CN112990271B - Parallel traffic simulation method and system based on traffic cluster - Google Patents
Parallel traffic simulation method and system based on traffic cluster Download PDFInfo
- Publication number
- CN112990271B CN112990271B CN202110187648.4A CN202110187648A CN112990271B CN 112990271 B CN112990271 B CN 112990271B CN 202110187648 A CN202110187648 A CN 202110187648A CN 112990271 B CN112990271 B CN 112990271B
- Authority
- CN
- China
- Prior art keywords
- traffic
- partition
- road
- updating
- nodes
- 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.)
- Active
Links
- 238000004088 simulation Methods 0.000 title claims abstract description 34
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000005192 partition Methods 0.000 claims abstract description 209
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 28
- 230000001360 synchronised effect Effects 0.000 claims abstract description 15
- 238000000638 solvent extraction Methods 0.000 claims description 12
- 230000005540 biological transmission Effects 0.000 claims description 8
- 230000003044 adaptive effect Effects 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 6
- 230000001131 transforming effect Effects 0.000 claims description 5
- 238000010845 search algorithm Methods 0.000 claims description 4
- 238000000265 homogenisation Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 abstract description 8
- 230000008859 change Effects 0.000 abstract description 4
- YCKRFDGAMUMZLT-UHFFFAOYSA-N Fluorine atom Chemical compound [F] YCKRFDGAMUMZLT-UHFFFAOYSA-N 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 239000003795 chemical substances by application Substances 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 229910052731 fluorine Inorganic materials 0.000 description 2
- 239000011737 fluorine Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
- 238000013316 zoning Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/40—Business processes related to the transportation industry
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/12—Edge-based segmentation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Business, Economics & Management (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Economics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention relates to the field of urban traffic systems, and discloses a parallel traffic simulation method and system based on traffic clusters, wherein an input traffic network is obtained; identifying a traffic cluster and a traffic cluster center point in a traffic network graph structure; respectively generating a plurality of initial subareas from the central points of all traffic clusters by using a weighted graph growth algorithm; judging whether the total synchronous message quantity and the load unbalance degree of the current subarea exceed the threshold values according to the subarea updating condition, and obtaining a subarea updating judgment result; and performing self-adaptive partition updating on each edge cutting node according to the partition updating judgment result. The invention effectively reduces the number of synchronous messages among the computing nodes, automatically determines the number of the partitions by identifying the traffic cluster, solves the problem of partition uncertainty, avoids a complicated computing process, enables the partition result to dynamically adapt to the change of flow, reduces manual intervention, can obtain better partitions, minimizes the total number of transmitted messages and improves the parallel traffic simulation efficiency.
Description
Technical Field
The invention relates to the field of urban traffic systems, in particular to a parallel traffic simulation method and system based on traffic clusters.
Background
Traffic simulation based on individuals is widely applied to analysis and evaluation of urban traffic systems. Such individual-based traffic simulation is mainly performed by tracking travel plans (such as departure, driving, and arrival) of each individual (such as a vehicle). In the traffic simulation process, a large amount of computing resources are occupied by creating agents, agent attribute configuration (such as vehicle size, capacity and speed), state change (such as transition from a driving state to an arrival state) and the like, and the simulation of tens of millions of individuals in a large city can cause the exhaustion of the computing resources, so that the computing time is too long, and the simulation result is unavailable. Parallel simulation has an important meaning in accelerating urban traffic simulation, and therefore receives more and more attention.
One common parallel traffic simulation method is to divide the road network into several sub-areas, and allocate the sub-areas to a series of computing nodes (such as CPU cores), and then each computing node is responsible for processing the individual travel plans occurring in the corresponding area. The efficiency of parallel traffic simulation depends to a large extent on the result of road network division. Introducing a topology to partition a road network is a common method that can effectively reduce the number of edge cuts (i.e., edges connected to two compute nodes, which can be considered "message pipes"). The other road network dividing method is based on a heuristic algorithm, and the load balancing partitions can be obtained by using the method. In terms of updating partitions, updating partitions by using a threshold value is a common technology, and dynamic update of the partitions can be realized. The road network division methods achieve certain achievement and improve the performance of parallel simulation. However, they still have some disadvantages: (1) most research focuses on reducing the number of edge cuts, but this may not minimize the total number of messages between compute nodes, since the number of edge cuts is not equal to the total number of messaging; furthermore, the number of synchronization messages on certain edge cuts may be very high; (2) the prior method needs a large amount of experiments to determine the number of partitions, which causes resource and time waste; (3) much a priori knowledge is required when updating the partitions.
Disclosure of Invention
The invention provides a parallel traffic simulation method and system based on traffic clusters, thereby solving the problems in the prior art.
