CN114254213A - Top-k path sequence query method and system under multiple backgrounds - Google Patents

Abstract

The invention discloses a Top-k path sequence query method and a system under multiple backgrounds, which relate to the field of path planning, and the method comprises the steps of defining a time-dependent road network based on a time function, and defining the time function, a strategy sequence and road attributes of the time-dependent road network; connecting adjacent nodes of the intersection nodes to obtain an intersection connection table; establishing a vertex index for each node in the time-dependent road network to form a vertex index set; inquiring adjacent strategy points of road nodes to obtain a plurality of paths which can reach the strategy points; defining the domination relation among the paths, and obtaining an optimal path sequence set containing a preset number of optimal paths based on a path domination algorithm under the multi-background road network. The method can improve the efficiency of road query and efficiently realize Top-k path sequence query in a time-dependent road network under multiple backgrounds.

Description

Top-k path sequence query method and system under multiple backgrounds

Technical Field

The invention relates to the field of path planning, in particular to a Top-k path sequence query method and a Top-k path sequence query system under multiple backgrounds.

Background

With the rapid development of mobile networks and GPS (Global Positioning System) technology, people can more easily obtain more detailed road information than before, and meanwhile, the accuracy of road network matching is improved by a large amount of high-precision track data. The increasing variety and number of vehicles brings more alternative routes for people to travel, and the functions of various navigation services, which give a travel path between a designated source point and destination point by calculation, are becoming more sophisticated. The query of the road network is also expanded from the path query that each road on the traditional road network has a fixed weight to the path query on the dynamic road network that the weight of the path on the road network changes along with the time; in query, the static weight (e.g., distance) and the dynamic weight (e.g., travel time, fuel consumption, travel cost, etc.) are also often used as the basis for Top-k ranking of each path. In a real road network, the optimal paths between the same set of source points and destinations are different when queries are made at different times, because the difference in query times results in a change in the dynamic weights required for the same path.

When the time-dependent road network is queried, the optimal path under the current condition cannot be obtained frequently, in the normal travel process, the road traffic condition is changed along with the time, the weather, the holidays and other factors, and in addition, the self attributes (such as expressways, main roads, branches and the like) of roads can also influence the selection of the people. Different degrees of congestion often occur in early peak hours and late peak hours of a working day, and how to reasonably plan a path to bypass a congested road section in a peak hour or reduce the number of the congested road sections as much as possible is a great challenge. The driving speed is affected by weather conditions, for example, the driving speed in rainy and snowy days and in heavy foggy days becomes slow, so that the whole travel time becomes long, the selection of road attributes is changed, the time for reaching each node is changed, and the subsequently planned path has to be adjusted. In addition, there are time-limited traffic lanes, which are also taken into account during the inquiry. Therefore, under different query times and weather backgrounds, the results obtained after the user initiates the path query are different, and a more excellent query result can be obtained only by combining the current time-dependent road network, the weather factor and the road attribute in the query process.

Optimal path sequence query is also a long-standing research hotspot, finding a path with minimal dynamic cost or static cost through different strategy sequences (such as restaurant → gas station → movie theatre). This query approach has many practical applications in route planning, crisis management, supply chain management, and logistics transportation. The OSR (Optimal sequential routing) query only considers finding the shortest path sequence on the static road network, and the Top-k Optimal sequential routing query considers the situation that each person has different preferences for the same policy point, as shown in fig. 1, for example, the cost from the starting point s to supermarket a is 15, the cost to supermarket b is 20, but supermarket b has a discount activity, and at this time, if only the shortest path is recommended, the query may not meet the preferences of the user, so the query may return the first k Optimal paths meeting the given policy sequence. However, these query methods are only applicable to static road networks, and are not considered on time-dependent road networks with multiple backgrounds, and Top-k paths obtained when users query with the same strategy sequence under different backgrounds are different.

For a time-dependent road network with multiple backgrounds, the following problems need to be faced in the implementation process: (1) how to integrate the weather attribute and the road attribute into a time-dependent road network; (2) how to handle important traffic nodes such as crossroads; (3) how to perform efficient queries at different temporal contexts; (4) how to filter unnecessary exploration paths so as to effectively improve the path query efficiency.

Disclosure of Invention

Aiming at the defects in the prior art, the invention aims to provide a Top-k path sequence query method and a Top-k path sequence query system under multiple backgrounds, which can improve the efficiency of road query and efficiently realize the Top-k path sequence query in a time-dependent road network under multiple backgrounds.

In order to achieve the above object, the Top-k path sequence query method under multiple backgrounds provided by the present invention specifically includes the following steps:

defining a time-dependent road network based on a time function, defining the time function, a strategy sequence and road attributes of the time-dependent road network, and defining time-dependent road network path sequence query by taking a weather factor as a query condition;

preprocessing the intersection nodes of the road network by combining the time-dependent road network based on an intersection end point connection algorithm, and connecting adjacent nodes of the intersection nodes to obtain an intersection connection table;

establishing a vertex index for each node in a time-dependent road network to form a vertex index set, wherein the vertex index comprises a reverse connection lookup table and a forward connection lookup table, the reverse connection lookup table comprises all adjacent nodes which can reach the current node and a time function and road attributes of a road which passes through when the current node is reached, and the forward connection lookup table comprises all adjacent nodes which can be reached by the current node and a time function and road attributes of a road which passes through when the adjacent nodes are reached;

according to the adjacent strategy point query algorithm, vertex indexes established for each node in the time-dependent road network are integrated, and adjacent strategy points of road nodes are queried to obtain a plurality of paths which can reach the strategy points;

defining the domination relation among the paths, and inputting a source point, a destination, departure time, a strategy sequence, weather during departure and a required path sequence number of a query based on a path domination algorithm under the multi-background road network to obtain an optimal path sequence set containing the optimal paths with preset number.

