202110848678A CN 113554467 A CN113554467 A CN 113554467A
Authority
CN
China
Prior art keywords
sub
individuals
slope
dimensional linear
group
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110848678.5A
Other languages
Chinese (zh)
Other versions
CN113554467B (en
Inventor
缪鹍
周启航
冯倩
戴炎林
李洵贻
李挺
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Central South University
Original Assignee
Central South University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Central South University filed Critical Central South University
Priority to CN202110848678.5A priority Critical patent/CN113554467B/en
Publication of CN113554467A publication Critical patent/CN113554467A/en
Application granted granted Critical
Publication of CN113554467B publication Critical patent/CN113554467B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data
    • G06Q30/0206—Price or cost determination based on market factors
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/29—Geographical information databases
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Strategic Management (AREA)
  • Development Economics (AREA)
  • Theoretical Computer Science (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Data Mining & Analysis (AREA)
  • Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Human Resources & Organizations (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Tourism & Hospitality (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Quality & Reliability (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Operations Research (AREA)
  • Remote Sensing (AREA)
  • Feedback Control In General (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a railway three-dimensional linear intelligent design method based on coevolution, which comprises the steps of firstly establishing two three-dimensional linear design objective functions, and simultaneously solving the quantity and the parameters of plane intersection points and the quantity and the parameters of slope changing points of a three-dimensional linear based on an ArcGIS platform by adopting a coevolution differential algorithm according to the line grade of a railway, the constraint conditions of planes and longitudinal sections and unit price cost to realize the railway three-dimensional linear intelligent design; the three-dimensional linear shape of the intelligent design method can comprehensively consider engineering cost and operation cost, well avoid obstacles and adapt to terrains, and provide early design reference for engineering designers.