In a first aspect, the invention provides a parallel traffic simulation method based on traffic clusters, which comprises the following steps:
s1) acquiring an input traffic network, and converting the input traffic network into a traffic network diagram structure;
s2) automatically identifying traffic clusters and traffic cluster center points in the traffic network graph structure by using an improved density-based clustering algorithm;
s3) respectively generating a plurality of initial subareas from the central points of all traffic clusters by using a weighted graph growth algorithm, and respectively adding all road nodes into the plurality of initial subareas to obtain initial subarea results;
s4) acquiring m edge cutting nodes according to the initial partition result, establishing a partition updating condition, judging whether the total synchronous message quantity and the load unbalance degree of the current partition exceed the threshold value according to the partition updating condition, and acquiring a partition updating judgment result;
s5) according to the partition updating judgment result, carrying out self-adaptive partition updating on each edge cutting node.
Further, in step S1), the input traffic network is converted into a traffic network graph structure, which includes using the road nodes as vertices of the graph in the traffic network graph structure, using the roads as edges of the graph in the traffic network graph structure, and using the traffic flow on the roads as weights of the edges in the traffic network graph structure.
Further, in step S2), a traffic cluster and a traffic cluster center point in the traffic network graph structure are automatically identified by using an improved density-based clustering algorithm, including identifying a plurality of traffic clusters and a road node with the highest local traffic flow density in the plurality of traffic clusters, and using the road node with the highest local traffic flow density in the plurality of traffic clusters as the traffic cluster center points of the plurality of traffic clusters respectively.
Further, identifying a plurality of traffic clusters and road nodes with highest local traffic flow density in the plurality of traffic clusters, respectively using the road nodes with highest local traffic flow density in the plurality of traffic clusters as traffic cluster center points of the plurality of traffic clusters, and further using vehicles on a road as homogenization points of the road nodes and clustering the vehicles on the road nodes; and calculating the gamma value of the road node according to the shortest road network distance between two road nodes and the local flow density of the road node during traffic flow clustering, acquiring k road nodes with high gamma values, and taking the k road nodes with high gamma values as the central points of the traffic cluster, wherein k is more than or equal to 1.
Further, the ith road NodeiThe gamma value of (a) is gamma (i) delta (i), which indicates the ith road NodeiD, the traffic flow within the road network distance; delta (i) denotes the ith road node NodeiAnd the jth road NodejThe shortest road network distance between the nodes, i is not equal to j, the jth road NodejThe traffic flow in the d-step road network distance is greater than the ith road NodeiAnd d, traffic flow within the road network distance.
Further, in step S3), generating a plurality of initial sub-areas starting from all the traffic cluster center points respectively by using a weighted graph growth algorithm, adding all the road nodes into the plurality of initial sub-areas respectively, and obtaining initial sub-area results, including generating a plurality of initial sub-areas { P corresponding to all the traffic cluster center points respectively by using a weighted graph growth algorithm1,P2,P3,…,PkAdding adjacent road nodes with the maximum traffic flow between each initial partition in parallel; when the adjacent road nodes with the maximum traffic flow between the adjacent road nodes and each initial subarea are added in parallel, if the initial subareas { P }1,P2,P3,…,PkWhen the maximum traffic flow is reached on the z-th road node at the same time, selecting an initial partition with the maximum traffic flow between the initial partition and the z-th road node as a partition to which the z-th road node belongs, wherein z is 1, 2, … and r, the total number of the road nodes is r, and the initial partition with the maximum traffic flow between the initial partition and the z-th road node is argmax ({ flow)1,flow2,flow3,…,flowk})，flowkRepresenting the traffic flow between the kth initial partition and the z < th > road node; and adding all road nodes into each partition in parallel to obtain an initial partition result.
Further, in step S4), a partition updating condition is established, and whether the total synchronization message amount and the load imbalance degree of the current partition exceed the threshold is determined according to the partition updating condition, so as to obtain a partition updating determination result, including the following steps:
s41) obtaining a plurality of classified partitions according to the initial partitioning result, and obtaining m edge cutting nodes according to the plurality of classified partitions;
s42) calculating the objective function value for measuring the current traffic state at the time t according to the m edge cutting nodes
E denotes the set of edge cuts, floweRepresenting the traffic on the e-th edge cut, N representing the set of partitions, FlownThe traffic volume of the nth zone is indicated,
representing the average traffic flow in each zone;
s43) according to the objective function value f used for measuring the current traffic statetEstablishing a partition update condition ft-fl>α*flWherein f islThe target function value after the last traffic state updating is obtained, and alpha is a proportionality coefficient; judging whether the total synchronous message quantity and the load unbalance degree of the current subarea exceed the threshold value according to the subarea updating condition, and if the traffic state updating condition ft-fl>α*flWhen the total synchronization message amount and the load imbalance degree of the current partition exceed the threshold, the step S5 is executed; when the traffic state updates the condition ft-fl>α*flIf not, the process returns to step S41).