On the basis of the above technical solution, the time-dependent road network specifically includes:

G_m＝(V_m，E_m，F_m，C_m，R_m)

wherein G is_mShowing time dependence of road network, V_mRepresenting a set of nodes, each node representing a different strategy point in the road network, E_mRepresenting a set of edges, each edge representing a path connecting two different nodes, for which edge e ═ v_i，v_j) Existence time boxNumber F_m，v_iAnd v_jNodes, C, corresponding to two vertices representing the edge e_mPolicy attributes for storage nodes when there is a node V ∈ V_mWhen it is time, the policy attribute function C (v) returns the policy attribute of node v, R_mFor storing road attributes, when there is an edge E ∈ E_mThen the road attribute function r (e) returns the road attribute of edge e.

On the basis of the above technical solution, the time-dependent road network is defined by a time function, a strategy sequence and road attributes, wherein:

the definition process of the time function is as follows:

defining a time function f on a time-dependent network_e(t) is a discrete function, and each edge e ═ v (v) in the time-dependent road network_i，v_j)∈E_mAll have a time function f_e(t_i)＝t_eIs used to indicate at time t_iSlave node v_iInitiating a query to reach node v_jThe required time is t_e；

The definition process for the policy sequence is:

in a real road network, each node has a policy attribute, a set of policy attributes of all vertices on the road network is a policy set S, and when nodes are visited in sequence by a specific sequence in the travel process, the sequence is called a policy sequence C, and the policy sequence C is called a policy sequence C<C₁，C₂，...，C_j>Indicates that C is required in the process of query₁，C₂，...，C_jThis sequence accesses in turn j corresponding attribute policy points, C_jRepresenting the jth policy point attribute, for each policy attribute in the policy sequence there is C_i∈S(1≤i≤j)，C_iRepresenting the ith strategy point attribute;

the definition process for the road attribute is as follows:

connection path between two nodes e ═ v_i，v_j) Then, the road attribute R (e) of the route e is R_i，R_iIndicating the road attribute of the route e.

On the basis of the technical proposal, the device comprises a shell,

the influence of the weather factors on road selection comprises traffic influence and travel time influence;

defining a weather factor influence coefficient for representing the influence of different weather on the running time of the roads with different attributes, wherein the weather factor influence coefficient is the ratio of the running time of a certain attribute road in normal weather to the running time of the same attribute road when the weather is w, and w represents the weather type.

On the basis of the technical scheme, the concrete steps of obtaining the intersection connection table comprise:

sequentially traversing nodes v of road network_iIf node v_iIf there is a crossing node v in the adjacent points, the node v will be connected_iThe edge of the intersection node v is e₁；

Traversing each adjacent node v of intersection nodes v_jLet the edge e₂I.e. (v, v)_j) Join to tuple e_iIn (1), a new path is formed<v_i，v，v_j>；

Will be edge e₁And edge e₂And merging the time functions, and reserving the road attribute to obtain the intersection connection table.

On the basis of the technical scheme, the reverse connection lookup table specifically comprises:

L_in(v_j)＝(v_i，f_e(t)，R(e))

wherein L is_in(v_j) Indicating a reverse connection look-up table, L_in(v_j) In, v_iIndicating that node v can be reached_jThe edge e ═ v_i，v_j)，f_e(t) represents a time function of the edge e, and R (e) represents a road attribute of the edge e.

On the basis of the above technical solution, the forward connection lookup table specifically includes:

L_out(v_j)＝(v_i，f_e(t)，R(e))

wherein L is_out(v_j) Indicating a forward link look-up table, L_out(v_j) In, v_iRepresenting a node v_jReachable node, edge e ═ v_j，v_i)，f_e(t) represents a time function of the edge e, and R (e) represents a road attribute of the edge e.

On the basis of the technical scheme, the method comprises the following specific steps of inquiring the adjacent strategy points of the road nodes according to the adjacent strategy point inquiry algorithm by integrating the vertex indexes established for each node in the time-dependent road network to obtain a plurality of paths which can reach the strategy points, wherein the specific steps comprise:

inputting the initial node v to be inquired currently_iNode v to be queried_iThe method comprises the following steps of (1) adjacent strategy attributes, query time t and current weather w;

establishing a path set BIFS _ list, wherein the path set BIFS _ list is used for storing a node v_iAll paths to the next policy attribute node and the tuple consisting of the arrival time;

traversing all nodes in the road network, and finding out the node v with the strategy attribute as the target attribute C_jStarting a loop, repeatedly traversing the node v_iForward link look-up table L_out(v_i) And node v_jReverse connection look-up table L_in(v_j) Finding a node v_a∈L_out(v_i) And node v_b∈L_in(v_j) When v is_a＝v_jTime, node v_iThe next adjacent node is the required strategy point, the connecting node v_iAnd node v_j(ii) a When v is_a＝v_bConnecting nodes, paths<v_i，v_a，v_a，v_j>Obtaining a path for the query; otherwise, the node v is connected_aAs the next query node L_out(v_i) Inquiring until all nodes are traversed;

inquiring to obtain the road attribute R (v) under the current weather w_i，v_j) Obtaining corresponding weather factor influence coefficient according to road attribute and weather

Calculating to obtain the time of reaching the next strategy point, and adding the path into a path set BIFS _ list;

and sorting the paths in the path set BIFS _ list according to the sequence of the arrival time from low to high to obtain a sorted path set BIFS _ list, and selecting the kth path which can reach the strategy point according to the sorted path set BIFS _ list.

The invention provides a Top-k path sequence query system under multiple backgrounds, which comprises:

the system comprises a definition module, a time function module and a query module, wherein the definition module is used for defining a time-dependent road network based on a time function, defining the time function, a strategy sequence and road attributes of the time-dependent road network and defining the time-dependent road network path sequence query by taking a weather factor as a query condition;

the connection module is used for preprocessing the intersection nodes of the road network based on an intersection end point connection algorithm and in combination with the time-dependent road network, and connecting the adjacent nodes of the intersection nodes to obtain an intersection connection table;

the system comprises an establishing module, a searching module and a judging module, wherein the establishing module is used for establishing a vertex index for each node in a time-dependent road network to form a vertex index set, the vertex index comprises a reverse connection lookup table and a forward connection lookup table, the reverse connection lookup table comprises all adjacent nodes which can reach the current node and a time function and road attributes of a road which passes through when the current node is reached, and the forward connection lookup table comprises all adjacent nodes which can reach the current node and a time function and road attributes of a road which passes through when the adjacent node is reached;

the query module is used for querying the adjacent strategy points of the road nodes according to the adjacent strategy point query algorithm and by integrating the vertex indexes established for each node in the time-dependent road network to obtain a plurality of paths which can reach the strategy points;

and the determining module is used for defining the domination relationship among the paths, inputting a source point, a destination, a departure time, a strategy sequence, weather during departure and a required path sequence number of the query based on a path domination algorithm under the multi-background road network, and obtaining an optimal path sequence set containing the optimal paths with the preset number.

On the basis of the above technical solution, the time-dependent road network specifically includes:

G_m＝(V_m，E_m，F_m，C_m，R_m)

wherein G is_mShowing time dependence of road network, V_mRepresenting a set of nodes, each node representing a different strategy point in the road network, E_mRepresenting a set of edges, each edge representing a path connecting two different nodes, for which edge e ═ v_i，v_j) Existence of a function of time F_m，v_iAnd v_jNodes, C, corresponding to two vertices representing the edge e_mPolicy attributes for storage nodes when there is a node V ∈ V_mWhen it is time, the policy attribute function C (v) returns the policy attribute of node v, R_mFor storing road attributes, when there is an edge E ∈ E_mThen the road attribute function r (e) returns the road attribute of edge e.

Compared with the prior art, the invention has the advantages that: firstly, simplifying the complex query problem caused by the existence of crossroads by a crossroad connection algorithm, secondly establishing forward and reverse query label indexes of nodes, taking the connection label indexes as a basic algorithm for querying proximity strategy points, and finally, providing a pruning algorithm based on path domination, and reasonably pruning according to the path domination relationship among roads so as to improve the efficiency of road query and efficiently realize Top-k path sequence query in a time-dependent road network under multiple backgrounds.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

FIG. 1 is a schematic diagram of a road network according to an embodiment of the present invention;

FIG. 2 is a flowchart of a Top-k path sequence query method under multiple contexts according to an embodiment of the present invention;

fig. 3 is a schematic diagram of an intersection according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments.

Referring to fig. 2, a Top-k path sequence query method under multiple backgrounds provided by the embodiment of the present invention specifically includes the following steps:

s1: defining a time-dependent road network based on a time function, defining the time function, a strategy sequence and road attributes of the time-dependent road network, and defining time-dependent road network path sequence query by taking a weather factor as a query condition;

the path weights of the time-dependent road network are mostly expressed by simple discrete functions, and the expression mode can make the expression of the path weights more diversified and intuitive, so the path weights of the time-dependent road network are expressed by adopting the time functions.

S2: preprocessing the intersection nodes of the road network by combining the time-dependent road network based on an intersection end point connection algorithm, and connecting adjacent nodes of the intersection nodes to obtain an intersection connection table;

s3: establishing a vertex index for each node in a time-dependent road network to form a vertex index set, wherein the vertex index comprises a reverse connection lookup table and a forward connection lookup table, the reverse connection lookup table comprises all adjacent nodes which can reach the current node and a time function and road attributes of a road which passes through when the current node is reached, and the forward connection lookup table comprises all adjacent nodes which can be reached by the current node and a time function and road attributes of a road which passes through when the adjacent nodes are reached;

s4: according to the adjacent strategy point query algorithm, vertex indexes established for each node in the time-dependent road network are integrated, and adjacent strategy points of road nodes are queried to obtain a plurality of paths which can reach the strategy points;

s5: defining the domination relation among the paths, and inputting a source point, a destination, departure time, a strategy sequence, weather during departure and a required path sequence number of a query based on a path domination algorithm under the multi-background road network to obtain an optimal path sequence set containing the optimal paths with preset number.

In the embodiment of the present invention, the time-dependent road network specifically includes:

G_m＝(V_m，E_m，F_m，C_m，R_m)

wherein G is_mShowing time dependence of road network, V_mRepresenting a set of nodes, each node representing a different strategy point in the road network, E_mRepresenting a set of edges, each edge representing a path connecting two different nodes, for which edge e ═ v_i，v_j) Existence of a function of time F_m，v_iAnd v_jNodes, C, corresponding to two vertices representing the edge e_mPolicy attributes for storage nodes when there is a node V ∈ V_mWhen it is time, the policy attribute function C (v) returns the policy attribute of node v, R_mFor storing road attributes, when there is an edge E ∈ E_mThen the road attribute function r (e) returns the road attribute of edge e.

That is, a time-dependent road network under multiple backgrounds can be described as G_m＝(V_m，E_m，F_m，C_m，R_m). Referring to fig. 1, the policy attributes are nodes a and c of a supermarket, the policy attributes are nodes B and d of a bank, the policy attributes are nodes e and f of a restaurant, the nodes a and B represent intersections of intersections, and the nodes s and t represent source points and destinations, respectively.

In the embodiment of the present invention, a time function, a strategy sequence and a road attribute are defined for the time-dependent road network, where:

(1) the definition process of the time function is as follows:

defining a time function f on a time-dependent network_e(t) is a discrete function, and each edge e ═ v (v) in the time-dependent road network_i，v_j)∈E_mAll have a time function f_e(t_i)＝t_eIs used to indicate at time t_iSlave node v_iInitiating a query to reach node v_jThe required time is t_e；

Time function f of paths on time-dependent road network_e(t) is a simple discrete function, and in actual traffic conditions, the time-dependent function of the vehicle traffic flow on the same path follows a first-in first-out characteristic. In a real road network, the traffic conditions of roads are often distributed according to a certain rule along with time;

for f_e(t_i)＝t_eCan also be expressed as

I.e. at node v_jTime of departure minus at node v_iAnd at node v_jTime of stay to reach node v_jCan be expressed as Arr (v)_j)＝t_i+f_e(t_i)，Arr(v_j) Representing an arriving node v_jThe time of (d);

a path p is typically a set of edges p ═<e₁＝(v₁，v₂)，e₂＝(v₂，v₃)，...，e_k＝(v_i，v_j)>The travel time may be represented as trt (p) ═ Arr (v)_j)-t_iTrt (p) represents the travel time of the route p;

(2) the definition process for the policy sequence is:

in a real road network, each node has a policy attribute, a set of policy attributes of all vertexes on the road network is a policy set S, and a specific sequence is used for sequentially visiting the nodes in a trip process, so that the sequence is called a policy sequenceC, strategy sequence C ═<C₁，C₂，...，C_j>Indicates that C is required in the process of query₁，C₂，...，C_jThis sequence accesses in turn j corresponding attribute policy points, C_jRepresenting the jth policy point attribute, for each policy attribute in the policy sequence there is C_i∈S(1≤i≤j)，C_iRepresenting the ith strategy point attribute;

in a real road network, each node often has different policy attributes, such as restaurants, gas stations, movie theaters and the like, and a set of the policy attributes of all nodes on the road network is called a policy set S; defining | C | as the total number of strategies in the strategy sequence, and | C |_iI is a policy attribute of C_iThe number of nodes of (c);

(3) the definition process for the road attribute is as follows:

connection path between two nodes e ═ v_i，v_j) Then, the road attribute R (e) of the route e is R_i，R_iIndicating the road attribute of the route e.

The road attribute may be classified into an expressway, a trunk road, a branch road, a residential road, and the like, for example, for a road attribute set R ═ R<R₁，R₂，R₃>Each of which stands for highway, trunk and branch, e.g. e ═ s, a, with R (e) ═ R₃This road from node s to node a is denoted as a branch.

In the embodiment of the invention, the influence of weather factors on road selection comprises traffic influence and travel time influence; defining a weather factor influence coefficient for representing the influence of different weather on the running time of the roads with different attributes, wherein the weather factor influence coefficient is the ratio of the running time of a certain attribute road in normal weather to the running time of the same attribute road when the weather is w, and w represents the weather type.

The influence of weather conditions on daily trip schemes is large, so that the influence of weather on road selection is generally divided into two types: the first is traffic influence, in extreme weather conditions, people can give up roads with some attributes, for example, in rainy and snowy weather, people often cannot select a highway when going out; the second is a travel time influence, where different weather conditions have different influences on the travel time of roads with different attributes, for example, the travel speed on a road with different attributes in rainy days is slower than that in normal weather, so that the overall travel time is longer, the time to reach the next policy point (node) is also changed, and the subsequent path expansion is also influenced.

For example, the travel time of a road of a certain attribute in normal weather is

When the weather is w, the travel time of the path is

Weather factor influence coefficient

I.e. the velocity influence ratio

The coefficient reflects the influence of different weather on the driving time of roads with different attributes, and an influence coefficient set delta (delta) is established for the influence coefficient delta of the weather factor₁，δ₂，...，δ_n) Delta in the set_nRepresenting the influence coefficients under different weather and road properties.

For example, for time-dependent road network G in multiple contexts_m＝(V_m，E_m，F_m，C_m，R_m) The query in (C) above may be defined as Q ═ s, d, t, C, w, k, which indicates that, when the time t and the weather condition are w, the first k path sets from the vertex s to the vertex d and with less travel time are found

Assuming that the set of influence coefficients of weather at the time of query is δ ═ (δ)₁，δ₂，...，δ_n) Then each path in the set of paths

The travel time can be expressed as

For Path Presence in Path set ζ

At the same time, the path

And v₁＝s，v_qThe vertex traversed by d needs to satisfy the requirements of the policy sequence.

For example, a Top-k path sequence query Q ═ in a multiple context is initiated (s, t, 10:00,<su，ba，re>suw, 2), su shows supermarket, ba shows bank, re shows restaurant, snow shows weather is snow day, query returns from node s at 10:00 am on snow day, passes through supermarket, bank, restaurant, the policy sequence reaches the end point and takes 2 paths before the time, see fig. 1, multiple paths P conforming to the policy sequence can be obtained₁＝<s，a，d，f，t>、P₂＝<s，A，c，d，B，e，t>、P₃＝<s，a，A，b，e，t>Assuming that the traveling on snowy days does not pass through the highway, the influence coefficients delta of the weather on the main road and the branch road are 0.6 and 0.5 respectively, and the stay time at each strategy point does not exceed one hour, then due to P₁Passing through the highway, pruning can be carried out, P₂And P₃Respectively Trt (P)₂)＝216、Trt(P₃) 247, so the set of query paths is returned as { P₂，P₃}。

In the embodiment of the invention, the concrete steps of obtaining the intersection connection table comprise:

s201: sequentially traversing nodes v of road network_iIf node v_iIf there is a crossing node v in the adjacent points, the node v will be connected_iThe edge of the intersection node v is e₁；

S202: traversing each adjacent node v of intersection nodes v_jLet the edge e₂I.e. (v, v)_j) Join to tuple e_iIn (1), a new path is formed<v_i，v，v_i>；

S203: will be edge e₁And edge e₂And merging the time functions, and reserving the road attribute to obtain the intersection connection table. By using the obtained intersection connection table IEL _ list, the time function and road attribute from the intersection starting point to the intersection connection point can be directly obtained during query processing.

The intersection is an important road form in a road network, and as an intersection of two roads, the intersection plays a great role in traffic trip, and at present, more and more algorithms are used for searching and reasonably calibrating the intersection on the real-time road network. However, in the conventional path query paper, the intersection is not well processed, and even the existence of the road attribute of the intersection is ignored, so the invention provides an effective algorithm for processing the special road condition.

Because of the special feature of connecting multiple roads at the intersection, it cannot be directly processed, see fig. 3, which shows a cross intersection and its topological graph in the road network, in fig. 3, the central point of the cross intersection is a, and the vertices on the roads passing through the point can all be connected to each other, for example, there are three paths from point a through the central point a:<a，A，b>、<a，A，c>、<a，A，d>in the inquiry process, if traversing the adjacent points of each crossroad brings unnecessary calculation time and affects the real-time inquiry processing, therefore, the invention preprocesses the crossroad in the step, connects the adjacent points, and establishes the method as (A), (B), (C) and (D) (C) in the step<a，A，b>，f_aAb(t)，2)，(<a，A，c>，f_aAc(t)，2)，(<a，A，d>，f_aAd(t, 2)) a connection table in which f_aAb(t)、f_aAc(t)、f_aAd(t) each represent a function of time over the respective current path.

In the embodiment of the present invention, the reverse connection lookup table specifically includes:

L_in(v_j)＝(v_i，f_e(t)，R(e))

wherein L is_in(v_j) Indicating a reverse connection look-up table, L_in(v_j) In, v_iIndicating that node v can be reached_jThe edge e ═ v_i，v_j)，f_e(t) represents a time function of the edge e, and R (e) represents a road attribute of the edge e.

The forward connection lookup table specifically includes:

L_out(v_j)＝(v_i，f_e(t)，R(e))

wherein L is_out(v_j) Indicating a forward link look-up table, L_out(v_j) In, v_iRepresenting a node v_jReachable node, edge e ═ v_j，v_i)，f_e(t) represents a time function of the edge e, and R (e) represents a road attribute of the edge e.

The actual road network scale is often large, a large amount of time is consumed in the query process, and in order to improve the query processing efficiency, the method establishes the label index set for the nodes in the road network graph to assist in query. The reachable nodes of all the nodes occupy more storage space in the processing process, especially the space overhead is huge when the reachable nodes are used on a large-scale road network in reality, and the influence of weather on the road is difficult to calculate due to the combination of roads with different attributes, so that the method can only connect the road junction nodes, and establish a label index set for other nodes.

In the embodiment of the invention, according to an adjacent strategy point query algorithm, vertex indexes established for each node in a time-dependent road network are integrated, and adjacent strategy points of road nodes are queried to obtain a plurality of paths which can reach the strategy points, and the specific steps comprise:

s401: inputting the initial node v to be inquired currently_iNode v to be queried_iThe method comprises the following steps of (1) adjacent strategy attributes, query time t and current weather w; after the vertex index set of each node in the road network graph is obtained, the vertex index set can be established according to the node index setTwo labels L of_out(v) And L_in(v) And inquiring the adjacent strategy points of the road nodes.

S402: establishing a path set BIFS _ list, wherein the path set BIFS _ list is used for storing a node v_iAll paths to the next policy attribute node and the tuple consisting of the arrival time;

s403: traversing all nodes in the road network, and finding out the node v with the strategy attribute as the target attribute C_jStarting a loop, repeatedly traversing the node v_iForward link look-up table L_out(v_i) And node v_jReverse connection look-up table L_in(v_j) Finding a node v_a∈L_out(v_i) And node v_b∈L_in(v_j) When v is_a＝v_jTime, node v_iThe next adjacent node is the required strategy point, the connecting node v_iAnd node v_j(ii) a When v is_a＝v_bConnecting nodes, paths<v_i，v_a，v_a，v_j>Obtaining a path for the query; otherwise, the node v is connected_aAs the next query node L_out(v_i) Inquiring until all nodes are traversed;

s404: inquiring to obtain the road attribute R (v) under the current weather w_i，v_j) Obtaining corresponding weather factor influence coefficient according to road attribute and weather

Calculating to obtain the time of reaching the next strategy point, and adding the path into a path set BIFS _ list; is calculated by the formula

S405: and sorting the paths in the path set BIFS _ list according to the sequence of the arrival time from low to high to obtain a sorted path set BIFS _ list, and selecting the kth path which can reach the strategy point according to the sorted path set BIFS _ list.

When v is_a＝v_bThen, it is necessary to carry out routing on both ends (v)_i，v_a) And (v)_a，v_j) Connecting, calculating the path set by using the formula, and adding (p, t) into the path set BIFS _ list after processing; when the adjacent point is not searched, v is_aIs added to the next L_out(v_i) The item is queried in the above way while preserving the path<v_i，v_a>And reaches v_aT, all L are traversed in the above manner_out(v_i) And L_in(v_j) And (4) sorting all strategy point paths which can reach the corresponding attributes and are found finally according to the arrival time from low to high to obtain a sorted path set BIFS _ list, and then selecting the kth path which can reach the strategy points according to the sorted path set.

In the present invention, as for the dominance relationship between paths, for example, in fig. 1, there are two paths p₁＝<s，a，d>And p₂＝<s，A，c，d>When the query is made at a certain time t with the weather w, the travel time of the two paths is Trt (p)₁)＝25，Trt(p₂) In this context, it can be seen that path p is 40₁Required travel time ratio p₂Less, so path p can be considered in this context₁Dominating path p at node d₂。

In a multi-background road network, when the weather is assumed to be w, starting at a certain time t, a policy sequence C is satisfied<C₁，C₂，...，C_j>The path set of (1) comprises two paths

If it is not

And Trt (p)₁)≤Trt(p₂) Then say path p in the context of time tmeaty w₁At the vertex

To the dominant path p₂Can be represented as p₁＜p₂Term "p₁For dominant or skyline paths, called p₂Is the dominated path.