Description

Railway three-dimensional linear intelligent design method based on co-evolution
Technical Field
The invention relates to the field of railway line selection, in particular to a railway three-dimensional line intelligent design method based on coevolution.
Background
The intelligent design of three-dimensional linear shape of railway means that the intelligent calculation method is adopted to solve the constructed three-dimensional linear model so as to realize the automatic wiring and automatic design of the three-dimensional linear shape of railway and simultaneously require the minimum objective function of the scheme. Since its objective function is obtained by interpolation of the digital elevation model and involves various costs, the function is generally not trivial. When the railway line is long and has more intersection points and slope changing points, numerous decision variables need to be involved in the process of three-dimensional linear intelligent design, numerous complex factors in project construction are considered, and the factors interact and mutually restrict, so that the railway three-dimensional linear intelligent design is a large-scale complex problem. In the prior art, the intelligent design of the line selection problem is usually to solve all decision variables of a line simultaneously, so that the calculation efficiency is low, and even the local optimal solution is easy to fall into.
The differential algorithm has excellent global optimization performance, better parallelism and robustness, can execute evolution operation according to the distribution condition of individuals in a group, better solves the problem of early convergence of the algorithm, but has slower convergence speed and depends very on the adopted variation strategy and parameter setting. Meanwhile, the coevolution framework can carry out grouping solution on the variables, and efficiently solves the problem of large-scale variable optimization.
Because the existing railway three-dimensional linear intelligent design method can not directly identify, analyze and utilize geographic information, the line selection result is not satisfactory, and the practical application is difficult to realize. The Geographic Information System (GIS) has strong spatial retrieval and analysis capability, can reasonably and quickly process and analyze complex and various surface feature, terrain and landform in the railway three-dimensional linear intelligent design process, and can be visually expressed. However, the prior art does not provide a railway three-dimensional linear intelligent design method combining GIS and co-evolution.
Disclosure of Invention
In order to solve the problems in the existing route selection design, the invention aims to provide a railway three-dimensional linear intelligent design method based on coevolution. The method utilizes the excellent performance of the coevolution in the large-scale multi-target design problem to more efficiently solve the problem of complex route selection design.
The intelligent design method for the three-dimensional linear shape of the railway based on the coevolution comprises the steps of firstly establishing two target functions for the three-dimensional linear shape design, and simultaneously solving the quantity and the parameters of the intersection points of the three-dimensional linear shape based on the ArcGIS platform and the quantity and the parameters of the slope changing points of the longitudinal section by adopting a coevolution differential algorithm according to the line grade of the railway, the constraint conditions of the plane and the longitudinal section and the unit price cost, so that the intelligent design of the three-dimensional linear shape of the railway is realized.
The number and parameters of the three-dimensional linear plane intersection points and the number and parameters of the vertical section slope changing points refer to the number, coordinates, radiuses and easement curves of the plane intersection points, the number, coordinates, radiuses and easement curves of the vertical section intersection points;
the three-dimensional linear plane intersection point and vertical section slope-changing point parameters refer to plane coordinates (x, y), radius R and slow length l of the plane intersection point, and vertical section slope-changing point mileage lzAnd elevation z.
The three-dimensional linear intelligent design method of the railway based on the coevolution is characterized by comprising the following steps:
step 1: respectively establishing two target functions F of railway three-dimensional line shape1And F2：
Figure BDA0003179025000000021
F2＝KN·LCp+365·(NH+η·NK)·∑ε1 (2)
Wherein, F1The engineering cost index is a railway three-dimensional linear engineering cost index; l is the cross section serial number; b is the serial number of the bridge; t is the serial number of the tunnel; l is the number of cross sections in the linear design interval; b is the number of bridge seats in the linear design interval; t is the number of the tunnel seats in the design interval; cLThe unit price of the earthwork project is; cBThe unit price of each linear meter of the bridge; cTThe unit price of each linear meter of the tunnel; d is the distance between two adjacent cross sections; a. thelFor the first one is providedMeasuring the area of the cross section; a. thel+1The area of the designed cross section is the (l + 1) th; l isbThe length of the b-th bridge; l istThe length of the t-th seat tunnel;
F2the operation cost index is a railway three-dimensional linear operation cost index; kNIs an unrelated rate; LC (liquid Crystal)pThe total length of the line; n is a radical ofHThe number of the rows of the trucks is; eta is a travel cost conversion coefficient of the travelling train; n is a radical ofKThe number of passenger train rows; sigma epsilon1The sum of the running fees of each freight train for one round trip;
step 2: setting a three-dimensional linear plane constraint condition, a longitudinal section constraint condition and a three-dimensional linear design target area constraint condition;
(1) plane constraint condition
(a) Number of intersections constraint, number of intersections n of three-dimensional line shapepThe specification or a human-defined allowed number should be met, i.e.: n ispmin≤np≤npmax；
Wherein n ispminA minimum number of intersections specified for specification or human; n ispmaxA maximum number of intersections specified for normative or artificial;
(b) minimum circular curve radius constraint, plane intersection point PiRadius of the circular curve RJDiNot less than the minimum circular curve radius RminNamely: rmin≤RJDi(i＝1，2，…，np-1)；
(c) Minimum circular curve length constraint, plane intersection point PiLength L of the circular curveRiNot less than the minimum circular curve length LRminNamely: l isRmin≤LRi＝RJDi·|αi|(i＝1，2，…，np-1)；
Wherein alpha isiIs a plane intersection point PiThe steering angle of (d);
(d) minimum relaxed curve length constraint, plane intersection point PiLength of the relaxation curve l0JDiNot less than the minimum relief curve length l0minNamely: l0min≤l0JDi(i＝1，2，…，np-1)；
(e) Minimum sizeThe length of the clamping line is restricted, and the length L of the clamping line between two adjacent easement curves at different intersection pointsjiNot less than the minimum clip line length LjminNamely: l isjmin≤Lji＝Li，i+1-Ti-Ti+1(i＝1，2，…，np-1)；
