CN114125711B

CN114125711B - Indoor wireless equipment deployment method and system based on local search algorithm

Info

Publication number: CN114125711B
Application number: CN202111385992.0A
Authority: CN
Inventors: 蔡少伟; 雷震东; 何兵
Original assignee: Institute of Software of CAS
Current assignee: Institute of Software of CAS
Priority date: 2021-11-22
Filing date: 2021-11-22
Publication date: 2022-10-14
Anticipated expiration: 2041-11-22
Also published as: CN114125711A

Abstract

The invention discloses an indoor wireless equipment deployment method and system based on a local search algorithm, relates to a wireless equipment deployment method, can quickly generate an equipment deployment scheme aiming at a specified target area on the premise of considering the influence of an obstacle on a signal, can flexibly set the signal quantity requirement received at each position and other constraint conditions, is suitable for solving a small area and is also suitable for deploying a large-scale area.

Description

Indoor wireless equipment deployment method and system based on local search algorithm

Technical Field

The invention relates to a wireless device deployment method, in particular to an indoor wireless device deployment method based on a local search algorithm.

Background

With the rapid development of the internet of things and the mobile internet technology, the indoor positioning service has become a focus of attention in the current location service research field and industry. Indoor wireless equipment such as bluetooth equipment can provide the function of indoor location, however because receive in the indoor location environment influence such as many factors such as non-line of sight, multipath effect, the indoor space object such as barrier, wall body, furniture has great interference to the propagation of bluetooth signal. Therefore, how indoor wireless devices are deployed plays a crucial role in the effectiveness and accuracy of indoor positioning.

The existing wireless device deployment methods can be mainly divided into two types, and the first type is a geometric deployment method, such as a regular quadrilateral deployment method and a regular hexagon deployment method. The regular quadrilateral deployment method divides the whole area to be covered by an inscribed square of a circle, and the position of the center of each circle is the deployment position of the selected equipment. Similarly, the regular hexagon deployment divides the entire area by a regular hexagon, the center of which is the device deployment location. The second deployment method is to model the equipment deployment problem as a linear programming problem and then solve the problem by using an existing linear programming solver. For example, chinese patent application CN109963254A discloses a bluetooth positioning beacon deployment method based on an indoor map, the scheme of which is mainly to divide an indoor positioning area grid according to the indoor map, divide a positioning area according to an indoor space structure data type, and extract key nodes of the indoor positioning area; randomly generating a beacon deployment scheme according to a certain deployment rule according to the key nodes; and performing global optimization of Bluetooth beacon deployment by adopting a genetic algorithm to obtain an optimal deployment scheme.

The existing wireless device deployment method mainly has the following limitations:

1) Most of the current deployment methods are directed to open areas, that is, assuming that there is no obstacle in the target area, the signal of the device can propagate unimpeded, so that the signal coverage of the device is circular at each position. However, the actual indoor environment is usually complex, and there are obstacles such as doors and windows, walls, etc., and the signal passes through these obstacles and then has a large attenuation, but the problem is not considered in the existing deployment mode, which may cause a large difference between the theoretical solution result and the actual effect.

2) Most of the current deployment methods can only ensure that each position is covered by at least one equipment signal, but in order to improve the precision, each position is expected to receive at least a plurality of signal information. At present, a geometric deployment method is researched and designed, so that signals of 4 devices can be received at each position, but the solution is complex, and the requirement of the minimum signal quantity cannot be set at will.

3) In the first type of geometry deployment, once the model is determined, it can only be used for one actual situation. When there are other device deployment requirements in other situations, their invention does not work well if the deployment requirements (e.g., the number of coverage per location) change.

4) The second type is modeling into a linear programming problem and then calling a linear programming solver to do so. The linear programming solver is expensive, the code is closed source, and personalized customization cannot be realized.

Disclosure of Invention

Aiming at the defects of the prior art, the invention provides an indoor wireless equipment deployment method based on a local search algorithm, which can quickly generate an equipment deployment scheme aiming at a specified target area on the premise of considering the influence of an obstacle on a signal, can flexibly set the signal quantity requirement received at each position and other constraint conditions (such as the distance between any two pieces of equipment cannot be too close), and is suitable for solving in a small area and deploying in a large-scale area.