Further, in step S42), the objective function value f is calculatedtRespectively normalizing the traffic flow on the edge cut of the e-th edge and the traffic flow of the nth subarea according to a Max-Min normalization formula to obtain the normalized traffic flow on the edge cut of the e-th edge and the normalized traffic flow of the nth subarea; the Max-Min normalization formula is
x represents the traffic flow on the e-th edge cut or the traffic flow of the nth zone; x is the number ofminRepresenting the minimum value of the traffic flow on the edge cut of the ith edge or the minimum value of the traffic flow of the nth subarea; x is a radical of a fluorine atommaxThe maximum value of the traffic flow on the e-th edge cut or the maximum value of the traffic flow of the nth subarea is shown.
Further, in step 5), performing adaptive partition updating on each edge cutting node according to the partition updating judgment result, including the following steps:
s51) searching neighborhood nodes of all edge cutting nodes by using a local search algorithm, and storing partitions of the neighborhood nodes;
s52) iteratively transforming the partition of the edge cutting node according to the total message transmission quantity and the load unbalance degree;
in step S52), iteratively transforming the partitions of the edge-cut node according to the total message transmission amount and the load imbalance degree, including the following steps:
s521) respectively replacing the partition to which the w-th edge cutting node belongs with partitions of a plurality of neighborhood nodes
Representing that the partition to which the w-th edge cutting node belongs is replaced by the partition of the q-th neighbor node; w is 1, 2, …, m;
s522) calculating the objective function value before replacing the partition to which the w-th edge cutting node belongs
And respectively replacing the partition to which the w-th edge cutting node belongs with a plurality of objective function values obtained by partitioning a plurality of neighborhood nodes
Representing the objective function value obtained by replacing the partition to which the w-th edge cutting node belongs with the partition of the q-th neighborhood node, and obtaining a plurality of objective function values
Respectively changing the target function value of the partition to which the w-th edge cutting node belongs
Comparing, if the objective function value before the partition to which the w-th edge cutting node belongs is replaced
Is at least larger than the target function values
Updating the partition to which the w-th edge cutting node belongs, wherein the updated partition to which the w-th edge cutting node belongs is
S523) setting a first partition updating condition and a second partition updating condition, wherein the first partition updating condition is
The flow of the subarea traffic before the subarea to which the w-th edge cutting node belongs is replaced,
representing the minimum traffic flow in all the partitions before updating; the second partition update condition is
The updated w-th edge cutting node belongs to the subarea traffic volume,
representing the maximum traffic flow in all the updated partitions; determining whether the first partition update condition and the second partition update condition are simultaneousIf yes, updating the partition to which the w-th edge cutting node belongs; and if not, not updating the partition to which the w-th edge cutting node belongs.
Further, the improved density-based clustering algorithm is the CFSFDP algorithm.
On the other hand, the invention provides a parallel traffic simulation system based on a traffic cluster, which comprises two parallel traffic map division modules, wherein the two parallel traffic map division modules comprise an automatic map growing module and an adaptive updating module;
the automatic map growing module is used for acquiring an input traffic network and converting the input traffic network into a traffic network map structure; automatically identifying traffic clusters and traffic cluster center points in the traffic network map structure by using an improved density-based clustering algorithm; respectively generating a plurality of initial subareas for the central points of all traffic clusters by using a weighted graph growth algorithm, and respectively adding all road nodes into the initial subareas to obtain initial subarea results;
the self-adaptive updating module is used for acquiring m edge cutting nodes according to the initial partitioning result, establishing a partitioning updating condition, judging whether the total synchronous message quantity and the load unbalance degree of the current partition exceed a threshold value according to the partitioning updating condition, and acquiring a partitioning updating judgment result; and performing self-adaptive partition updating on each edge cutting node according to the partition updating judgment result.
The invention has the beneficial effects that: the invention effectively reduces the number of synchronous messages among the calculation nodes, automatically determines the number of the partitions by identifying the traffic cluster, solves the problem of partition uncertainty and avoids a complicated simulation process. In addition, the invention updates the subareas by detecting the number of the synchronous messages and the load unbalance degree of the current subareas, so that the subarea result can dynamically adapt to the change of the flow, and the manual intervention is reduced. The invention can obtain better subareas and minimize the total message transmission quantity, and finally achieves the result of improving the parallel traffic simulation efficiency.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments are briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
Fig. 1 is a schematic flow chart of a parallel traffic simulation method based on traffic clusters according to a first embodiment of the present invention.
Fig. 2 is a schematic structural diagram of a parallel traffic simulation system based on traffic clusters according to a first embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating the invention, are intended for purposes of illustration only and are not intended to limit the scope of the invention. It is noted that the terms "comprises" and "comprising," and any variations thereof, in the description and claims of the present invention and the above-described drawings are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of elements is not necessarily limited to those elements, but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
In a first aspect, the present invention provides a parallel traffic simulation method based on traffic clusters, as shown in fig. 1, including the following steps:
s1), acquiring an input traffic network, and converting the input traffic network into a traffic network graph structure;
s2) automatically identifying traffic clusters and traffic cluster center points in the traffic network graph structure by using an improved density-based clustering algorithm; in the embodiment, a cfsfdp (clustering by fast search and find of diversity peaks) algorithm is adopted to automatically identify the traffic clusters and the traffic cluster center points in the traffic network graph structure.