Wherein L isi，i+1Is a plane intersection point PiTo the plane intersection point Pi+1The distance of (d); t isiIs a plane intersection point PiThe tangent length of (2); t isi+1Is a plane intersection point Pi+1The tangent length of (2);
(2) longitudinal section constraint condition
(a) The number of the variable slope points is restrained, and the number of the three-dimensional linear variable slope points is nzThe specification or a human-defined allowed number should be met, i.e.: n iszmin≤nz≤nzmax；
Wherein n iszminThe minimum number of the grade changing points is specified by a standard or a person; n iszmaxThe maximum number of the grade changing points is specified for the standard or the human;
(b) minimum slope segment length constraint, distance L between two adjacent slope change pointspjNot less than the minimum slope length LpminNamely: l ispmin≤Lpj＝XBPDj+1-XBPDj(j＝1，2，…，nz-1)；
Wherein, XBPDjTo change the slope point HjMileage of (1); xBPDj+1To change the slope point Hj+1Mileage of (1);
(c) longitudinal section slope constraint, slope i of two adjacent points of variationBPDjThe allowed gradient should be met, either by specification or by human regulation, i.e.:
Figure BDA0003179025000000031
wherein iminA minimum slope segment length specified for regulatory or human; i.e. imaxMaximum length of slope, Z, specified for regulation or for human purposesBPDjTo change the slope point HjElevation of (d); zBPDj+1To change the slope point Hj+1Elevation of (d);
(d) adjacent to each otherThe slope difference of the slope section is restrained, and the slope difference of the adjacent slope sections is delta iBPDjNot greater than a standard or artificially specified maximum gradient difference Δ imaxNamely: Δ iBPDj＝|iBPDj-iBPDj-1|≤Δimax；
(3) The three-dimensional linear design target area constraint, the three-dimensional linear design target area (x, y), should satisfy the range specified by the human, namely: x is the number ofmin≤x≤xmax，ymin≤y≤ymax
And step 3: searching for an optimal linear shape from a three-dimensional linear design target area by adopting a coevolution differential algorithm;
step a: setting population size NP, variation factors, cross factors, greedy factors, maximum function evaluation times and an iteration cycle of random grouping;
step b: taking the number and parameters of all plane intersection points and the number and parameters of slope-changing points of the longitudinal section on the three-dimensional line as individuals, wherein the individuals comprise n-dimensional variables, and randomly distributing the n-dimensional variables to s groups Pi(i ═ 1,2, …, s), each group containing m dimensions of variables, each group being assigned to children of size NP, each group being initialized according to the upper and lower limit values corresponding to the m dimensions of the variables for which each group is responsible for optimization;
step c: according to an inter-group cooperation mode, each sub-individual is cooperated with sub-individuals of other groups, the fitness value of the sub-individuals in each group is calculated by adopting a railway three-dimensional linear objective function, each group is subjected to rapid non-dominant sorting based on the fitness value, the non-dominant grade of each sub-individual is obtained, and the congestion distance value of each sub-individual in each group is calculated at the same time;
the specific implementation manner of inter-group cooperation is as follows:
because the sub-individuals in each sub-population only contain partial dimensionality of the three-dimensional linear variable, the fitness evaluation cannot be independently completed, and the sub-individuals in a certain sub-population must cooperate with other sub-populations to complete the fitness evaluation, namely, inter-group cooperation:
f(i，j)＝f(P1y，…Pj-1y，z，Pj+1y，…Psy) (3)
wherein f (i, j) represents the fitness of the jth individual in the ith sub-population; pjy represents the position of a random one of the other sub-populations at the highest non-dominant level; z represents the current position of the jth sub-individual in the ith sub-population;
step d: for all groups Pi(i-1, 2, …, s) performing a mutation operation and a crossover operation of a differential algorithm to generate a child group Qi(i-1, 2, …, s), calculating the fitness value of each sub-group neutron individual according to the target function of the railway three-dimensional linear shape by adopting an inter-group cooperation mode, and then updating an external file set;
wherein, the external archive set is updated as follows:
1) if the current group PiOf (2) dominates the progeny group QiSelecting the current group PiThe child individuals enter an external archive set; 2) if the child group QiDominates the current group PiSelecting a progeny group QiThe child individuals enter an external archive set; 3) if the child group QiAnd the current group PiThe two sub-individuals are selected and stored in an external file set;
step e: performing non-domination sorting on all the sub-individuals of the external archive set, calculating the non-domination level and the crowded distance value of each sub-individual, selecting NP sub-individuals as a new group according to the non-domination level and the crowded distance value of the sub-individuals, and entering next generation for updating;
wherein the non-dominant ordering is as follows:
1) for all the sub-individuals, finding out all non-dominated sub-individuals according to a Pareto domination relation, using the non-dominated sub-individuals as a first-level non-dominated layer, and recording the non-dominated ranking values Rank of the sub-individuals as 1;
2) removing the individuals which have obtained the non-dominant grade values from the whole external archive set, and carrying out next round of comparison on the remaining sub-individuals to obtain a second-level non-dominant layer; and so on until all individuals have obtained their non-dominant rank values;
step f: judging whether the current function evaluation times reach the maximum function evaluation times, if the current function evaluation times meet the termination condition, obtaining a final complete non-dominated solution set containing all variable dimensions from the optimal front edge (the sub-individual with the domination level of 1) of the external archive set, outputting the solution set, obtaining the optimal three-dimensional linear shape of the railway, and ending; if the algorithm termination condition is not met, judging whether the current function evaluation times meet the iteration period of the random grouping, if so, executing the updating group number s, returning to the step d to continue updating, and if not, directly returning to the step d to continue updating.
The invention provides a railway three-dimensional linear intelligent design method based on coevolution, which can be known by analyzing the test result of railway three-dimensional linear intelligent design of a research area as follows: the invention can fast and efficiently select and solve the line of the plane and the vertical section of the line, and the three-dimensional line shape of the intelligent design method can comprehensively consider the engineering cost and the operation cost, better avoid the obstacle and adapt to the terrain, and provide the early design reference for engineering designers. In addition, the invention has the following beneficial effects:
(1) the scheme has good diversity and high efficiency. And finally, generating an optimal scheme group which comprises different characteristics such as a bypass scheme and a bridge and tunnel erecting scheme instead of a single scheme. By utilizing the co-evolution, the diversity of the intelligent designed circuit scheme is ensured, more than selection schemes with reference values are provided for circuit designers, and the omission of feasible schemes due to subjective factors of the designers is avoided.
(2) Is suitable for large-scale and large-range line selection. The GIS has strong space analysis capability and mass data processing capability, can conveniently and quickly process large-scale landforms, and provides a strong support system for large-scale line selection. The three-dimensional line selection problem of the large-scale variable can be divided into a plurality of low-dimensional sub-problems by the co-evolution, and then each sub-problem is solved independently, so that the calculation efficiency of the three-dimensional line selection problem of the large-scale variable is improved.
Drawings
FIG. 1 is a flow chart of co-evolution;
FIG. 2 is a schematic diagram of an example terrain plan;
FIG. 3 is a flow chart of model optimization;
FIG. 4 is a schematic illustration of inter-group collaboration;
FIG. 5 is a schematic diagram of a three-dimensional linear design scheme set.
Detailed Description
The invention is further illustrated with reference to the following figures and examples:
the invention provides a railway three-dimensional linear intelligent design method based on coevolution, which comprises the steps of firstly establishing two three-dimensional linear design objective functions, and simultaneously solving the quantity and the parameters of plane intersection points and the quantity and the parameters of slope changing points of a three-dimensional linear based on an ArcGIS platform by adopting a coevolution differential algorithm according to the line grade of a railway, the constraint conditions of planes and longitudinal sections and unit price cost, so as to realize the railway three-dimensional linear intelligent design.
The number and parameters of the three-dimensional linear plane intersection points and the number and parameters of the vertical section slope changing points refer to the number, coordinates, radiuses and easement curves of the plane intersection points, the number, coordinates, radiuses and easement curves of the vertical section intersection points;
the three-dimensional linear plane intersection point and vertical section slope-changing point parameters refer to plane coordinates (x, y), radius R and slow length l of the plane intersection point, and vertical section slope-changing point mileage lzAnd elevation z.
The co-evolution flow chart is shown in fig. 1.
The example covers 220.8km with the east of Chongqing City in China2And taking the mountainous terrain with the altitude of 255-1445 meters as a research area, and testing the developed road three-dimensional linear intelligent design method, as shown in fig. 1. The specific topographic data of the research area is TIFF files of the topography acquired from the Google Earth, then Global Mapper is adopted to convert the TIFF files into contour lines in DWG format, and finally Arcmap is utilized to convert line elements of the contour lines into point elements which can be identified by ArcGIS.
The line selection is to apply the invention to design a three-level highway with the length of about 12km in the line selection area, and as shown in fig. 2, the line selection area can be seen as a mountainThe mountains and hills are heavy, the landforms and hills are fluctuated, and the gullies are vertical and horizontal, so that no bad geological condition exists. Setting parameters according to specific conditions: the speed per hour is designed to be 80 km/h; the range of the intersection point of the planes is [5, 15 ]](ii) a The minimum curve radius is 500 m; the minimum circular curve length is 50 m; a minimum relief curve length of 60 m; the minimum clip line length is 50 m; the range of the longitudinal gradient point is [5, 15 ]](ii) a The minimum slope section length is 200 m; maximum limit slope 22%; the maximum longitudinal slope difference is 22%; the unit price of the filled earthwork is 65 yuan/m3(ii) a The unit price of the square earth and stone is 35 yuan/m3(ii) a The unit price of the bridge is 30000 yuan/m2(ii) a The unit price of the tunnel is 28000 yuan/m2。
Fig. 3 shows a flow chart of model optimization.
Step 1: respectively establishing two target functions F of railway three-dimensional line shape1And F2：
Figure BDA0003179025000000061
F2＝KN·LCp+365·(NH+η·NK)·∑ε1 (5)
Wherein, F1The engineering cost index is a railway three-dimensional linear engineering cost index; l is the cross section serial number; b is the serial number of the bridge; t is the serial number of the tunnel; l is the number of cross sections in the linear design interval; b is the number of bridge seats in the linear design interval; t is the number of the tunnel seats in the design interval; cLThe unit price of the earthwork project is; cBThe unit price of each linear meter of the bridge; cTThe unit price of each linear meter of the tunnel; d is the distance between two adjacent cross sections; a. thelThe area of the first design cross section; a. thel+1The area of the designed cross section is the (l + 1) th; l isbThe length of the b-th bridge; l istThe length of the t-th seat tunnel;
F2the operation cost index is a railway three-dimensional linear operation cost index; kNIs an unrelated rate; LC (liquid Crystal)pThe total length of the line; n is a radical ofHThe number of the rows of the trucks is; eta is a travel cost conversion coefficient of the travelling train; n is a radical ofKThe number of passenger train rows; sigma epsilon1To and from each freight trainSum of one-time running cost;
step 2: setting a three-dimensional linear plane constraint condition, a longitudinal section constraint condition and a three-dimensional linear design target area constraint condition;
(1) plane constraint condition
(a) Number of intersections constraint, number of intersections n of three-dimensional line shapepThe specification or a human-defined allowed number should be met, i.e.: n ispmin≤np≤npmax；
Wherein n ispminA minimum number of intersections specified for specification or human; n ispmaxA maximum number of intersections specified for normative or artificial;
(b) minimum circular curve radius constraint, plane intersection point PiRadius of the circular curve RJDiNot less than the minimum circular curve radius RminNamely: rmin≤RJDi(i＝1，2，…，np-1)；
(c) Minimum circular curve length constraint, plane intersection point PiLength L of the circular curveRiNot less than the minimum circular curve length LRminNamely: l isRmin≤LRi＝RJDi·|αi|(i＝1，2，…，np-1)；
Wherein alpha isiIs a plane intersection point PiThe steering angle of (d);
(d) minimum relaxed curve length constraint, plane intersection point PiLength of the relaxation curve l0IDiNot less than the minimum relief curve length l0minNamely: l0min≤l0JDi(i＝1，2，…，np-1)；
(e) Minimum clip line length constraint, clip line length L between two adjacent easement curves at different intersection pointsjiNot less than the minimum clip line length LjminNamely: l isjmin≤Lji＝Li，i+1-Ti-Ti+1(i＝1，2，…，np-1)；
Wherein L isi，i+1Is a plane intersection point PiTo the plane intersection point Pi+1The distance of (d); t isiIs a plane intersection point PiIs tangent toLength; t isi+1Is a plane intersection point Pi+1The tangent length of (2);
(2) longitudinal section constraint condition
