CN107883956A - A kind of routing resource and system based on dijkstra's algorithm - Google Patents

Abstract

The invention discloses a path selection method and system based on the Dijkstra algorithm. Firstly, an indoor path model is obtained. The indoor path model includes the weights of each selection factor between each node, and the selection factors in the indoor path model include path distance. and the congestion degree of the path, and then calculate the comprehensive weight between each node of the indoor path model, and then replace the weight of the path distance in the Dijkstra algorithm with the corresponding comprehensive weight to perform the Dijkstra algorithm, and select a total comprehensive weight the smallest path. The present invention comprehensively considers multiple selection factors affecting path planning, and can obtain a comprehensive optimal path that is more in line with the actual indoor environment and individual needs of users.

Description

A method and system for path selection based on Dijkstra algorithm

technical field

The present invention relates to the field of navigation, in particular to the aspect of path selection in the navigation process, and more specifically, to a path selection method and system based on Dijkstra algorithm.

Background technique

With the acceleration of urbanization, there are more and more large indoor buildings, and their internal structures are becoming more and more complex. People spend more and more time indoors, so indoor navigation path planning has practical significance. Among the path planning algorithms, the Dijkstra algorithm is an effective algorithm for finding the shortest path from a source point to other points in a directed graph. The algorithm adapts to network topology changes and has stable performance. It is a classic path planning algorithm. algorithm.

The idea of Dijkstra's algorithm is: Let G=(V, E) be a weighted directed graph, divide the set V of each vertex in the figure into two groups, and the first group is the set of vertices for which the shortest path has been obtained (represented by S, Initially, there is only one source point in S, and each time a shortest path is obtained, it will be added to the set S until all vertices are added to S, and the algorithm is over), the second group is the rest of the undetermined shortest path For the vertex set (denoted by U), the vertices of the second group are added to S in the ascending order of the shortest path length. In the process of joining, always keep the shortest path length from the source point v to each vertex in S not greater than the shortest path length from the source point v to any vertex in U. In addition, each vertex corresponds to a distance. The distance of the vertex in S is the shortest path length from v to this vertex. The distance of the vertex in U is from v to this vertex, only including the current vertex in S as the intermediate vertex. Shortest path length.

Dijkstra algorithm steps:

a. Initially, S only contains the source point, that is, S={v}, and the distance of v is 0. U contains other vertices except v, that is: U={other vertices}, if v has an edge with vertex u in U, then <u, v> has a weight not equal to ∞, if u is not adjacent to the outgoing edge of v point, then the weight of <u, v> is ∞.

b. Select a vertex k with the smallest distance v from U, and add k to S (the selected distance is the shortest path length from v to k).

c. Take k as the newly considered intermediate point, modify the distance of each vertex in U; if the distance from source point v to vertex u (passing vertex k) is shorter than the original distance (not passing vertex k), modify the distance of vertex u The distance value, the modified distance value is the distance of vertex k plus the weight on the edge.

d. Repeat steps b and c until all vertices are included in S.

The above-mentioned Dijkstra algorithm can obtain the shortest path from the starting point to the end point, but in the actual indoor navigation path planning problem, due to the complexity and diversity of the indoor environment, the actual indoor path selection is also personalized and diverse, making People's choice of indoor paths is not only based on path distance as the only criterion, but also some other factors that affect path selection. Only considering path distance cannot satisfy people's indoor path selection needs.

Contents of the invention

The technical problem to be solved by the present invention is that, in view of the complexity and diversity of the above-mentioned existing indoor environment, indoor path selection presents individuality and diversification, and the existing path planning method only uses the path distance as the only standard for indoor path planning Navigation, the technical defect that only considering the path distance cannot satisfy people's indoor path selection needs, provides a path selection method and system based on Dijkstra algorithm.