S3) respectively generating a plurality of initial subareas from the central points of all traffic clusters by using a weighted graph growth algorithm, and respectively adding all road nodes into the plurality of initial subareas to obtain initial subarea results;
s4) acquiring m edge cutting nodes according to the initial partition result, establishing a partition updating condition, judging whether the total synchronous message quantity and the load unbalance degree of the current partition exceed the threshold value according to the partition updating condition, and acquiring a partition updating judgment result;
s5) according to the partition updating judgment result, carrying out self-adaptive partition updating on each edge cutting node.
In step S1), the input traffic network is converted into a traffic network graph structure, including taking the road nodes as vertices of the graph in the traffic network graph structure, taking the roads as edges of the graph in the traffic network graph structure, and taking the traffic flow on the roads as weights of the edges in the traffic network graph structure.
In step S2), automatically identifying a traffic cluster and a traffic cluster center point in the traffic network graph structure by using an improved density-based clustering algorithm, including identifying a plurality of traffic clusters and a road node with the highest local traffic flow density in the plurality of traffic clusters, and respectively using the road node with the highest local traffic flow density in the plurality of traffic clusters as the traffic cluster center point (i.e., the road node with the highest local traffic flow density) of the plurality of traffic clusters.
Identifying a plurality of traffic clusters and road nodes with highest local traffic flow density in the plurality of traffic clusters, respectively taking the road nodes with highest local traffic flow density in the plurality of traffic clusters as traffic cluster center points of the plurality of traffic clusters, and further taking vehicles on a road as homogenization points of the road nodes and clustering the vehicles on the road nodes; and calculating the gamma value of the road node according to the shortest road network distance between two road nodes and the local flow density of the road node during traffic flow clustering, acquiring k road nodes with high gamma values, and taking the k road nodes with high gamma values as the central points of the traffic cluster, wherein k is more than or equal to 1.
Ith road NodeiThe gamma value of (a) is gamma (i) delta (i), which indicates the ith road NodeiDistance of d-step road networkThe traffic flow therein; delta (i) represents the ith road NodeiAnd the jth road NodejThe shortest road network distance between the nodes, i is not equal to j, the jth road NodejThe traffic flow in the d-step road network distance is larger than that of the ith road NodeiD equals 4.
In step S3), generating a plurality of initial subareas respectively starting from all the traffic cluster center points by using a weighted graph growth algorithm, adding all the road nodes into the plurality of initial subareas respectively, and obtaining initial subarea results, including generating a plurality of initial subareas { P) respectively corresponding to all the traffic cluster center points by using a weighted graph growth algorithm1,P2,P3,…,PkAdding adjacent road nodes with the maximum traffic flow between each initial partition in parallel; when the adjacent road nodes with the maximum traffic flow between the adjacent road nodes and each initial subarea are added in parallel, if the initial subareas { P }1,P2,P3,…,PkWhen the maximum traffic flow is reached on the z-th road node at the same time, selecting an initial partition with the maximum traffic flow between the initial partition and the z-th road node as a partition to which the z-th road node belongs, wherein z is 1, 2, … and r, the total number of the road nodes is r, and the initial partition with the maximum traffic flow between the initial partition and the z-th road node is argmax ({ flow)1,flow2,flow3,…,flowk})，flowkRepresenting the traffic flow between the kth initial partition and the z < th > road node; and adding all road nodes into each partition in parallel to obtain an initial partition result.
In step S4), a partition updating condition is established, and whether the total synchronization message amount and the load imbalance degree of the current partition exceed the threshold is determined according to the partition updating condition, so as to obtain a partition updating determination result, including the following steps:
s41) obtaining a plurality of classified partitions according to the initial partitioning result, and obtaining m edge cutting nodes according to the plurality of classified partitions;
s42) calculating the current traffic state at the time t according to the m edge cutting nodesTarget function value of
E denotes the set of edge cuts, floweRepresenting the traffic Flow on the e-th edge cut, N representing the set of partitions, FlownThe traffic volume of the nth zone is shown,
representing the average traffic flow in each zone.
In step S42), the objective function value f is calculatedtRespectively normalizing the traffic flow on the edge cut of the e-th edge and the traffic flow of the nth subarea according to a Max-Min normalization formula to obtain the normalized traffic flow on the edge cut of the e-th edge and the normalized traffic flow of the nth subarea; the Max-Min normalization formula is
x represents the traffic flow on the e-th edge cut or the traffic flow of the nth zone; x is the number ofminRepresenting the minimum value of the traffic flow on the edge cut of the ith edge or the minimum value of the traffic flow of the nth subarea; x is a radical of a fluorine atommaxThe maximum value of the traffic flow on the e-th edge cut or the maximum value of the traffic flow of the nth subarea is shown.