(a) The number of the variable slope points is restrained, and the number of the three-dimensional linear variable slope points is nzThe specification or a human-defined allowed number should be met, i.e.: n iszmin≤nz≤nzmax；
Wherein n iszminThe minimum number of the grade changing points is specified by a standard or a person; n iszmaxThe maximum number of the grade changing points is specified for the standard or the human;
(b) minimum slope segment length constraint, distance L between two adjacent slope change pointspjNot less than the minimum slope length LpminNamely: l ispmin≤Lpj＝XBPDj+1-XBPDj(j＝1，2，…，nz-1)；
Wherein, XBPDjTo change the slope point HjMileage of (1); xBPDj+1To change the slope point Hj+1Mileage of (1);
(c) longitudinal section slope constraint, slope i of two adjacent points of variationBPDjThe allowed gradient should be met, either by specification or by human regulation, i.e.:
Figure BDA0003179025000000071
wherein iminA minimum slope segment length specified for regulatory or human; i.e. imaxMaximum length of slope, Z, specified for regulation or for human purposesBPDjTo change the slope point HjElevation of (d); zBPDj+1To change the slope point Hj+1Elevation of (d);
(d) the slope difference of the adjacent slope sections is restrained, and the slope difference of the adjacent slope sections is delta iBPDjNot greater than a standard or artificially specified maximum gradient difference Δ imaxNamely: Δ iBPDj＝|iBPDj-iBPDj-1|≤Δimax；
(3) The three-dimensional linear design target area constraint, the three-dimensional linear design target area (x, y), should satisfy the range specified by the human, namely: x is the number ofmin≤x≤xmax，ymin≤y≤ymax
And step 3: searching for an optimal linear shape from a three-dimensional linear design target area by adopting a coevolution differential algorithm;
step a: setting population size NP, variation factors, cross factors, greedy factors, maximum function evaluation times and an iteration cycle of random grouping;
step b: taking the number and parameters of all plane intersection points and the number and parameters of slope-changing points of the longitudinal section on the three-dimensional line as individuals, wherein the individuals comprise n-dimensional variables, and randomly distributing the n-dimensional variables to s groups Pi(i ═ 1,2, …, s), each group containing m dimensions of variables, each group being assigned to children of size NP, each group being initialized according to the upper and lower limit values corresponding to the m dimensions of the variables for which each group is responsible for optimization;
step c: according to an inter-group cooperation mode, each sub-individual is cooperated with sub-individuals of other groups, the fitness value of the sub-individuals in each group is calculated by adopting a railway three-dimensional linear objective function, each group is subjected to rapid non-dominant sorting based on the fitness value, the non-dominant grade of each sub-individual is obtained, and the congestion distance value of each sub-individual in each group is calculated at the same time;
the specific implementation manner of inter-group cooperation is as follows:
because the sub-individuals in each sub-population only contain partial dimensionality of the three-dimensional linear variable, the fitness evaluation cannot be independently completed, and the sub-individuals in a certain sub-population must cooperate with other sub-populations to complete the fitness evaluation, namely, inter-group cooperation:
f(i，j)＝f(P1y，…Pj-1y，z，Pj+1y，…Psy) (6)
wherein f (i, j) represents the fitness of the jth individual in the ith sub-population; pjy represents the position of a random one of the other sub-populations at the highest non-dominant level; z represents the current position of the jth sub-individual in the ith sub-population;
step d: for all groups Pi(i-1, 2, …, s) performing a mutation operation and a crossover operation of a difference algorithm to generate offspringGroup Qi(i-1, 2, …, s), calculating the fitness value of each sub-group neutron individual according to the target function of the railway three-dimensional linear shape by adopting an inter-group cooperation mode, and then updating an external file set;
wherein, the external archive set is updated as follows:
1) if the current group PiOf (2) dominates the progeny group QiSelecting the current group PiThe child individuals enter an external archive set; 2) if the child group QiDominates the current group PiSelecting a progeny group QiThe child individuals enter an external archive set; 3) if the child group QiAnd the current group PiThe two sub-individuals are selected and stored in an external file set;
step e: performing non-domination sorting on all the sub-individuals of the external archive set, calculating the non-domination level and the crowded distance value of each sub-individual, selecting NP sub-individuals as a new group according to the non-domination level and the crowded distance value of the sub-individuals, and entering next generation for updating;
wherein the non-dominant ordering is as follows:
1) for all the sub-individuals, finding out all non-dominated sub-individuals according to a Pareto domination relation, using the non-dominated sub-individuals as a first-level non-dominated layer, and recording the non-dominated ranking values Rank of the sub-individuals as 1;
2) removing the individuals which have obtained the non-dominant grade values from the whole external archive set, and carrying out next round of comparison on the remaining sub-individuals to obtain a second-level non-dominant layer; and so on until all individuals have obtained their non-dominant rank values;
step f: judging whether the current function evaluation times reach the maximum function evaluation times, if the current function evaluation times meet the termination condition, obtaining a final complete non-dominated solution set containing all variable dimensions from the optimal front edge (the sub-individual with the domination level of 1) of the external archive set, outputting the solution set, obtaining the optimal three-dimensional linear shape of the railway, and ending; if the algorithm termination condition is not met, judging whether the current function evaluation times meet the iteration period of the random grouping, if so, executing the updating group number s, returning to the step d to continue updating, and if not, directly returning to the step d to continue updating.
FIG. 4 is a schematic diagram of inter-group collaboration.
After the basic parameters of line selection are set, the three-dimensional linear design scheme set of the line is generated by inputting the coordinates of the starting point and the coordinates of the end point and adopting the method provided by the invention, as shown in fig. 5. The scheme set comprises 26 schemes which can avoid plane elevation obstacles well.
Through the steps, information of the whole line can be obtained, and table 2 shows the target values and partial information of the first seven schemes of the obtained 26 schemes; taking the first solution as an example, the plane intersection data of the first solution is shown in table 3, and the vertical section gradient point data of the first solution is shown in table 4.
TABLE 2 target values and partial information for seven schemes
Figure BDA0003179025000000091
Table 3 plan one intersection data
Figure BDA0003179025000000092
Table 4 scheme one vertical section slope point changing data
Figure BDA0003179025000000101