According to one aspect of the present invention, the present invention provides a kind of route selection method based on Dijkstra algorithm for solving its technical problem, comprises the following steps:

S1. Obtain an indoor path model, the indoor path model includes the weights of each selection factor between each node, and the selection factors in the indoor path model include path distance and path congestion;

S2. Calculate the comprehensive weight between each node of the indoor path model, wherein the comprehensive weight between any node p and q is calculated by the following formula:

f＝ω ₁ ×f ₁ +ω ₂ ×f ₂ +...+ω _n ×f _n ,

In the formula, f represents the comprehensive weight, n represents the total number of selected factors, f ₁ , f ₂ , ... and f _n are the weights of each factor between node p and node q respectively, ω ₁ , ω ₂ , ... and ω _n are weights corresponding to f ₁ , f ₂ , ... and f _n respectively.

S3. Perform the Dijkstra algorithm with the weight corresponding to the comprehensive weight replacing the path distance in the Dijkstra algorithm, and select a path with the smallest total comprehensive weight.

Further, in the path selection method based on the Dijkstra algorithm of the present invention, the normalized weight of the kth selection factor is obtained by the following method:

In the formula, k=1, 2, ..., n, S _k is to utilize 1～9 levels to judge the matrix standard degree, obtain the value of the matrix standard degree of the kth selection factor respectively with respect to all selection factors respectively, then will The sum obtained by adding all matrix standard degree values for this selection factor.

Further, in the route selection method based on Dijkstra algorithm of the present invention, the above-mentioned multiple selection factors can be summarized into three selection factors, which are respectively the distance to the route, the congestion degree of the route and the preference degree of the route.

Further, in the path selection method based on Dijkstra algorithm of the present invention, the weight of path distance is ω ₁ =0.633, the weight of path congestion is ω ₂ =0.261, and the weight of path preference is ω ₃ =0.106.

Further, in the path selection method based on the Dijkstra algorithm of the present invention, in the process of obtaining the weights of each selection factor, a step is also included: checking the calculated normalized weights by judging the consistency of the normalized weights Whether the final weight conforms to the actual importance between the weight types, if so, use the weight calculated this time as the final value of the weight, otherwise, obtain the value of the re-selected matrix standard degree to calculate the weight;

The method of making a consistency judgment is as follows:

judge Whether the value of is less than 0.1, if it is consistent, otherwise it is not consistent;

in, RI is equal to the value of the n-order matrix in the average random consistency table, λ _max is the largest characteristic root of the judgment matrix, and the element of the i-th row and the j-th column of the judgment matrix is the matrix of the i-th selection factor relative to the j-th selection factor The value of standard degree, i=1, 2,..., n, j=1, 2,..., n.

According to another aspect of the present invention, the present invention also provides a kind of path selection system based on Dijkstra algorithm for solving its technical problem, comprises the following steps:

A path model acquisition module, used to obtain an indoor path model, the indoor path model includes the weight of each selection factor between each node, and the selection factors in the indoor path model include path distance and path congestion;

The comprehensive weight calculation module is used to calculate the comprehensive weight between each node of the indoor path model, wherein the comprehensive weight between any node p and q is calculated by the following formula:

f＝ω ₁ ×f ₁ +ω ₂ ×f ₂ +...+ω _n ×f _n ,

In the formula, f represents the comprehensive weight, n represents the total number of selected factors, f ₁ , f ₂ , ... and f _n are the weights of each factor between node p and node q respectively, ω ₁ , ω ₂ , ... and ω _n are weights corresponding to f ₁ , f ₂ , ... and f _n respectively.

The Dijkstra algorithm module is used to replace the weight of the path distance in the Dijkstra algorithm corresponding to the comprehensive weight to perform the Dijkstra algorithm, and select a path with the smallest total comprehensive weight.

Further, in the path selection system based on the Dijkstra algorithm of the present invention, the normalized weight of the kth selection factor is obtained by the following method:

In the formula, k=1, 2, ..., n, S _k is to utilize 1～9 grades to judge the matrix standard degree, obtain the value of the matrix standard degree of the kth selection factor respectively relative to all selection factors respectively, and then The sum obtained by adding all matrix standard degree values for this selection factor.

Further, in the route selection system based on the Dijkstra algorithm of the present invention, the above multiple selection factors can be summarized into three selection factors, which are respectively route distance, route congestion degree and route preference degree.

Further, in the route selection system based on Dijkstra algorithm of the present invention, the weight of route distance is ω ₁ =0.633, the weight of route congestion is ω ₂ =0.261, and the weight of route preference is ω ₃ =0.106.

Further, in the path selection system based on the Dijkstra algorithm of the present invention, in the process of obtaining the weights of each selection factor, a step is also included: checking the normalized weight obtained by the calculation by performing consistency on the normalized weight Whether the weight of the weight conforms to the actual importance between the weight types, if so, the weight calculated this time is used as the final value of the weight, otherwise, the value of the standard degree of the re-selected matrix is obtained to calculate the weight;

The method of making a consistency judgment is as follows:

judge Whether the value of is less than 0.1, if it is consistent, otherwise it is not consistent;

in, RI is equal to the value of the n-order matrix in the average random consistency table, λ _max is the largest characteristic root of the judgment matrix, and the element of the i-th row and the j-th column of the judgment matrix is the matrix of the i-th selection factor relative to the j-th selection factor The value of standard degree, i=1, 2,..., n, j=1, 2,..., n.

The route selection method and system based on the Dijkstra algorithm of the present invention first obtains the indoor route model, the indoor route model includes the weights of each selection factor between each node, and the selection factors in the indoor route model include the route distance and the weight of the route Then calculate the comprehensive weight between each node of the indoor path model, and then replace the weight of the path distance in the Dijkstra algorithm with the corresponding comprehensive weight to perform the Dijkstra algorithm, and select a path with the smallest total comprehensive weight . The present invention comprehensively considers multiple selection factors affecting path planning, and can obtain a comprehensive optimal path that is more in line with the actual indoor environment and individual needs of users.

Description of drawings

The present invention will be further described below in conjunction with accompanying drawing and embodiment, in the accompanying drawing:

Fig. 1 is the flowchart of an embodiment of the path selection method based on Dijkstra algorithm of the present invention;

Fig. 2 is a path network topology diagram of the present invention.

Detailed ways

In order to have a clearer understanding of the technical features, purposes and effects of the present invention, the specific implementation manners of the present invention will now be described in detail with reference to the accompanying drawings.

Referring to FIG. 1 , it is a flowchart of an embodiment of a path selection method based on the Dijkstra algorithm of the present invention. The path selection method of the present embodiment mainly includes the following steps:

S1. Obtain an indoor path model, the indoor path model includes the weights of each selection factor between each node, and there are at least two selection factors.

Due to the diversity and complexity of indoor environments, users are the main body of indoor activities, and indoor path selection is often more flexible and autonomous. Indoor navigation path planning aims to plan a suitable route from the starting point to the target point for users , so that users can walk shorter paths, use less time, and better fit people's different preferences for paths. In this embodiment, the weight of the path distance, the weight of the congestion degree of the path, and the weight of the preference degree of the path are selected as the three weight indicators that affect indoor path planning for illustration, so as to obtain a more comprehensive and more in line with the user's indoor It should be understood that the influencing factors of the present invention include but are not limited to these three influencing factors.

The path distance reflects the length of the road section, and is the most direct factor in the navigation path planning problem. It is simple and intuitive, and can be calculated through the relevant measurement of the road section. The weight value of path distance is the basis of all navigation path planning, and it is a factor that users will definitely consider when making path selection. Usually, the smaller the value is, the more favorable it is for the path selection. The traditional Dijkstra algorithm selects the path distance as the only weight for navigation path planning to obtain the shortest path from the starting point to the destination point. However, due to the complexity and diversity of the indoor environment, users often do not only consider the weight of the path distance as a factor for path selection. Therefore, it is necessary to analyze other influencing factors of indoor navigation path planning and plan a more comprehensive comprehensive optimization path. According to the length of the path, the path can be divided into multiple levels, and each level corresponds to a weight value. The longer the path, the larger the value.

The degree of congestion is a conceptual value that comprehensively reflects the smoothness or congestion of the road network. Combined with the weight of the degree of congestion for path optimization, it can effectively avoid congested roads and save travel time. Beijing released the "Urban Road Traffic Operation Evaluation Index System" in April 2011 and officially implemented it in August of the same year. The traffic operation index is used to comprehensively reflect the traffic congestion of the road network. The urban road traffic operation evaluation index system will traffic congestion The degree is qualitatively divided into five levels, the value ranges from 0 to 10, and every two numbers are divided into one level, corresponding to "smooth traffic", "basically smooth traffic", "slight congestion" and "moderate congestion" , "serious congestion", the smaller the value, the smoother the traffic road, and the higher the value, the more serious the traffic congestion. Referring to the definition scheme of congestion degree in the traffic evaluation index system in outdoor navigation, the specific definition rules of indoor road congestion degree for indoor navigation path planning can be determined. In the field of outdoor navigation, the traffic congestion mainly reflects the real-time traffic congestion situation through the size of the traffic flow. Similarly, in the indoor route navigation, the present invention reflects the indoor congestion through the size of the human flow. In practical applications, the indoor road congestion index can be obtained by statistically analyzing a large amount of data by actually measuring the flow of people in a certain building, and then reflecting the characteristics of the indoor building congestion. With reference to the definition rules of urban road traffic congestion, this embodiment divides the indoor congestion evaluation index into five grades, and the evaluation indicators are: "smooth", "basically smooth", "slight congestion", "moderate congestion", "Severe congestion". The value range of the congestion degree index is [0, 10], and every interval of two numbers is divided into a grade. The value of the congestion degree index can be directly used as the weight of the influencing factor. The closer the value of the congestion degree is to 0, the smoother the road section is, and the closer the value is to 10, the more congested the road section is. The corresponding relationship between congestion degree and congestion index is shown in Table 1:

Table 1 Indoor road network congestion evaluation table

In actual indoor navigation path planning, due to the diversity and complexity of indoor building structures and different use scenarios, users are the main body of path decision-making, and users' indoor activities often have strong autonomy and individuality. Preference, user preference will also affect the route selection of indoor navigation, so the user's personalized preference cannot be ignored, for example: in large indoor shopping malls, some users prefer to choose the main road or the road section near some shops, elevators, etc. The user preference represents the user's liking for the road section, and is a subjective decision-making index provided to the user. Find a path that better meets user preferences under a certain time or distance constraint, and find a balance between path consumption (time or distance) and user individual needs. This planning idea combines the characteristics of the indoor environment and enhances the path. The sense of experience reflects the humanized thought. Due to the complexity of the indoor environment and the diversification and personalization of indoor activities of users, the selection of indoor routes is highly subjective and real-time. At the same time, the present invention mainly focuses on the analysis of the influence of user preference on indoor route planning. An indoor path navigation model with comprehensive weights is established. Therefore, the present invention adopts a rule-based delineation method to qualitatively divide user preference into nine levels, and the value assignment range is 1 to 9, and each value corresponds to a first-level preference The corresponding relationship between user preference degree and evaluation index is as follows: the smaller the value of user preference degree, the higher the degree of user preference, that is, the more the user likes it, the greater the user's choice tendency; the larger the value of user preference degree, the worse the degree of preference , that is, the more disliked, the smaller the user's tendency to choose. In specific indoor navigation route planning, different users can be obtained to set corresponding preference values according to the subjective weighting method, and the preference value is also the weight value of the degree of preference of the route in the present invention.

S2. Calculate the comprehensive weight between each node of the indoor path model, wherein the comprehensive weight between any node p and q is calculated by the following formula:

f＝ω ₁ ×f ₁ +ω ₂ ×f ₂ +...+ω _n ×f _n ,

In the formula, f represents the comprehensive weight, n represents the total number of selected factors, f ₁ , f ₂ , ... and f _n are the weights of each factor between node p and node q respectively, ω ₁ , ω ₂ , ... and ω _n are weights corresponding to f ₁ , f ₂ , ... and f _n respectively. If ω ₁ , ω ₂ , and ω ₃ are normalized, then ω ₁ +ω ₂ +ω ₃ =1.

The present invention uses the analytic hierarchy process to first qualitatively determine the importance of each pair of weights, and then convert the qualitative analysis into quantitative data calculation to obtain the impact of specific weight types on the comprehensive weight. Using AHP to solve the multi-objective linear weighted weight types in the indoor navigation path planning problem has an influence on the comprehensive weight. The specific steps are as follows:

To construct the judgment matrix of each weight type, first of all, the standard degree of the judgment matrix must be introduced, and the standard degree of the judgment matrix from 1 to 9 levels proposed by T.L.Saaty et al. (see the book "The Analytic Hierarchy Process", author T.L.Saaty), as follows Table 2 shows:

Table 2 1st to 9th grade judgment matrix standard degree

According to the judgment matrix standard degree table, the judgment matrix of each weight type is constructed, as shown in Table 3 below:

Table 3 Judgment matrix of each weight type

As can be seen from the table, this embodiment presets the relative importance of each pair of weights, such as: the weight of the path distance is more than the weight of the congestion degree of the path and the weight of the preference degree of the path The degree of influence on the comprehensive weight is higher, and the weight of the path congestion degree is slightly higher than the degree of influence of the path preference degree.

Then use the canonical column average method to calculate, and the calculation steps to find the combination weight coefficient are as follows:

First, the judgment matrix is summed column by column, and the following table 4 is obtained:

Table 4 Matrix column summation

f f1 f2 f3 f1 1 3 5 f2 1/3 1 3 f3 1/5 1/3 1 sum 23/15 13/3 9

Normalize each column and sum the rows to obtain the feature vector, as shown in Table 5 below:

Table 5 column normalization

f f1 f2 f3 sum f1 15/23 9/13 5/9 1.9 f2 5/23 3/13 1/3 0.782 f3 3/23 1/13 1/9 0.318

Then normalize the eigenvectors to obtain the combination weight coefficient, that is, the value of each influence degree:

ω ₁ =0.633, ω ₂ =0.261, ω ₃ =0.106.

The consistency judgment is made on the influence coefficient, and the consistency judgment is used to check whether the calculated relative weight coefficient conforms to the actual importance between the weight types, so as to check the correctness of the multi-weight fusion. Introduce a unified average random consistency table (see the article "Beynon MJ. An analysis of distributions of priority values from alternative comparison scales within AHP" in the journal "European Journal of Operational Research", 2002,140(1):104-117.) , as shown in Table 6:

Table 6 Average Random Consistency Table

Table 6 average random consistency table (continued)

Table 6 average random consistency table (continued)

The formula for checking consistency is:

in

n is the number of parameters, that is, the number of weight types, RI is equal to the value of the n-order matrix in the average random consistency table, and λ _max is the largest characteristic root of the judgment matrix. If CR<0.1, the judgment matrix is consistent, otherwise it is inconsistent. When they are consistent, the weight calculated this time is used as the final value of the weight, otherwise, the value of the standard degree of the reselected matrix is obtained to calculate the weight. because

It can be obtained that λ _max = 3.0556. Substituting λ _max and n into the above formula can obtain CI = 3.0556-3/3-1 = 0.0535, and substituting CI into the above formula can obtain CR = 0.0535/0.52 = 0.053<0.1. It can be seen that the judgment matrix obtained through the above calculation satisfies the consistency, which is consistent with the actual importance of each weight value, and its correctness is verified. To sum up, the comprehensive weight of the improved indoor navigation path planning model obtained from the calculation formula of the combined weight coefficient comprehensive weight is as follows:

F＝0.633*f ₁ +0.261*f ₂ +0.106*f ₃

S3. Perform the Dijkstra algorithm with the weight corresponding to the comprehensive weight replacing the path distance in the Dijkstra algorithm, and select a path with the smallest total comprehensive weight. This step is completely consistent with the Dijkstra algorithm, except that the weight value of the path distance in the Dijkstra algorithm is replaced by the comprehensive weight value.

The present invention adopts the following indoor path network topology diagram to carry out specific example simulation tests, and the path network topology diagram is shown in Figure 2 below. Among them, node 0 in the road network is set as the starting node, and the three values marked on the road section between two nodes represent the path distance, congestion degree, and user preference weight respectively.

First, the above road network information data is preprocessed to obtain a multidimensional array of three weights of path distance, congestion degree, and user preference degree, as shown in the following formula, where if two nodes are not adjacent, the value is defined as infinity.

Substituting the weights of each road section between two adjacent nodes into the comprehensive weight calculation formula of the established indoor navigation path planning model: F＝0.633*f ₁ +0.261*f ₂ +0.106*f ₃ , the two adjacent nodes can be obtained The comprehensive weight of each road section between nodes is obtained to obtain the matrix formula of the comprehensive weight of the path network topology map. As shown in the following formula:

The value in the formula represents the comprehensive weight of the section between two nodes in the indoor navigation path planning model, and the comprehensive weight of non-adjacent nodes is infinite. Through the above matrix formula, combined with the Dijkstra algorithm, the path with the smallest comprehensive weight is obtained, that is, a comprehensive optimal path that is more in line with the individual needs of users is planned from the start node to the target node with multiple weight types. Set 0 as the source node, and the simulation test results are shown in Table 7 below:

Table 7 Improved model test results

Table 7 shows the comprehensive optimal path from the source node to each node. Taking the starting node 0 to the destination node 9 as an example, the optimal path that combines multiple weight types is a path from node 0, passing through node 3, passing through node 5, then going to node 6, and finally reaching node 9.

The traditional Dijkstra algorithm only considers the only weight of the path length for path planning, and expands from the starting point to the outer layer until it reaches the end point, and obtains a shortest path from the starting point to the end point. The traditional Dijkstra algorithm path planning model is used to solve the planned path from the starting node 0 to each target node in the path network topology graph, and the results are shown in Table 8 below.

Table 8 Traditional model test results

Also taking from the starting node 0 to the destination node 9 as an example, the planned path obtained by the traditional Dijkstra algorithm path planning model is a path from node 0, passing through node 2, passing through node 5, passing through node 6, and finally reaching node 9.

The test results of the optimized path obtained by the improved indoor navigation path planning model and the optimized path obtained by the traditional Dijkstra algorithm path planning are compared and analyzed. Taking from the starting node 0 to the destination node 4 as an example, the optimized path obtained under the improved model is from the starting node 0 through node 3 and finally reaches the destination node 4; the path obtained under the traditional Dijkstra algorithm model is from the starting node 0 passes through node 1 and finally reaches the destination node 4. It can be seen that when the path distance is considered as the only factor, 0-1-4 is the shortest path from the starting node 0 to the destination node 4. In this case, only the length of the path is considered, but due to the indoor environment Complexity and diversity. As the main body of indoor activities, users have increasingly strong individual needs. Road congestion and user preference are important factors in indoor route navigation problems and need to be taken into consideration. The optimized path planned by the improved indoor navigation planning model is 0-3-4. Although this path is not the shortest, its congestion and user preference are relatively small, that is, the road is relatively smooth and users prefer it. Therefore, the The path 0-3-4 in the road network is a comprehensive optimal path from the starting node 0 to the destination node 4. In the indoor environment, the user's choice of the indoor path is not only to choose the shortest path, but also consider the road congestion, personal preference and other important factors on the basis of the path distance. The improved indoor navigation planning model It avoids the singleness problem that the traditional Dijkstra algorithm only considers the path distance, and comprehensively considers the influence of path distance, congestion, and user preference on indoor navigation path planning, which is more comprehensive and can better meet user needs.

The present invention also correspondingly provides a path selection system based on Dijkstra algorithm, comprising:

A route model acquiring module, configured to acquire an indoor route model, the indoor route model includes the weights of each selection factor between each node, and there are at least two selection factors;

The comprehensive weight calculation module is used to calculate the comprehensive weight between each node of the indoor path model, wherein the comprehensive weight between any node p and q is calculated by the following formula:

f＝ω ₁ ×f ₁ +ω ₂ ×f ₂ +...+ω _n ×f _n ,

In the formula, f represents the comprehensive weight, n represents the total number of selected factors, f ₁ , f ₂ , ... and f _n are the weights of each factor between node p and node q respectively, ω ₁ , ω ₂ , ... and ω _n are weights corresponding to f ₁ , f ₂ , ... and f _n respectively.

The Dijkstra algorithm module is used to replace the weight of the path distance in the Dijkstra algorithm corresponding to the comprehensive weight to perform the Dijkstra algorithm, and select a path with the smallest total comprehensive weight.

Embodiments of the present invention have been described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementations, and the above-mentioned specific implementations are only illustrative, rather than restrictive. Those of ordinary skill in the art will Under the enlightenment of the present invention, many forms can also be made without departing from the gist of the present invention and the protection scope of the claims, and these all belong to the protection of the present invention.

Claims (10)

1. a path selection method based on Dijkstra algorithm, is characterized in that, comprises the following steps: S1. Obtain an indoor path model, the indoor path model includes the weights of each selection factor between each node, and the selection factors in the indoor path model include path distance and path congestion; S2. Calculate the comprehensive weight between each node of the indoor path model, wherein the comprehensive weight between any node p and q is calculated by the following formula: f＝ω ₁ ×f ₁ +ω ₂ ×f ₂ +...+ω _n ×f _n , In the formula, f represents the comprehensive weight, n represents the total number of selected factors, f ₁ , f ₂ , ... and f _n are the weights of each factor between node p and node q respectively, ω ₁ , ω ₂ , ... and ω _n are the weights corresponding to f ₁ , f ₂ , ... and f _n respectively; S3. Perform the Dijkstra algorithm with the weight corresponding to the comprehensive weight replacing the path distance in the Dijkstra algorithm, and select a path with the smallest total comprehensive weight. 2. path selection method according to claim 1, is characterized in that, the weight after the normalization of the k selection factor obtains by following method: <mrow><msub><mi>&amp;omega;</mi><mi>k</mi></msub><mo>=</mo><mfrac><msub><mi>S</mi><mi>k</mi></msub><mrow><msub><mi>S</mi><mn>1</mn></msub><mo>+</mo><msub><mi>S</mi><mn>2</mn></msub><mo>+</mo><mo>...</mo><mo>+</mo><msub><mi>S</mi><mi>n</mi></msub></mrow></mfrac><mo>,</mo></mrow> In the formula, k=1, 2, ..., n, S _k is to utilize 1～9 grades to judge the matrix standard degree, obtain the value of the matrix standard degree of the kth selection factor respectively relative to all selection factors respectively, and then The sum obtained by adding all matrix standard degree values for this selection factor. 3 . The path selection method according to claim 2 , wherein the plurality of selection factors are three selection factors, which are respectively path distance, path congestion degree, and path preference degree. 4 . 4. The path selection method according to claim 3, wherein the weight of path distance is ω ₁ =0.633, the weight of path congestion is ω ₂ =0.261, and the weight of path preference is ω ₃ =0.106 . 5. The path selection method according to claim 2, characterized in that, in the process of obtaining the weight of each selection factor, it also includes the step of: checking the calculated normalized weight by performing a consistency judgment on the normalized weight Whether the optimized weight is in line with the actual importance between the weight types, if so, the weight calculated this time is used as the final value of the weight, otherwise, the value of the standard degree of the reselected matrix is obtained to calculate the weight; The method of making a consistency judgment is as follows: judge Whether the value of is less than 0.1, if it is consistent, otherwise it is not consistent; in, RI is equal to the value of the n-order matrix in the average random consistency table, λ _max is the largest characteristic root of the judgment matrix, and the element of the i-th row and the j-th column of the judgment matrix is the matrix of the i-th selection factor relative to the j-th selection factor The value of standard degree, i=1, 2,..., n, j=1, 2,..., n. 6. A path selection system based on Dijkstra algorithm, characterized in that, comprising: The path model acquisition module is used to obtain an indoor path model, the indoor path model includes the weight of each selection factor between each node, and the selection factors in the indoor path model include path distance and path congestion; The comprehensive weight calculation module is used to calculate the comprehensive weight between each node of the indoor path model, wherein the comprehensive weight between any node p and q is calculated by the following formula: f＝ω ₁ ×f ₁ +ω ₂ ×f ₂ +...+ω _n ×f _n , In the formula, f represents the comprehensive weight, n represents the total number of selected factors, f ₁ , f ₂ , ... and f _n are the weights of each factor between node p and node q respectively, ω ₁ , ω ₂ , ... and ω _n are weights corresponding to f ₁ , f ₂ , ... and f _n respectively. The Dijkstra algorithm module is used to replace the weight of the path distance in the Dijkstra algorithm corresponding to the comprehensive weight to perform the Dijkstra algorithm, and select a path with the smallest total comprehensive weight. 7. path selection system according to claim 6, is characterized in that, the weight after the normalization of the k selection factor obtains by following method: <mrow><msub><mi>&amp;omega;</mi><mi>k</mi></msub><mo>=</mo><mfrac><msub><mi>S</mi><mi>k</mi></msub><mrow><msub><mi>S</mi><mn>1</mn></msub><mo>+</mo><msub><mi>S</mi><mn>2</mn></msub><mo>+</mo><mo>...</mo><mo>+</mo><msub><mi>S</mi><mi>n</mi></msub></mrow></mfrac><mo>,</mo></mrow> In the formula, k=1, 2, ..., n, S _k is to utilize 1～9 grades to judge the matrix standard degree, obtain the value of the matrix standard degree of the kth selection factor respectively relative to all selection factors respectively, and then The sum obtained by adding all matrix standard degree values for this selection factor. 8. The path selection system according to claim 7, wherein the plurality of selection factors are three selection factors, namely path distance, path congestion degree and path preference degree. 9. The path selection system according to claim 8, wherein the weight of path distance is ω ₁ =0.633, the weight of path congestion is ω ₂ =0.261, and the weight of path preference is ω ₃ =0.106 . 10. The path selection system according to claim 7, characterized in that, in the process of obtaining the weights of each selection factor, it also includes the step of: checking the calculated normalization by judging the consistency of the normalized weights Whether the optimized weight is in line with the actual importance between weight types, if so, the weight calculated this time will be used as the final value of the weight, otherwise, the value of the standard degree of the re-selected matrix will be obtained to calculate the weight; The method of making a consistency judgment is as follows: judge Whether the value of is less than 0.1, if it is consistent, otherwise it is not consistent; in, RI is equal to the value of the n-order matrix in the average random consistency table, λ _max is the largest characteristic root of the judgment matrix, and the element of the i-th row and the j-th column of the judgment matrix is the matrix of the i-th selection factor relative to the j-th selection factor The value of standard degree, i=1, 2,..., n, j=1, 2,..., n.