S43) according to the objective function value f used for measuring the current traffic statetEstablishing a partition update condition ft-fl>α*flWherein f islThe target function value after the last traffic state updating is obtained, and alpha is a proportionality coefficient; judging whether the total synchronous message quantity and the load unbalance degree of the current subarea exceed the threshold value according to the subarea updating condition, and if the traffic state updating condition ft-fl>α*flWhen the total synchronization message amount and the load imbalance degree of the current partition exceed the threshold, the step S5 is executed; when the traffic state updates the condition ft-fl>α*flIf not, the process returns to step S41).
Step S43), the result value (objective function of the current traffic state) calculated in step S42) is compared with the result value updated last time (i.e., the objective function value updated last traffic state), and if the current result value exceeds the result value updated last time by a certain ratio, the operation of updating the edge cutting node partition is performed.
In step 5), performing adaptive partition updating on each edge cutting node according to the partition updating judgment result, including the following steps:
s51) searching neighborhood nodes of all edge cutting nodes by using a local search algorithm, and storing partitions of the neighborhood nodes;
s52) iteratively transforming the partitions of the edge cutting nodes according to the total message transmission quantity and the load unbalance degree;
in step S52), iteratively transforming the partitions of the edge cutting node according to the total message transmission amount and the load imbalance degree, the method includes the following steps:
s521) respectively replacing the partition to which the w-th edge cutting node belongs with partitions of a plurality of neighborhood nodes
Representing that the partition to which the w-th edge cutting node belongs is replaced by the partition of the q-th neighborhood node; w is 1, 2, …, m;
s522) calculating the objective function value before the partition to which the w-th edge cutting node belongs is replaced
And respectively replacing the partition to which the w-th edge cutting node belongs with a plurality of objective function values after the partitions of a plurality of neighborhood nodes are replaced
Indicates that the partition to which the w-th edge cut node belongs is changed to the q-th oneThe objective function values after the partition of the neighborhood nodes, and the objective function values
Respectively changing the target function value of the partition to which the w-th edge cutting node belongs before replacement
Comparing, if the objective function value before the partition to which the w-th edge cutting node belongs is replaced
Is at least greater than the plurality of objective function values
In the above, the partition to which the w-th edge cutting node belongs is updated, and the updated partition to which the w-th edge cutting node belongs is
S523) setting a first partition updating condition and a second partition updating condition, wherein the first partition updating condition is that
The flow of the subarea traffic before the subarea to which the w-th edge cutting node belongs is replaced,
representing the minimum traffic flow in all the partitions before updating; the second partition update condition is
Represents the w-th after updateThe zoning traffic volume to which the edge cutting node belongs,
representing the maximum traffic flow in all the updated partitions; judging whether the first partition updating condition and the second partition updating condition are simultaneously satisfied, if so, saving the update of the partition to which the w-th edge cutting node belongs; and if not, not updating the partition to which the w-th edge cutting node belongs.
In this embodiment, the partition to which the w-th edge cutting node belongs before updating is denoted as h, and the partition to which the w-th edge cutting node belongs after updating is denoted as f. The traffic flow of the partition h to which the w-th edge cut node belongs before updating must be higher than the minimum threshold (i.e. the traffic flow of the partition h is higher than the minimum threshold)
) And the traffic flow of the partition f to which the updated w-th edge cutting node belongs must be lower than the maximum threshold. And when the first partition updating condition and the second partition updating condition are both satisfied, keeping the update of the partition to which the w-th edge cutting node belongs, and otherwise, canceling the update of the partition to which the w-th edge cutting node belongs.
On the other hand, the invention provides a parallel traffic simulation system based on a traffic cluster, which comprises two parallel traffic map division modules, wherein the two parallel traffic map division modules comprise an automatic map growing module and a self-adaptive updating module;
the automatic map growing module is used for acquiring an input traffic network and converting the input traffic network into a traffic network map structure; automatically identifying traffic clusters and traffic cluster center points in the traffic network map structure by using an improved density-based clustering algorithm; respectively generating a plurality of initial subareas for all traffic cluster central points by using a weighted graph growth algorithm, and respectively adding all road nodes into the initial subareas to obtain initial subarea results;
the self-adaptive updating module is used for acquiring m edge cutting nodes according to the initial partitioning result, establishing a partitioning updating condition, judging whether the total synchronous message quantity and the load unbalance degree of the current partition exceed a threshold value according to the partitioning updating condition, and acquiring a partitioning updating judgment result; and performing self-adaptive partition updating on each edge cutting node according to the partition updating judgment result.