Claims (3)

1. A railway three-dimensional linear intelligent design method based on coevolution is characterized in that two three-dimensional linear design objective functions are established, and according to the line grade of a railway, the constraint conditions of planes and longitudinal sections and unit price cost, a coevolution differential algorithm is adopted to simultaneously solve the quantity and parameters of plane intersection points and the quantity and parameters of longitudinal section variable slope points of three-dimensional linear based on an ArcGIS platform, so that the railway three-dimensional linear intelligent design is realized;
the number and parameters of the three-dimensional linear plane intersection points and the number and parameters of the vertical section slope changing points refer to the number, coordinates, radiuses and easement curves of the plane intersection points, the number, coordinates, radiuses and easement curves of the vertical section intersection points;
the three-dimensional linear plane intersection point and vertical section slope changing point parameters refer to plane coordinates (x, y), radius R and slow length l of the plane intersection point, and vertical section slope changing point mileage lzAnd elevation z.
2. The cooperative evolution-based railway three-dimensional linear intelligent design method according to claim 1, characterized by comprising the following steps:
step 1: respectively establishing two target functions F of railway three-dimensional line shape1And F2：
Figure FDA0003179024990000011
F2＝KN·LCp+365·(NH+η·NK)·∑ε1 (2)
Wherein, F1The engineering cost index is a railway three-dimensional linear engineering cost index; l is the cross section serial number; b is the serial number of the bridge; t is the serial number of the tunnel; l is the number of cross sections in the linear design interval; b is the number of bridge seats in the linear design interval; t is the number of the tunnel seats in the design interval; cLThe unit price of the earthwork project is; cBThe unit price of each linear meter of the bridge; cTThe unit price of each linear meter of the tunnel; d is the distance between two adjacent cross sections; a. thelThe area of the first design cross section; a. thel+1The area of the designed cross section is the (l + 1) th; l isbThe length of the b-th bridge; l istThe length of the t-th seat tunnel;
F2the operation cost index is a railway three-dimensional linear operation cost index; kNIs an unrelated rate; LC (liquid Crystal)pThe total length of the line; n is a radical ofHThe number of the rows of the trucks is; eta is a travel cost conversion coefficient of the travelling train; n is a radical ofKThe number of passenger train rows; sigma epsilon1The sum of the running fees of each freight train for one round trip;
step 2: setting a three-dimensional linear plane constraint condition, a longitudinal section constraint condition and a three-dimensional linear design target area constraint condition;
(1) plane constraint condition
(a) Number of intersections constraint, number of intersections n of three-dimensional line shapepThe specification or a human-defined allowed number should be met, i.e.: n ispmin≤np≤npmax；
Wherein n ispminA minimum number of intersections specified for specification or human; n ispmaxA maximum number of intersections specified for normative or artificial;
(b) minimum circular curve radius constraint, plane intersection point niRadius of the circular curve RJDiNot less than the minimum circular curve radius RminNamely: rmin≤RJDi(i＝1,2,…,np-1)；
(c) Minimum circular curve length constraint, plane intersection point PiLength L of the circular curveRiNot less than the minimum circular curve length LRminNamely: l isRmin≤LRi＝RJDi·|αi|(i＝1,2,…,np-1)；
Wherein alpha isiIs a plane intersection point PiThe steering angle of (d);
(d) minimum relaxed curve length constraint, plane intersection point PiLength of the relaxation curve l0JDiNot less than the minimum relief curve length l0minNamely: l0min≤l0JDi(i＝1,2,…,np-1)；
(e) Minimum clip line length constraint, clip line length L between two adjacent easement curves at different intersection pointsjiNot less than the minimum clip line length LjminNamely: l isjmin≤Lji＝Li,i+1-Ti-Ti+1(i＝1,2,…,np-1)；
Wherein L isi,i+1Is a plane intersection point PiTo the plane intersection point Pi+1The distance of (d); t isiIs a plane intersection point PiThe tangent length of (2); t isi+1Is a plane intersection point Pi+1The tangent length of (2);
(2) longitudinal section constraint condition
(a) The number of the variable slope points is restrained, and the number of the three-dimensional linear variable slope points is nzThe specification or a human-defined allowed number should be met, i.e.: n iszmin≤nz≤nzmax；
Wherein n iszminThe minimum number of the grade changing points is specified by a standard or a person; n iszmaxThe maximum number of the grade changing points is specified for the standard or the human;
(b) minimum slope segment length constraint, distance L between two adjacent slope change pointspjNot less than the minimum slope length LpminNamely: l ispmin≤Lpj＝XBPDj+1-XBPDj(j＝1,2,…,nz-1)；
Wherein, XBPDjTo change the slope point HjMileage of (1); xBPDj+1To change the slope point Hj+1Mileage of (1);
(c) longitudinal section slope constraint, slope i of two adjacent points of variationBPDjThe allowed gradient should be met, either by specification or by human regulation, i.e.:
Figure FDA0003179024990000021
wherein iminA minimum slope segment length specified for regulatory or human; i.e. imaxMaximum length of slope, Z, specified for regulation or for human purposesBPDjTo change the slope point HjElevation of (d); zBPDj+1To change the slope point Hj+1Elevation of (d);
(d) the slope difference of the adjacent slope sections is restrained, and the slope difference of the adjacent slope sections is delta iBPDjNot greater than a standard or artificially specified maximum gradient difference Δ imaxNamely: Δ iBPDj＝|iBPDj-iBPDj-1|≤Δimax；
(3) The three-dimensional linear design target area constraint, the three-dimensional linear design target area (x, y), should satisfy the range specified by the human, namely: x is the number ofmin≤x≤xmax，ymin≤y≤ymax
And step 3: searching for an optimal linear shape from a three-dimensional linear design target area by adopting a coevolution differential algorithm;