After a set of dominant route and dominated route is obtained, road expansion can be continued on the basis of the dominant route in priority, and road expansion on the dominated route is not considered for the moment. Assume two paths

And

existence of dominating relationship p₁＜p₂，Trt(p₁)≤Trt(p₂) After the two paths are continuously expanded on the basis of meeting the strategy sequence, the two expanded paths are obtained

And the expansion paths of the two paths are both p^*＝<v_i+1，v_i+2，...，t>. The travel time of the post-expansion route may be represented as Trt (p'₁)＝Trt(p₁)+Trt(p^*) And Trt (p'₂)＝Trt(p₂)+Trt(p^*) Because the travel time of the two routes on the extended route is the same, and the route p₁And p₂Since there is a dominant relationship, Trt (p'₁)≤Trt(p′₂). By the method, after the expansion path of the dominant path becomes the Top-k optimal path sequence, the dominant path is expanded, and because the expansion path generated by the dominant path is not superior to the path obtained by the expansion of the dominant path, after an optimal path is found, the dominant path is searched, and whether the expansion path of the path can become the next optimal path is calculated. The method for pruning in the query process can greatly reduce the calculation times and improve the qualityAnd searching efficiency.

When the path domination algorithm is used on the time-dependent road network, the algorithm may fail, because when the query is performed on the time-dependent road network, the difference of the query time background may cause the spent travel time to be different. Suppose two paths p₁，p₂Existence of dominating relationship p₁＜p₂When the query time is t, the travel time of the two routes is Trt (p)₁) And Trt (p)₂) When the two roads are expanded, the time for initiating the query is t₁＝t+Trt(p₁)，t₂＝t+Trt(p₂) From the dominance relation, t₁≤t₂. When t is₁＝t₂At path p^*Query time is the same in the upper extension, so Trt (p'₁)＝Trt(p′₂) Must be true. But when t is₁<t₂For time function on extended path

Because the query time background is different, we cannot determine

And

the relationship between the two expanded path weights Trt (p'₁)，Trt(p′₂) The magnitude relationship between them. This path-dominant relationship is therefore not applicable to the scenario described above.

In real life, the time function generated on the time-dependent road network is generally a linear function with a slow gradient, and the travel time does not change too much in a short time. When the background time difference of two queries is within an acceptable range, we can consider that the different query times do not affect the expansion after the dominant path, so a threshold value xi is set, and when the weight difference between the dominant path and the dominated path is less than the threshold value, i.e. Trt (p)₂)-Trt(p₁)≤Xi, we ignore the influence of different time backgrounds; when the threshold is exceeded, the two paths are considered to be not in a dominant relationship.

In the present invention, the Top-k path query algorithm based on the dominance relationship specifically includes the following steps:

for a Top-k path sequence query Q ═ (s, d, t, C, w, k) in a multiple background, the algorithm inputs the query's source and destination (s, d), departure time t, and policy sequence C ═<C₁，C₂，...，C_j>The weather w when the vehicle starts and the required path sequence number k, and outputting the first k optimal path sequence sets MCDR _ L. First, establish an initial query queue Q ←: (<s>T, 1), starting from the start s, the first sequence of paths at time t is queried. And then each node V E V establishes a dominant table and a dominated table V_＜cAnd v.HT_＞cAnd initializes both tables to null. The initialization query sequence set MCDR _ L is empty and the policy identification num is defined to be 0. When the query is started, it is first determined whether the queue Q is empty, and the number of sequences in the list L cannot be more than k. We mark the node popped from queue Q as ρ ═ (p ═<v₀，v₁，...，v_i>T, x), wherein p ═<v₀，v₁，...，v_i>Is a current path sequence which is already traveled, and t is a slave point v_iThe time of departure, x, represents the weighted ranking of the selected route from the last strategic point to the next. For the current path sequence p ═<v₀，v₁，...，v_i>First from point v_iExpanding, searching the nearest vertex meeting the policy requirement C, and recording the found point as v_i+1. If v is_i+1Node domination table v_i+1，HT_＜cFor an empty set, the expanded path p is defined as<v₀，v₁，...，v_i，v_i+1>Joining to node dominance table v_i+1，HT_＜cIn (1).

V of when path sequence_iD, after the strategy sequence is traversed, the path is the searched path sequence, and rho is inserted into the sequence L. Is connected withThen all the nodes are searched for in the dominated table v_j，HT_＞cIf the dominated path table of a certain node is not found to be empty, the sequence is taken out and is marked as (p' ═ m<v₀，v′₁，...，v_i>Trt (p), -) and adds it to queue Q. After expanding to a new node, if the path dominance table of the node is not empty, the path p weight in the dominance table is compared with the weight of the current path p ', if the travel time difference Trt (p ') -Trt (p) ≦ ξ of two paths, the two paths form a dominance relation, and the path p ' is added to the dominated table v of the current node_i+1，HT_＞cIn (1). The sequence is inserted into queue Q if no dominance is constructed. After the judgment, if the node number | p! p of the current node sequence p is non-existent>1, find v again_iFind it from v_i-1The strategy point with the x +1 being close is marked as p<v₀，v₁，...，v′_i>And adds the path to queue Q. After the query is finished, the algorithm returns the first k optimal path sequence sets MCDR _ L.

The invention processes the special road attribute of the crossroad and reduces the calculation redundancy brought by the crossroad in an efficient mode; the time-dependent road network definition is extended, expanded from the time dimension, added with road attributes, and the problem of path sequence query on the road network is solved; the path sequence query is expanded from a static road network to a time-dependent road network under multiple backgrounds, an optimized road domination algorithm is provided, and the problem of Top-k path sequence query on the time-dependent road network under multiple backgrounds is successfully solved.

According to the Top-k path sequence query method under the multiple backgrounds, firstly, a crossroad connection algorithm is used for simplifying the complex query problem caused by the existence of crossroads, secondly, forward and reverse query label indexes of nodes are established, the connection label indexes are used as a basic algorithm for querying adjacent strategy points, and finally, a pruning algorithm based on path domination is provided, and reasonable pruning is carried out through the path domination relation among the roads, so that the efficiency of road query is improved, and Top-k path sequence query in a time-dependent road network under the multiple backgrounds is efficiently realized.

The Top-k path sequence query system under the multiple backgrounds comprises a definition module, a connection module, an establishment module, a query module and a determination module.

The defining module is used for defining a time-dependent road network based on a time function, defining the time function, a strategy sequence and road attributes of the time-dependent road network, and defining the time-dependent road network path sequence query by taking a weather factor as a query condition; the connection module is used for preprocessing the intersection nodes of the road network based on an intersection end point connection algorithm and in combination with the time-dependent road network, and connecting adjacent nodes of the intersection nodes to obtain an intersection connection table; the establishing module is used for establishing a vertex index for each node in the time-dependent road network to form a vertex index set, wherein the vertex index comprises a reverse connection lookup table and a forward connection lookup table, the reverse connection lookup table comprises all adjacent nodes which can reach the current node and a time function and road attributes of a road which passes through when the current node is reached, and the forward connection lookup table comprises all adjacent nodes which can be reached by the current node and a time function and road attributes of a road which passes through when the adjacent nodes are reached; the query module is used for querying the adjacent strategy points of the road nodes according to the adjacent strategy point query algorithm and by integrating the vertex indexes established for each node in the time-dependent road network, so as to obtain a plurality of paths which can reach the strategy points; the determining module is used for defining the domination relation among the paths, inputting a source point, a destination, a departure time, a strategy sequence, weather during departure and a required path sequence number of the query based on a path domination algorithm under the multi-background road network, and obtaining an optimal path sequence set containing the optimal paths with the preset number.

In the embodiment of the present invention, the time-dependent road network specifically includes:

G_m＝(V_m，E_m，F_m，C_m，R_m)

wherein G is_mShowing time dependence of road network, V_mRepresenting a set of nodes, each node representing a different strategy point in the road network, E_mRepresenting a set of edges, each edge representing a path connecting two different nodes, for which edge e ═ v_i，v_j) Existence of a function of time F_m，v_iAnd v_jNodes, C, corresponding to two vertices representing the edge e_mPolicy attributes for storage nodes when there is a node V ∈ V_mWhen it is time, the policy attribute function C (v) returns the policy attribute of node v, R_mFor storing road attributes, when there is an edge E ∈ E_mThen the road attribute function r (e) returns the road attribute of edge e.

The above description is merely exemplary of the present application and is presented to enable those skilled in the art to understand and practice the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

Claims (10)

1. A Top-k path sequence query method under multiple backgrounds is characterized by comprising the following steps:

defining a time-dependent road network based on a time function, defining the time function, a strategy sequence and road attributes of the time-dependent road network, and defining time-dependent road network path sequence query by taking a weather factor as a query condition;

preprocessing the intersection nodes of the road network by combining the time-dependent road network based on an intersection end point connection algorithm, and connecting adjacent nodes of the intersection nodes to obtain an intersection connection table;

establishing a vertex index for each node in a time-dependent road network to form a vertex index set, wherein the vertex index comprises a reverse connection lookup table and a forward connection lookup table, the reverse connection lookup table comprises all adjacent nodes which can reach the current node and a time function and road attributes of a road which passes through when the current node is reached, and the forward connection lookup table comprises all adjacent nodes which can be reached by the current node and a time function and road attributes of a road which passes through when the adjacent nodes are reached;

according to the adjacent strategy point query algorithm, vertex indexes established for each node in the time-dependent road network are integrated, and adjacent strategy points of road nodes are queried to obtain a plurality of paths which can reach the strategy points;

defining the domination relation among the paths, and inputting a source point, a destination, departure time, a strategy sequence, weather during departure and a required path sequence number of a query based on a path domination algorithm under the multi-background road network to obtain an optimal path sequence set containing the optimal paths with preset number.

2. The Top-k path sequence query method under the multiple backgrounds of claim 1, wherein the time-dependent road network specifically comprises:

G_m＝(V_m，E_m，F_m，C_m，R_m)

wherein G is_mShowing time dependence of road network, V_mRepresenting a set of nodes, each node representing a road networkSame policy point, E_mRepresenting a set of edges, each edge representing a path connecting two different nodes, for which edge e ═ v_i，v_j) Existence of a function of time F_m，v_iAnd v_jNodes, C, corresponding to two vertices representing the edge e_mPolicy attributes for storage nodes when there is a node V ∈ V_mWhen it is time, the policy attribute function C (v) returns the policy attribute of node v, R_mFor storing road attributes, when there is an edge E ∈ E_mThen the road attribute function r (e) returns the road attribute of edge e.

3. The Top-k path sequence query method under multiple backgrounds of claim 2, wherein the defining of the time function, the strategy sequence and the road attribute is performed on the time-dependent road network, wherein:

the definition process of the time function is as follows:

defining a time function f on a time-dependent network_e(t) is a discrete function, and each edge e ═ v (v) in the time-dependent road network_i，v_j)∈E_mAll have a time function f_e(t_i)＝t_eIs used to indicate at time t_iSlave node v_iInitiating a query to reach node v_jThe required time is t_e；

The definition process for the policy sequence is:

in a real road network, each node has a policy attribute, a set of policy attributes of all vertexes on the road network is a policy set S, and when nodes are visited in sequence by a specific sequence in a trip process, the sequence is called a policy sequence C, and the policy sequence C is < C₁，C₂，...，C_j>. The expression requires C in the query process₁，C₂，...，C_jThis sequence accesses in turn j corresponding attribute policy points, C_jRepresenting the jth policy point attribute, for each policy attribute in the policy sequence there is C_i∈S(1≤i≤j)，C_iRepresenting the ith strategy point attribute;

the definition process for the road attribute is as follows:

connection path between two nodes e ═ v_i，v_j) Then, the road attribute R (e) of the route e is R_i，R_iIndicating the road attribute of the route e.

4. The method of claim 3, wherein the Top-k path sequence query under multiple backgrounds comprises:

the influence of the weather factors on road selection comprises traffic influence and travel time influence;

defining a weather factor influence coefficient for representing the influence of different weather on the running time of the roads with different attributes, wherein the weather factor influence coefficient is the ratio of the running time of a certain attribute road in normal weather to the running time of the same attribute road when the weather is w, and w represents the weather type.

5. The method for searching for Top-k path sequences under multiple backgrounds as claimed in claim 4, wherein the step of obtaining the intersection connection table comprises:

sequentially traversing nodes v of road network_iIf node v_iIf there is a crossing node v in the adjacent points, the node v will be connected_iThe edge of the intersection node v is e₁；

Traversing each adjacent node v of intersection nodes v_jLet the edge e₂I.e. (v, v)_j) Join to tuple e_iIn (1), a new path is formed<v_i，v，v_j＞；

Will be edge e₁And edge e₂And merging the time functions, and reserving the road attribute to obtain the intersection connection table.

6. The method of claim 5, wherein the reverse link lookup table comprises:

L_in(v_j)＝(v_i，f_e(t)，R(e))

wherein L is_in(v_j) Indicating a reverse connection look-up table, L_in(v_j) In, v_iIndicating that node v can be reached_jThe edge e ═ v_i，v_j)，f_e(t) represents a time function of the edge e, and R (e) represents a road attribute of the edge e.

7. The method for querying a Top-k path sequence under multiple backgrounds as claimed in claim 6, wherein said forward direction connection lookup table specifically comprises:

L_out(v_j)＝(v_i，f_e(t)，R(e))

wherein L is_out(v_j) Indicating a forward link look-up table, L_out(v_j) In, v_iRepresenting a node v_jReachable node, edge e ═ v_j，v_i)，f_e(t) represents a time function of the edge e, and R (e) represents a road attribute of the edge e.

8. The method according to claim 7, wherein the method for querying Top-k path sequences under multiple backgrounds is characterized in that the method for querying the strategy points adjacent to the road nodes by integrating the vertex indexes established for each node in the time-dependent road network according to the adjacent strategy point query algorithm to obtain multiple paths which can reach the strategy points, and comprises the following specific steps:

inputting the initial node v to be inquired currently_iNode v to be queried_iThe method comprises the following steps of (1) adjacent strategy attributes, query time t and current weather w;

establishing a path set BIFS _ list, wherein the path set BIFS _ list is used for storing a node v_iAll paths to the next policy attribute node and the tuple consisting of the arrival time;

traversing all nodes in the road network, and finding out the node v with the strategy attribute as the target attribute C_jStarting a loop, repeatedly traversing the node v_iForward link look-up table L_out(v_i) And node v_jReverse connection look-up table L_in(v_j) Finding a node v_a∈L_out(v_i) And node v_b∈L_in(v_j) When v is_a＝v_jTime, node v_iThe next adjacent node is the required strategy point, the connecting node v_iAnd node v_j(ii) a When v is_a＝v_bConnecting nodes, paths<v_i，v_a，v_a，v_j>Obtaining a path for the query; otherwise, the node v is connected_aAs the next query node L_out(v_i) Inquiring until all nodes are traversed;

inquiring to obtain the road attribute R (v) under the current weather w_i，v_j) Obtaining corresponding weather factor influence coefficient according to road attribute and weather

Calculating to obtain the time of reaching the next strategy point, and adding the path into a path set BIFS _ list;

and sorting the paths in the path set BIFS _ list according to the sequence of the arrival time from low to high to obtain a sorted path set BIFS _ list, and selecting the kth path which can reach the strategy point according to the sorted path set BIFS _ list.

9. A Top-k path sequence query system under multiple backgrounds, comprising:

the system comprises a definition module, a time function module and a query module, wherein the definition module is used for defining a time-dependent road network based on a time function, defining the time function, a strategy sequence and road attributes of the time-dependent road network and defining the time-dependent road network path sequence query by taking a weather factor as a query condition;

the connection module is used for preprocessing the intersection nodes of the road network based on an intersection end point connection algorithm and in combination with the time-dependent road network, and connecting the adjacent nodes of the intersection nodes to obtain an intersection connection table;

the system comprises an establishing module, a searching module and a judging module, wherein the establishing module is used for establishing a vertex index for each node in a time-dependent road network to form a vertex index set, the vertex index comprises a reverse connection lookup table and a forward connection lookup table, the reverse connection lookup table comprises all adjacent nodes which can reach the current node and a time function and road attributes of a road which passes through when the current node is reached, and the forward connection lookup table comprises all adjacent nodes which can reach the current node and a time function and road attributes of a road which passes through when the adjacent node is reached;

the query module is used for querying the adjacent strategy points of the road nodes according to the adjacent strategy point query algorithm and by integrating the vertex indexes established for each node in the time-dependent road network to obtain a plurality of paths which can reach the strategy points;

and the determining module is used for defining the domination relationship among the paths, inputting a source point, a destination, a departure time, a strategy sequence, weather during departure and a required path sequence number of the query based on a path domination algorithm under the multi-background road network, and obtaining an optimal path sequence set containing the optimal paths with the preset number.

10. The Top-k path sequence query system under multiple backgrounds of claim 9, wherein the time-dependent road network specifically is:

G_m＝(V_m，E_m，F_m，C_m，R_m)

wherein G is_mShowing time dependence of road network, V_mRepresenting a set of nodes, each node representing a different strategy point in the road network, E_mRepresenting a set of edges, each edge representing a path connecting two different nodes, for which edge e ═ v_i，v_j) Existence of a function of time F_m，v_iAnd v_jNodes, C, corresponding to two vertices representing the edge e_mPolicy attributes for storage nodes when there is a node V ∈ V_mWhen it is time, the policy attribute function C (v) returns the policy attribute of node v, R_mFor storing road attributes, when there is an edge E ∈ E_mThen the road attribute function r (e) returns the road attribute of edge e.

Images