FIG. 2 shows that the parallel traffic simulation system based on the traffic cluster comprises two parallel traffic map division modules, wherein the two parallel traffic map division modules comprise an automatic map growing module and an adaptive updating module. In the automatic graph growth module, firstly, an improved density-based clustering algorithm is applied to automatically determine central points of several traffic clusters, and then a weighted graph growth algorithm is utilized to generate 4 initial partitions { P }in parallel1,P2,P3,P4}. The adaptive update module is accomplished by iteratively updating the partitions of the edge-cutting node: firstly, acquiring a partition to which an edge cutting node before updating belongs, and storing the edge cutting node in a queue; then sequentially accessing each edge cutting node in the queue, and finding out a neighborhood node corresponding to the edge cutting node through local search; finally, the partition of each edge-cutting node will be modified according to the degree of load imbalance and the total number of messages. And outputting the updated partition after all the edge cutting nodes complete the partition. In this embodiment, the adjacent partition P of the edge-cut node 5 is found by the local search algorithm2、P3And P4Then the partitions of the edge cutting node 5 are respectively divided from P1Change to P2、P3And P4And calculating the unbalanced workload degree and the total number of the messages after the partition is changed. Finally, as the updated traffic flow of the partition to which the edge cutting node 5 belongs is compared with the traffic flow before updating, the workload imbalance and the message quantity can reach the minimum, and the partition to which the edge cutting node 5 belongs is updated to be P3。
By adopting the technical scheme disclosed by the invention, the following beneficial effects are obtained: the invention effectively reduces the number of synchronous messages among the calculation nodes, automatically determines the number of partitions by identifying the traffic cluster, solves the problem of partition uncertainty and avoids a complicated simulation process. In addition, the invention updates the subareas by detecting the number of the synchronous messages and the load unbalance degree of the current subareas, so that the subarea result can dynamically adapt to the change of the flow, and the manual intervention is reduced. The invention can obtain better subareas and minimize the total message transmission quantity, and finally achieves the result of improving the parallel traffic simulation efficiency.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that it will be apparent to those skilled in the art that various modifications and improvements can be made without departing from the principle of the present invention, and such modifications and improvements should also be considered within the scope of the present invention.
Claims (6)
1. A parallel traffic simulation method based on traffic clusters is characterized by comprising the following steps:
s1), acquiring an input traffic network, and converting the input traffic network into a traffic network graph structure;
s2) automatically identifying traffic clusters and traffic cluster center points in the traffic network graph structure by using an improved density-based clustering algorithm;
s3) respectively generating a plurality of initial subareas from the central points of all traffic clusters by using a weighted graph growth algorithm, and respectively adding all road nodes into the plurality of initial subareas to obtain initial subarea results;
s4) acquiring m edge cutting nodes according to the initial partition result, establishing a partition updating condition, judging whether the total synchronous message quantity and the load unbalance degree of the current partition exceed the threshold value according to the partition updating condition, and acquiring a partition updating judgment result;
s5) carrying out self-adaptive partition updating on each edge cutting node according to the partition updating judgment result;
in step S2), automatically identifying a traffic cluster and a traffic cluster center point in the traffic network graph structure by using an improved density-based clustering algorithm, including identifying a plurality of traffic clusters and a road node with the highest local traffic flow density in the plurality of traffic clusters, and respectively using the road node with the highest local traffic flow density in the plurality of traffic clusters as the traffic cluster center point of the plurality of traffic clusters; the method also comprises the steps of taking vehicles on the road as homogenization points of road nodes and clustering the vehicles on the road nodes; calculating the gamma value of a road node according to the shortest road network distance between two road nodes and the local flow density of the road node during traffic flow clustering, acquiring k road nodes with high gamma values, and taking the k road nodes with high gamma values as the central points of traffic clusters, wherein k is more than or equal to 1;
in step S3), generating a plurality of initial subareas respectively starting from all the traffic cluster center points by using a weighted graph growth algorithm, adding all the road nodes into the plurality of initial subareas respectively, and obtaining initial subarea results, including generating a plurality of initial subareas { P) respectively corresponding to all the traffic cluster center points by using a weighted graph growth algorithm1,P2,P3,…,PkAdding adjacent road nodes with the maximum traffic flow between each initial partition in parallel; when the adjacent road nodes with the maximum traffic flow between the adjacent road nodes and each initial subarea are added in parallel, if the initial subareas { P }1,P2,P3,…,PkWhen the maximum traffic flow is reached on the z-th road node at the same time, selecting an initial partition with the maximum traffic flow between the initial partition and the z-th road node as a partition to which the z-th road node belongs, wherein z is 1, 2, … and r, the total number of the road nodes is r, and the initial partition with the maximum traffic flow between the initial partition and the z-th road node is argmax ({ flow)1,flow2,flow3,…,flowk})，flowkRepresenting the traffic flow between the kth initial partition and the z < th > road node; adding all road nodes into each partition in parallel to obtain an initial partition result;
in step S4), a partition updating condition is established, and whether the total synchronization message amount and the load imbalance degree of the current partition exceed the threshold is determined according to the partition updating condition, so as to obtain a partition updating determination result, including the following steps:
s41) obtaining a plurality of classified partitions according to the initial partitioning result, and obtaining m edge cutting nodes according to the plurality of classified partitions;
s42) calculating the objective function value for measuring the current traffic state at the time t according to the m edge cutting nodes
E denotes the edge cut set, floweRepresenting the traffic Flow on the e-th edge cut, N representing the set of partitions, FlownThe traffic volume of the nth zone is shown,
representing the average traffic flow in each zone;
s43) according to the objective function value f for measuring the current traffic statetEstablishing a partition update condition ft-fl>α*flWherein f islThe target function value after the last traffic state updating is obtained, and alpha is a proportionality coefficient; judging whether the total synchronous message quantity and the load unbalance degree of the current subarea exceed the threshold value according to the subarea updating condition, and if the traffic state updating condition ft-fl>α*flWhen the total synchronization message amount and the load imbalance degree of the current partition exceed the threshold, the step S5 is executed; when the traffic state updates the condition ft-fl>α*flIf not, returning to step S41);
in step 5), performing adaptive partition updating on each edge cutting node according to the partition updating judgment result, including the following steps:
s51) searching neighborhood nodes of all edge cutting nodes by using a local search algorithm, and storing partitions of the neighborhood nodes;
s52) iteratively transforming the partitions of the edge-cut nodes according to the total message transmission amount and the load imbalance degree.