step a: setting population size NP, variation factors, cross factors, greedy factors, maximum function evaluation times and an iteration cycle of random grouping;
step b: taking the number and parameters of all plane intersection points and the number and parameters of slope-changing points of the longitudinal section on the three-dimensional line as individuals, wherein the individuals comprise n-dimensional variables, and randomly distributing the n-dimensional variables to s groups Pi(i ═ 1,2, …, s), each group containing m dimensions of variables, each group being assigned to children of size NP, each group being initialized according to the upper and lower limit values corresponding to the m dimensions of the variables for which each group is responsible for optimization;
step c: according to an inter-group cooperation mode, each sub-individual is cooperated with sub-individuals of other groups, the fitness value of the sub-individuals in each group is calculated by adopting a railway three-dimensional linear objective function, each group is subjected to rapid non-dominant sorting based on the fitness value, the non-dominant grade of each sub-individual is obtained, and the congestion distance value of each sub-individual in each group is calculated at the same time;
the specific implementation manner of inter-group cooperation is as follows:
because the sub-individuals in each sub-population only contain partial dimensionality of the three-dimensional linear variable, the fitness evaluation cannot be independently completed, and the sub-individuals in a certain sub-population must cooperate with other sub-populations to complete the fitness evaluation, namely, inter-group cooperation:
f(i,j)＝f(P1y,…Pj-1y,z,Pj+1y,…Psy) (3)
wherein f (i, j) represents the fitness of the jth individual in the ith sub-population; pjy represents the position of a random one of the other sub-populations at the highest non-dominant level; z represents the current position of the jth sub-individual in the ith sub-population;
step d: for all groups Pi(i-1, 2, …, s) performing a mutation operation and a crossover operation of a differential algorithm to generate a child group Qi(i 1,2, …, s) root in an inter-group collaborationCalculating the fitness value of each sub-group neutron individual according to the target function of the railway three-dimensional linear shape, and then updating an external archive set;
wherein, the external archive set is updated as follows:
1) if the current group PiOf (2) dominates the progeny group QiSelecting the current group PiThe child individuals enter an external archive set; 2) if the child group QiDominates the current group PiSelecting a progeny group QiThe child individuals enter an external archive set; 3) if the child group QiAnd the current group PiThe two sub-individuals are selected and stored in an external file set;
step e: performing non-domination sorting on all the sub-individuals of the external archive set, calculating the non-domination level and the crowded distance value of each sub-individual, selecting NP sub-individuals as a new group according to the non-domination level and the crowded distance value of the sub-individuals, and entering next generation for updating;
wherein the non-dominant ordering is as follows:
1) for all the sub-individuals, finding out all non-dominated sub-individuals according to a Pareto domination relation, using the non-dominated sub-individuals as a first-level non-dominated layer, and recording the non-dominated ranking values Rank of the sub-individuals as 1;
2) removing the individuals which have obtained the non-dominant grade values from the whole external archive set, and carrying out next round of comparison on the remaining sub-individuals to obtain a second-level non-dominant layer; and so on until all individuals have obtained their non-dominant rank values;
step f: judging whether the current function evaluation times reach the maximum function evaluation times, if the current function evaluation times meet the termination condition, obtaining a final complete non-dominated solution set containing all variable dimensions from the optimal front edge (the sub-individual with the domination level of 1) of the external archive set, outputting the solution set, obtaining the optimal three-dimensional linear shape of the railway, and ending; if the algorithm termination condition is not met, judging whether the current function evaluation times meet the iteration period of the random grouping, if so, executing the updating group number s, returning to the step d to continue updating, and if not, directly returning to the step d to continue updating.
3. The method for designing the three-dimensional linear shape of the railway based on the coevolution as claimed in claim 2, wherein the variation operation and the crossover operation of the difference algorithm of step d in the step 3 are as follows:
1) mutation operation: the differential algorithm generates variant individuals for each individual in the population through a variant strategy, namely:
Figure FDA0003179024990000041
wherein the content of the first and second substances,
Figure FDA0003179024990000042
represents the ith individual
Figure FDA0003179024990000043
Variant individuals at passage G;
Figure FDA0003179024990000044
are randomly selected individuals from the current population,
Figure FDA0003179024990000045
is an individual randomly selected from the current population and an external archive set;
Figure FDA0003179024990000046
is an individual randomly selected from an algorithm non-dominated solution set; f is a variation factor; r, rnd are two between [0, 1 ]]A random number of ranges;
2) and (3) cross operation: after performing mutation operation to obtain a variant individual, combining the target individual and the variant individual by adopting a binomial crossover operator to obtain an offspring individual, namely:
Figure FDA0003179024990000047
wherein randi,j() Is [0, 1 ]]Random numbers, j, uniformly distributed over the rangerandIs [0, n]Random integers uniformly distributed in the range, and n is the dimension of the variable.
CN202110848678.5A 2021-07-26 2021-07-26 Railway three-dimensional linear intelligent design method based on co-evolution Active CN113554467B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110848678.5A CN113554467B (en) 2021-07-26 2021-07-26 Railway three-dimensional linear intelligent design method based on co-evolution