The invention adopts the following technical scheme:

an indoor wireless equipment deployment method based on a local search algorithm comprises the following steps:

1) Dividing an indoor target area according to grids, taking each intersection point of grid lines as a target point covered by at least k wireless equipment signals, and determining coordinates of the target point;

2) Extracting three-dimensional space structure information of a target area and coordinate information of an obstacle, and processing to obtain a structure information graph of the target area;

3) Constructing a coverage set aiming at each target point according to a structural information graph of a target area, wherein elements in the coverage set are candidate deployment points, and the candidate deployment points are all other target points of the target points corresponding to the coverage set covered by wireless equipment signals;

4) Selecting any two candidate deployment points from the coverage set, wherein the distance between the two candidate deployment points is required to be greater than a preset value;

5) And taking the steps 3) and 4) as two constraint conditions, selecting the minimum candidate deployment point from the coverage set of all target points through an optimization solution algorithm, and taking the minimum candidate deployment point as a deployment point for deploying the wireless equipment.

Further, the step of determining candidate deployment points is:

calculating the distance L between any two target points according to the coordinates of the target points;

according to the radius D of a signal coverage area of the wireless equipment, the attenuation distance D after each obstacle is encountered in signal propagation and the number of obstacles between two points, if L is less than or equal to D-m x D, the two target points are judged as candidate deployment points of each other, and the candidate deployment points are added into a coverage set.

Further, whether an obstacle exists between the two target points is judged according to whether a connecting line segment between the two target points in the structural information graph of the target area is intersected with the obstacle line segment.

Further, whether an obstacle exists between the two target points is judged through a rapid filtering method or a straddle experiment method; wherein the content of the first and second substances,

the rapid filtration method comprises the following steps: if the two line segments of the connecting line segment between the two target points and the barrier line segment are not intersected in the x-axis direction or in the y-axis direction, the two line segments are not intersected, otherwise, the two line segments are intersected;

a straddle experimental method: if two line segments, namely a connecting line segment between two target points and an obstacle line segment, span, namely two end points of one line segment are positioned at two ends of the other line segment, the two line segments are intersected; whether the straddling is judged by vector multiplication, and the method comprises the following steps: for the two line segments AB and CD, if the following two vector equations (CA × CD) (CB × CD) < =0 and (AC × AB) ((AD × AB) < = 0), where × represents a vector product and × represents a digital product, it is determined that the two line segments intersect.

Further, the optimization solution algorithm comprises a dynamic weighting mechanism and a heuristic function, and the heuristic function is preferably a local search algorithm.

Further, the optimization solution algorithm is as follows: firstly, taking candidate deployment points in all coverage sets as variables, and randomly or greedy endowing each variable with a weight value to construct an initial solution; then the algorithm enters a loop, the current solution is iteratively improved in the loop until a termination condition is reached, and the found optimal solution is returned; in the cyclic process of the algorithm, if a better solution is found, the current optimal solution is updated, otherwise, the algorithm selects a variable to turn over and endows a weight value; if the algorithm enters local optimization and the variable cannot be selected, updating the weight according to a dynamic weighting mechanism weighting-PMS, randomly selecting a variable and turning the weight value given by the variable.

Further, the method is realized by engineering through the following four steps:

1) Problem modeling step: modeling an indoor wireless device deployment problem as an extended maximum satisfiability problem, comprising the specific steps of:

1.1 Divide the target area into a grid of d, each intersection of grid lines as a target point covered by k wireless devices, a set of target pointsIs collectively denoted as { x ₁ ,x ₂ ,…,x _n In which x _i =1 representing the deployment of a wireless device on the candidate deployment point, x _i =0 represents no device deployed at the candidate deployment point;

1.2 Modeling by defining hard clauses, let collections

For each coverage target point x _i For each target point x _i The following hard clauses were constructed:

in the formula, V represents extraction; the meaning of this formula is that at least k target points are true;

1.3 Define two candidate deployment points x by constructing the following hard clauses _i And x _j The minimum distance between the wireless devices is not less than d, the wireless devices are deployed at the same time when d is hour, and the wireless devices are deployed at the same time when d is hour:

in the formula (I), the compound is shown in the specification,

represents the operator NOT;

1.4 For each candidate deployment point x) _i Constructing a soft clause:

λ(s _i )＝1，

in the form of soft clause s _i Is shown in position

Without deployment of wireless devices, λ(s) _i ) As soft clauses s _i Weight of (a), the equationIndicating if the soft clause is not satisfied, i.e. at s _i Calculating a penalty value according to the weight if the wireless device is deployed;

establishing constraint conditions through the hard clauses and the soft clauses constructed in the steps 1.2), 1.3) and 1.4);

2) A data preparation step:

2.1 Three-dimensional space structure information of the target area and coordinate information of the obstacle are drawn in a CAD file in advance, and the information is analyzed from the CAD file when in use;

2.2 Modeling according to the information analyzed by the CAD file and the constraint conditions established in the step 1), and determining the coordinates of the target points and the coverage set of each target point;

2.3 Output constraint data in a specified format, each line corresponding to a clause representing the constraint, each line specifying the output format as: the weight of the clauses, at least how many characters in the clauses need to be true, and which characters are related in the clauses;

3) Algorithm design and implementation steps: firstly, calculating weights and penalty values of clauses, and scoring variables, wherein the score is the sum of all self-owned penalty values reduced by turning over the assignment of the variables; secondly, guiding a local search algorithm to search according to the scores, searching out a solution which meets the hard clause and has the minimum punishment value, and obtaining a deployment scheme which uses the minimum number of wireless devices under the condition that each point is covered at least k times; finally, when the local search algorithm is trapped in local optimum, updating the weight of the clause according to a dynamic weighting mechanism weighting-PMS;

4) A tool step: and making a visualization tool, inputting a CAD file and constraint condition parameters of the target area, and generating a visualized wireless device deployment scheme aiming at the target area.

Further, the dynamic weighting mechanism weighting-PMS is as follows:

the weight for all unsatisfied hard clauses is increased by 1;

the weight for all unsatisfied soft clauses with weights less than 1000 is increased by 1.

An indoor wireless device deployment system based on a local search algorithm comprises a memory on which a computer program is stored and a processor which, when executing the program, implements the steps of the above method.

A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the steps of the above-mentioned method.

Drawings

FIG. 1 is a flow chart of an indoor wireless device deployment method based on a local search algorithm of the present invention

Fig. 2 is a flow chart of an engineering implementation of an indoor wireless device deployment method based on a local search algorithm according to the present invention.

FIG. 3 is a schematic diagram of the present invention for determining whether two straight lines straddle by vector multiplication.

Detailed Description

In order to make the aforementioned and other features and advantages of the invention more comprehensible, embodiments accompanied with figures are described in detail below.

The invention discloses an indoor wireless equipment deployment method based on a local search algorithm, the flow of which is shown in figure 1, and the method specifically comprises the following steps:

2) Extracting three-dimensional space structure information of the target area and coordinate information of the obstacle, and processing to obtain a structure information graph of the target area;

In a specific embodiment, the overall process of the engineering implementation of the method of the present invention includes four steps of problem modeling, data preparation, algorithm design and implementation, and toolization, the core work is mainly in the data preparation and algorithm design part, as shown in fig. 2, and each step is detailed as follows:

1. problem modeling: the indoor wireless device deployment problem is modeled as an extended maximum satisfiability problem.

Hard clauses and soft clauses are defined in the extended maximum satisfiability problem: hard clause c: = ({ l) ₁ ∨…∨l _m K) for the set l ₁ ∨…∨l _m At least k of m words in the Chinese character are true. Soft clause s: = ({ l) ₁ ∨…∨l _m -if there are less than k true words in k), a penalty value is introduced: penalty [ s ]]＝λ(s)*(k-sum[s]). Wherein sum [ s ]]＝l ₁ +l ₂ +…+l _m And λ(s) is the weight of the soft clause. The hard clauses and the soft clauses are expressed by using a main disjunctive normal form, wherein l represents characters, and V represents disjunctions. Given an extended maximum satisfiable problem instance, which contains a series of hard and soft clauses, the goal is to find a variable assignment such that all hard clauses are satisfied (i.e., at least k are true) and the sum of all unsatisfied soft clause penalty values is minimal.

The specific modeling method for modeling the minimum equipment deployment problem into the expanded maximum satisfiability problem is as follows:

(1) Dividing the target area by a grid with the side length of d, wherein each intersection point of grid lines is used as a target point to be covered by k wireless devices. While each target point is also a candidate deployment point for deploying the wireless device. For convenience, the points are labeled as x sequentially from left to right, top to bottom, in order ₁ ,x ₂ ,…,x _n 。x _i =1 represents deployment of wireless device on this candidate deployment point, x _i =0 represents that no device is deployed at the candidate deployment point.

(2) Hard clause modeling 1: due to the fact thatEach target point x _i At least needs to be covered by k wireless devices. Order set

For all the coverable target points x _i The set of candidate deployment points of (2). For each target point x _i The following hard clause is constructed:

the above-mentioned formula is shown in the specification,

the representation may cover point x _i Requiring each point to be covered by at least k wireless devices, with hard clause c _i Meaning that at least k words are true (value 1), i.e.

At least k of which are 1.

(3) Hard clause modeling 2: in practice, if two devices are deployed at too small a distance, mutual signal interference may occur. It is therefore desirable to limit the minimum distance d between deployed devices. Thus for any two candidate deployment points x _i And x _j If the distance between them is less than d, then the two candidate deployment points may not deploy the device at the same time. The following hard clause is constructed:

the above-mentioned formula is shown in the specification,

the operator "not" is represented in a disjunctive normal form; this hard clause represents the deployment point x for any two candidates _i And x _j If the distance between them is less than d, it is required that the two points cannot deploy the device at the same time. From the truth table it can be seen that only x _i And x _j When the two are not true simultaneously, the hard clauses can be satisfied.

(4) Modeling soft clauses: the optimization goal of the method is to want to meet the above hard constraint requirement with as little equipment as possible. Thus for each deployment point x _i Constructing a soft clause:

λ(s _i )＝1

the above expression, soft clause s _i Is shown in position

Without deployment of wireless devices, λ(s) _i ) As soft clauses s _i If the soft clause is not satisfied, i.e. at s _i With the device deployed, a penalty value is calculated based on this weight.

According to the mode, the corresponding data are modeled into an expanded maximum satisfiability problem, and a series of hard clauses and soft clauses are obtained, namely an expanded maximum satisfiability problem example.

2. Preparing data: and reading the CAD file of the designated area, extracting the structural information of the target area from the CAD file, and generating constraint condition data. The method mainly comprises two tasks: CAD file parsing and constraint data modeling.

1) CAD file parsing

The dxfgrabber library in python is used to analyze the CAD drawing stored in the DXF format, and the structure information of the target area, the coordinate information of the obstacle, and the like are extracted from the drawing layer of the partition such as the wall, the door, the window, and the like. And processing the extracted original information to generate a structural information graph of the target area. DXF is a CAD data file format developed by Autodesk for CAD data exchange between AutoCAD and other software, and its interior is ASCII code, so that different types of computers can exchange DXF files to achieve the purpose of exchanging graphics. The Dxfgrabber is a library provided by python and used for reading the DXF file, and layers, entities and the like in the DXF file can be conveniently read.

2) Constraint data modeling

And performing data modeling on the constraint conditions on the basis of the CAD analysis result to generate a constraint condition data file as the input of a subsequent algorithm.

The main objective of the invention is to flexibly set the signal quantity requirement to be received by each position and to rapidly produce the equipment deployment scheme. Among the most critical constraints are: the main steps for data modeling of this constraint are as follows:

(1) Determining coordinates x of a target point _i : the problem modeling module divides the target area into grids and generates a target point set, and only on the basis, a coordinate system is established to determine the coordinates of each target point.

(2) Determining each target point x _i Is covered with _i : first assume that the effective coverage area of a device is a circle centered at the device and having a radius of D. And assume that the signal travels a distance d after encountering an obstacle (e.g., a wall) in that direction. For any two target points x _i And x _j The actual distance between them can be calculated from their actual coordinates as L, assuming that the number of obstacles between them is m. If L is less than or equal to D-m x D, x is added _i Joining to the overlay set S _j In (1), x _j Joining to the overlay set S _i In (x) _i And x _j Each being a candidate deployment point). And traversing all the point pairs to obtain a coverage set corresponding to each target point.

The parameters D and D need to be actually measured according to the equipment and the actual obstacle condition. Whether an obstacle exists between any two target points is judged by judging whether a connecting line segment of the two points is intersected with an obstacle line segment, wherein the position information of the obstacle line segment is obtained by a CAD file analysis module. The specific method comprises the following steps:

a rapid filtration method: two line segments do not intersect if they do not intersect in the x-axis direction or in the y-axis direction. Taking the x-axis as an example, if the minimum x-coordinate of one of the segments is greater than the maximum x-coordinate of the other segment, the two segments do not intersect in the x-axis direction.

A transnational experimental method: if two line segments intersect, then a stride is necessary. Straddling means that two end points of one line segment are at two ends of the other line segment. Whether the crossover is caused or not can be judged by the vector product, and if (CA × CD) (CB × CD) < =0 (× represents the vector product, and × represents the digital product), it indicates that the directions of the vector CA and the vector CB with respect to the vector CD are different, and the points a and B are located on both sides of the CD, respectively, as shown in fig. 3. Similarly, it can be determined whether C and D are located on both sides of AB.

Therefore, if (CA × CD) (CB × CD) < =0 and (AC × AB) ((AD × AB) < = 0), the two line segments intersect.

(3) Outputting constraint condition data: and outputting constraint condition data according to a specified format, wherein each line corresponds to a clause representing the constraint condition. Each row specifies an output format as: the weight of the clause, at least how many characters in the clause need to be true, and which characters are involved in the clause. The data are separated by spaces.

3. Algorithm design and implementation:

at present, algorithms for solving the set coverage problem are more, but the algorithms cannot adapt to various requirements in practical situations well. The invention designs an expanded solving algorithm of the maximum satisfiability problem. The algorithm belongs to a heuristic search algorithm, and comprises a dynamic weighting mechanism and a heuristic function. The local search algorithm framework is as follows:

the local search algorithm is a heuristic function that solves the optimization problem. Firstly, constructing an initial solution randomly or greedily, then entering a loop by the algorithm, iteratively improving the current solution in the loop until a termination condition is reached, and finally returning to the best solution found in the process. In the course of iterative improvement, the algorithm will perform a local operation once at a time to change the current solution.

The input to the algorithm is an example of the maximum satisfiability problem of the encoded extension described above. Wherein the variable is x ₁ ,x ₂ ,…,x _n . In the initialization stage, each variable is randomly assigned with 0 and 1, wherein 0 represents no deployment and 1 represents deployment, so that an initial solution is constructed. The algorithm then enters a loop to iteratively refine the current solution. In the algorithm loop process, whenever a better solution is found, the optimal solution is updated. Otherwise, the algorithm selects a variable to flip its assignment according to a heuristic function, i.e., from 0 to 1, or from 1 to 0. If the algorithm enters local optimization, namely when the variables cannot be selected according to the heuristic function, the algorithm updates the weights according to a dynamic weighting mechanism weighting-PMS, then the algorithm randomly selects one variable and then turns the value of the variable. The specific steps are described in detail below:

(1) Weight of clause: for any clause c (which may be a hard clause or a soft clause), an integer w [ c ] is given as its weight. In the initial stage, the weights of all clauses are initialized to 1.

(2) Penalty of clauses: for any clause c (which may be a hard clause and a soft clause), if it is unsatisfied, it introduces a penalty value:

penalty[c]＝w(c)*(k-sum[c])。

where sum [ c ] corresponds to the number of all genuine words in the clause, and m represents the number of clauses. If the clause is satisfied, then penalty [ c ] =0.

(3) Score of the variables: for any variable x, its score [ x ] is the sum of the penalty values of all clauses reduced by flipping the variable's valuation. The scores are used for guiding a heuristic search algorithm to search, and when the hard clauses are met, solutions with punishment values as small as possible are searched. The problem is to use as few deployment scenarios as possible in the case that each point is covered at least k times.

(4) Dynamic weighting mechanism weighting-PMS: each time the algorithm falls into local optimum, i.e. no variable scores greater than 0. The algorithm will update the weight of the clause according to the following rule:

a) The weight for all unsatisfied hard clauses is increased by 1.

b) The weight for all unsatisfied soft clauses with weights less than 1000 is increased by 1.

The algorithm will cycle through until the cutoff time T is reached. Where T is a parameter considered to be given according to actual needs. The algorithm finally outputs a variable x ₁ ,x ₂ ,…,x _n Is assigned to the global value. Wherein x _i A point of =1, a device needs to be deployed at its corresponding coordinates. So there are as many devices as needed for a variable with a value of 1.

4. A tool module:

the module generates a visual tool which can be directly used by using an open-source Web framework Django, and inputs a CAD file of a target area and related constraint parameters, so that a device deployment scheme aiming at the area can be rapidly and visually generated. Django is an open-source Web application framework written by python, which is a framework that follows the MVC (model, view, controller) design pattern. By using the framework, the high-quality, easy-to-maintain and database-driven application program can be conveniently and quickly created. The tool is divided into a document uploading page and a result output page. Uploading the CAD file in the DXF format of the target area in a file uploading page, setting related parameters, and visually displaying a deployment result in a result page after clicking confirmation.

Although the present invention has been described with reference to the above embodiments, it should be understood that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.

Claims

1. An indoor wireless equipment deployment method based on a local search algorithm is characterized by comprising the following steps:

3) Constructing a coverage set aiming at each target point according to a structural information graph of a target area, wherein elements in the coverage set are candidate deployment points, and the candidate deployment points are all other target points of the target points corresponding to the coverage set covered by wireless equipment signals; the method for determining the candidate deployment point comprises the following steps: calculating the distance L between any two target points according to the coordinates of the target points; according to the radius D of a signal coverage area of the wireless equipment, the attenuation distance D after each obstacle is encountered in signal propagation and the number of obstacles between two points, if L is less than or equal to D-m x D, the two target points are judged as candidate deployment points of each other and are added into a coverage set;

5) Taking the steps 3) and 4) as two constraint conditions, selecting the minimum candidate deployment point from the coverage set of all target points through an optimization solution algorithm, and taking the minimum candidate deployment point as a deployment point for wireless equipment deployment; the optimization solving algorithm comprises the following steps: firstly, taking candidate deployment points in all coverage sets as variables, giving a weight value to each variable randomly or greedily, and constructing an initial solution; then the algorithm enters a loop, the current solution is iteratively improved in the loop until a termination condition is reached, and the found optimal solution is returned; in the cyclic process of the algorithm, if a better solution is found, the current optimal solution is updated, otherwise, the algorithm selects a variable to turn over and endows a weight value; and if the algorithm enters local optimization and the variable cannot be selected, updating the weight according to a dynamic weighting mechanism weighting-PMS, randomly selecting a variable and turning the weight value given by the variable.

2. The method according to claim 1, wherein whether the obstacle exists between the two target points is determined according to whether a link segment between the two target points in the structural information map of the target area intersects with the obstacle segment.

3. The method of claim 2, wherein whether an obstacle exists between two target points is determined by a rapid filtering method or a transnational experimental method; wherein the content of the first and second substances,

4. The method of claim 1, wherein the optimization solution algorithm comprises a dynamic weighting mechanism and a heuristic function, the heuristic function preferably being a local search algorithm.

5. The method of claim 1, wherein the method is engineered by four steps comprising:

1) Problem modeling step: modeling an indoor wireless device deployment problem into an extended maximum satisfiability problem, comprising the specific steps of:

1.1 Divide the target area into a grid of d, with each intersection of grid lines as a target point covered by k wireless devices, and the set of target points is denoted as { x } ₁ ，x ₂ ，…，x _n In which x is _i =1 representing the deployment of a wireless device on the candidate deployment point, x _i =0 for not deploying a device on the candidate deployment point;

1.2 Modeling by defining hard clauses, let set

wherein V represents an extraction; the meaning of this formula is that at least k target points are true;

1.3 Define two candidate deployment points x by constructing the following hard clauses _i And x _j When the minimum distance between the two is not less than d, the wireless devices are deployed at the same time, and when the d is not smaller than d, the wireless devices are deployed at the same time:

in the formula (I), the compound is shown in the specification,

representing an operator NOT;

1.4 For each candidate deployment point x) _i Constructing a soft clause:

λ(s _i )＝1，

in the form of soft clauses _i Is shown in position

There is no wireless device deployed and,λ(s _i ) As soft clauses s _i If the soft clause is not satisfied, i.e. at s _i Calculating a penalty value according to the weight if the wireless device is deployed;

2) A data preparation step:

4) A tool step: and making a visualization tool, inputting the CAD file and the constraint condition parameters of the target area, and generating a visualized wireless device deployment scheme aiming at the target area.

6. The method of claim 1 or 5, wherein the dynamic weighting mechanism weighting-PMS is:

the weight for all unsatisfied hard clauses is increased by 1;

7. An indoor wireless device deployment system based on a local search algorithm, comprising a memory having stored thereon a computer program and a processor which, when executing the program, carries out the steps of the method according to any of claims 1 to 6.

8. A computer-readable storage medium, characterized in that a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 6.