2. The parallel traffic simulation method based on traffic clusters according to claim 1, wherein in step S1), the input traffic network is converted into a traffic network graph structure, which comprises using road nodes as vertices of a graph in the traffic network graph structure, using roads as edges of the graph in the traffic network graph structure, and using traffic on the roads as weights of the edges in the traffic network graph structure.
3. The parallel traffic simulation method based on traffic cluster of claim 1, wherein the ith road Node is a Node of a road with a specific traffic clusteriThe gamma value of (a) is gamma (i) delta (i), which indicates the ith road NodeiD, traffic flow within the distance of the road network; delta (i) represents the ith road NodeiAnd the jth road NodejThe shortest road network distance between the nodes, i is not equal to j, the jth road NodejThe traffic flow in the d-step road network distance is greater than the ith road NodeiAnd d, traffic flow within the road network distance.
4. The parallel traffic simulation method based on traffic clusters according to claim 1, characterized in that in step S42), the objective function value f is calculatedtRespectively normalizing the traffic flow on the edge cut of the e-th edge and the traffic flow of the nth subarea according to a Max-Min normalization formula to obtain the normalized traffic flow on the edge cut of the e-th edge and the normalized traffic flow of the nth subarea; the Max-Min normalization formula is
x represents the traffic flow on the e edge cut or the traffic flow of the nth zone; x is the number ofminRepresenting the minimum value of the traffic flow on the edge cut of the ith edge or the minimum value of the traffic flow of the nth subarea; x is the number ofmaxThe maximum value of the traffic flow at the e-th edge cut or the maximum value of the traffic flow at the n-th zone is shown.
5. The parallel traffic simulation method based on traffic cluster according to claim 1, wherein in step S52), the partition of the edge cutting node is iteratively transformed according to the total message transmission amount and the load imbalance degree, and the method comprises the following steps:
S521) respectively replacing the partition to which the w-th edge cutting node belongs with partitions of a plurality of neighborhood nodes
Representing that the partition to which the w-th edge cutting node belongs is replaced by the partition of the q-th neighborhood node; w is 1, 2, …, m;
s522) calculating the objective function value before the partition to which the w-th edge cutting node belongs is replaced
And respectively replacing the partition to which the w-th edge cutting node belongs with a plurality of objective function values after the partitions of a plurality of neighborhood nodes are replaced
Representing the objective function value obtained by replacing the partition to which the w-th edge cutting node belongs with the partition of the q-th neighborhood node, and obtaining a plurality of objective function values
Respectively changing the target function value of the partition to which the w-th edge cutting node belongs before replacement
Comparing, if the objective function value before the partition to which the w-th edge cutting node belongs is replaced
At leastGreater than the number of objective function values
In the above, the partition to which the w-th edge cutting node belongs is updated, and the updated partition to which the w-th edge cutting node belongs is
S523) setting a first partition updating condition and a second partition updating condition, wherein the first partition updating condition is
Shows the subarea traffic volume before the subarea to which the w-th edge cutting node belongs is replaced,
representing the minimum traffic flow in all the partitions before updating; the second partition update condition is
The updated w-th edge cutting node belongs to the subarea traffic volume,
representing the maximum traffic flow in all the updated partitions; judging whether the first partition updating condition and the second partition updating condition are simultaneously satisfied, if so, saving the partition to which the w-th edge cutting node belongsUpdating of (1); and if not, not updating the partition to which the w-th edge cutting node belongs.
6. A parallel traffic simulation system based on traffic clusters, which is suitable for the parallel traffic simulation method based on traffic clusters according to any one of claims 1 to 5, and is characterized by comprising two parallel traffic map division modules, wherein the two parallel traffic map division modules comprise an automatic map growing module and an adaptive updating module;
the automatic map growing module is used for acquiring an input traffic network and converting the input traffic network into a traffic network map structure; automatically identifying traffic clusters and traffic cluster center points in the traffic network map structure by using an improved density-based clustering algorithm; respectively generating a plurality of initial subareas for the central points of all traffic clusters by using a weighted graph growth algorithm, and respectively adding all road nodes into the initial subareas to obtain initial subarea results;
the self-adaptive updating module is used for acquiring m edge cutting nodes according to an initial partition result, establishing a partition updating condition, judging whether the total synchronous message quantity and the load unbalance degree of the current partition exceed a threshold value according to the partition updating condition, and acquiring a partition updating judgment result; and performing self-adaptive partition updating on each edge cutting node according to the partition updating judgment result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110187648.4A CN112990271B (en) | 2021-02-18 | 2021-02-18 | Parallel traffic simulation method and system based on traffic cluster |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110187648.4A CN112990271B (en) | 2021-02-18 | 2021-02-18 | Parallel traffic simulation method and system based on traffic cluster |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112990271A CN112990271A (en) | 2021-06-18 |
| CN112990271B true CN112990271B (en) | 2022-07-08 |
Family
ID=76393514
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110187648.4A Active CN112990271B (en) | 2021-02-18 | 2021-02-18 | Parallel traffic simulation method and system based on traffic cluster |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112990271B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016169290A1 (en) * | 2015-04-21 | 2016-10-27 | 华南理工大学 | Decision-making supporting system and method oriented towards emergency disposal of road traffic accidents |
| CN108320511A (en) * | 2018-03-30 | 2018-07-24 | 江苏智通交通科技有限公司 | Urban highway traffic sub-area division method based on spectral clustering |
| CN112149287A (en) * | 2020-09-08 | 2020-12-29 | 安徽中科龙安科技股份有限公司 | Traffic simulation road network graphical segmentation method and system oriented to load balancing |
-
2021
- 2021-02-18 CN CN202110187648.4A patent/CN112990271B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016169290A1 (en) * | 2015-04-21 | 2016-10-27 | 华南理工大学 | Decision-making supporting system and method oriented towards emergency disposal of road traffic accidents |
| CN108320511A (en) * | 2018-03-30 | 2018-07-24 | 江苏智通交通科技有限公司 | Urban highway traffic sub-area division method based on spectral clustering |
| CN112149287A (en) * | 2020-09-08 | 2020-12-29 | 安徽中科龙安科技股份有限公司 | Traffic simulation road network graphical segmentation method and system oriented to load balancing |
Non-Patent Citations (2)
| Title |
|---|
| Building a Spatially-Embedded Network of Tourism Hotspots From Geotagged Social Media Data;Xinyu Wu et al.;《IEEE Access》;20180419;第06卷;全文 * |
| 城市中观交通仿真的交通流车簇模型;万凌松等;《计算机工程与设计》;20100228(第04期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112990271A (en) | 2021-06-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103179052B (en) | A kind of based on the central virtual resource allocation method and system of the degree of approach | |
| CN112995289B (en) | Internet of vehicles multi-target computing task unloading scheduling method based on non-dominated sorting genetic strategy | |
| CN109617826A (en) | A kind of storm dynamic load balancing method based on cuckoo search | |
| CN109102124B (en) | Dynamic multi-target multi-path induction method and system based on decomposition and storage medium | |
| CN111538867B (en) | Method and system for dividing bounded incremental graph | |
| CN106021560A (en) | Data stream-oriented low-delay memory B + tree index construction method | |
| CN105704031B (en) | A kind of data transmission path determination and device | |
| CN111405634B (en) | Method and device for self-adaptive clustering of wireless sensor network | |
| CN105677447A (en) | Clustering-based delay bandwidth minimization virtual machine deployment method in distributed cloud | |
| CN112200336A (en) | Method and device for planning vehicle driving path | |
| CN101051972A (en) | Network resource route selecting method | |
| Gao et al. | A deep learning framework with spatial-temporal attention mechanism for cellular traffic prediction | |
| CN112990271B (en) | Parallel traffic simulation method and system based on traffic cluster | |
| CN112511652B (en) | Cooperative computing task allocation method under edge computing | |
| CN1992673B (en) | Method of implementing fast packet flow recognition in high-speed router and firewall | |
| CN114154685A (en) | Electric energy data scheduling method in smart power grid | |
| CN117749795A (en) | Vehicle edge server deployment method based on reinforcement learning algorithm | |
| CN108413980B (en) | Traffic itinerant path planning method for reducing path branches | |
| CN109889573A (en) | Based on the Replica placement method of NGSA multiple target in mixed cloud | |
| CN115438451A (en) | Photovoltaic module serial line arrangement determining method and device and electronic equipment | |
| CN114827933A (en) | Multipath routing method for wireless sensor network | |
| CN111860700B (en) | Energy consumption classification method and device, storage medium and equipment | |
| CN114741191A (en) | Multi-resource allocation method for compute-intensive task relevance | |
| CN112948087A (en) | Task scheduling method and system based on topological sorting | |
| CN112487187A (en) | News text classification method based on graph network pooling |
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 |