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110848678.5A CN113554467B (en) 2021-07-26 2021-07-26 Railway three-dimensional linear intelligent design method based on co-evolution

Publications (2)

Publication Number Publication Date
CN113554467A true CN113554467A (en) 2021-10-26
CN113554467B CN113554467B (en) 2022-07-22

Family

ID=78132878

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110848678.5A Active CN113554467B (en) 2021-07-26 2021-07-26 Railway three-dimensional linear intelligent design method based on co-evolution

Country Status (1)

Country Link
CN (1) CN113554467B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114154382A (en) * 2021-11-18 2022-03-08 中国公路工程咨询集团有限公司 High-resolution remote sensing three-dimensional visual railway route selection auxiliary decision making system and method
CN116305678A (en) * 2022-12-08 2023-06-23 西南交通大学 Fine optimization method for high-density urban railway line based on low-carbon emission
CN116796915A (en) * 2023-04-18 2023-09-22 中南大学 Proxy model method for intelligent railway route selection
CN117708961A (en) * 2024-02-05 2024-03-15 中南大学 Integrated intelligent reconstruction method for three-dimensional space line position of existing railway
CN117726196A (en) * 2024-02-18 2024-03-19 中南大学 Comprehensive evaluation method for environment suitability for railway route selection

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101872376A (en) * 2010-06-04 2010-10-27 中国公路工程咨询集团有限公司 True three-dimensional road intelligent design method and system
CN103399849A (en) * 2013-06-24 2013-11-20 中南大学 Road three-dimensional linear automatic optimization method based on improved particle swarm optimization
CN105069255A (en) * 2015-08-31 2015-11-18 中南大学 Intersection position model construction method for three-dimensional highway or railway route selection
CN110820580A (en) * 2019-11-18 2020-02-21 广州揽睿路桥设计有限公司 Bridge rotation construction three-dimensional linear control technology based on BIM technology

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101872376A (en) * 2010-06-04 2010-10-27 中国公路工程咨询集团有限公司 True three-dimensional road intelligent design method and system
CN103399849A (en) * 2013-06-24 2013-11-20 中南大学 Road three-dimensional linear automatic optimization method based on improved particle swarm optimization
CN105069255A (en) * 2015-08-31 2015-11-18 中南大学 Intersection position model construction method for three-dimensional highway or railway route selection
CN110820580A (en) * 2019-11-18 2020-02-21 广州揽睿路桥设计有限公司 Bridge rotation construction three-dimensional linear control technology based on BIM technology

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
BO LIU: ""A Co-evolutionary Differential Evolution Algorithm for Constrained Optimization"", 《THIRD INTERNATIONAL CONFERENCE ON NATURAL COMPUTATION (ICNC 2007)》 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114154382A (en) * 2021-11-18 2022-03-08 中国公路工程咨询集团有限公司 High-resolution remote sensing three-dimensional visual railway route selection auxiliary decision making system and method
CN116305678A (en) * 2022-12-08 2023-06-23 西南交通大学 Fine optimization method for high-density urban railway line based on low-carbon emission
CN116305678B (en) * 2022-12-08 2023-12-15 西南交通大学 Fine optimization method for high-density urban railway line based on low-carbon emission
CN116796915A (en) * 2023-04-18 2023-09-22 中南大学 Proxy model method for intelligent railway route selection
CN116796915B (en) * 2023-04-18 2024-01-12 中南大学 Proxy model method for intelligent railway route selection
CN117708961A (en) * 2024-02-05 2024-03-15 中南大学 Integrated intelligent reconstruction method for three-dimensional space line position of existing railway
CN117708961B (en) * 2024-02-05 2024-04-30 中南大学 Integrated intelligent reconstruction method for three-dimensional space line position of existing railway
CN117726196A (en) * 2024-02-18 2024-03-19 中南大学 Comprehensive evaluation method for environment suitability for railway route selection
CN117726196B (en) * 2024-02-18 2024-05-14 中南大学 Comprehensive evaluation method for environment suitability for railway route selection

Also Published As

Publication number Publication date
CN113554467B (en) 2022-07-22

Similar Documents

Publication Publication Date Title
CN113554467B (en) Railway three-dimensional linear intelligent design method based on co-evolution
Pu et al. A three‐dimensional distance transform for optimizing constrained mountain railway alignments
CN104050817B (en) Speed limiting information base generation and speed limiting information detection method and system
CN104318766A (en) Bus GPS track data road network matching method
CN108010103A (en) The quick fine generation method of river with complicated landform
CN112487695B (en) Multi-target intelligent comprehensive line selection method for railway in complex environment
CN106981198A (en) Deep learning network model and its method for building up for predicting travel time
Maji et al. Multi‐objective highway alignment optimization using a genetic algorithm
CN112883474B (en) Layout method, system, terminal and readable storage medium for intelligent reconstruction of existing track line
CN105718632B (en) A kind of uncertainty underground water repairs multiple-objection optimization management method
CN103399849A (en) Road three-dimensional linear automatic optimization method based on improved particle swarm optimization
CN104731963A (en) Grid path recommending method and system based on internet of vehicle
CN107729663A (en) A kind of highway geometry three-dimensional optimized method based on multilayer constraint system
CN114706400B (en) Path planning method based on improved A-x algorithm in off-road environment
CN104462714A (en) Calculation method for optimizing municipal drainage network plane layout design
CN107180534A (en) The express highway section average speed method of estimation of support vector regression fusion
CN103698809A (en) Kirchhoff pre-stack time migration parallel method with no speed-up ratio bottleneck
CN108985518A (en) A kind of traffic congestion charging policy optimization method based on VISSIM emulation
CN106949897B (en) Method and device for generating roads in map
CN114723121A (en) GIS-based field complex terrain path planning method
CN102339473B (en) Genetic algorithm-based method for design of cartoon model
CN106408660A (en) Digital terrain model compression method and equipment for electric power path planning
CN111125862B (en) Following model emission measuring and calculating method based on genetic algorithm and specific power
CN111008730B (en) Crowd concentration prediction model construction method and device based on urban space structure
CN113012286B (en) Method for constructing road DEM based on dense point cloud data

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant