Scheduling method for directional mobile chargers in WRSNs

Abstract

The invention relates to a method for scheduling directional mobile chargers in WRSNs. The method comprises the following steps: acquiring a node to be charged, the directional mobile chargers and basic charging information; using a KCPG algorithm to cluster and determine a charging position set, and using a cMFRDS algorithm to select an optimal charging direction set; constructing an asymmetric energy consumption adjacency matrix and calculating a moving time length list through an improved LKH algorithm; constructing a linear programming problem, and calculating the charging time in each direction by using a CPlex solver; and constructing a complete scheduling list of the directional mobile chargers. According to the invention, it is ensured that the directional mobile charger can efficiently provide energy supply for each node to be charged in the WRSNs, the energy loss is reduced through a reasonable scheduling strategy, and the charging path is optimized.

Description

Technical Field

[0001] The present invention relates to the application field of wireless charging technology, and in particular to a scheduling method for a directional mobile charger in WRSNs. Background Art

[0002] Wireless Sensor Networks (WSNs) have been widely used in various industrial and daily scenarios. However, the main factor hindering the rapid development of WSNs is the limited energy storage capacity of nodes. In order to solve this problem, researchers are not only committed to improving battery storage capacity, but also to promoting wireless power transfer (WPT) technology. With the continuous development of WPT, researchers apply it to WSNs to replenish the energy of nodes, thus introducing the concept of WRSNs (Wireless Rechargeable Sensor Networks).

[0003] In WRSNs, MCs can be divided into omnidirectional MCs (OMCs) and directional mobile chargers (DMCs) according to whether the transmission device is omnidirectional (transmitting energy uniformly in all directions centered on the charger) or directional (concentrating wireless energy in a certain direction for transmission). Most of the existing WRSN charging scheduling strategies are based on omnidirectional charging, which has low charging efficiency. In reality, due to factors such as terrain and topographic restrictions, altitude differences, wind (if deployed in the air) and water flow (if deployed on or in the water), the path symmetry assumption often does not hold in many real-world scenarios. Path asymmetry makes the design for path symmetric environments invalid in actual path asymmetric scenarios, resulting in directional mobile chargers consuming different energies in different directions along the same path during movement, thereby affecting the charging efficiency. Summary of the invention

[0004] In order to solve the above problems, the present invention proposes a scheduling method for directional mobile chargers in WRSNs.

[0005] The specific plan is as follows:

[0006] A method for scheduling directional mobile chargers in WRSNs comprises the following steps:

[0007] S101, obtaining a node to be charged, a directional mobile charger and basic charging information;

[0008] S102, using a K-means-based charging group generation KCPG algorithm to cluster the acquired nodes to be charged, and determine a charging location set X for each cluster;

[0009] S103, based on the basic charging information and the charging position set X, use the minimum functional representative direction set cMFRDS algorithm to select the optimal charging direction set S equivalent to the omnidirectional infinite direction set dir ;

[0010] S104, construct asymmetric energy consumption adjacency matrices S and W, and use the improved LKH algorithm to obtain a set of mobile plan items

[0011] S105, construct a linear programming problem and use the CPlex solver to calculate the charging time length list

[0012] S106, based on the optimal charging direction set S dir and charging time length list Building a collection of charging scheduling projects A collection of charging scheduling items and a collection of mobile plan items Combined, the complete dispatch list S of directional mobile chargers is calculated MC .

[0013] Furthermore, in S101, the basic charging information includes:

[0014] The set of nodes to be charged U = {u1,u2,…,u n Energy expectation of Initial Energy List Storage capacity list The initial position x0 of the base station and the transmission sector angle of the directional mobile charger Transmission power p0 and moving speed v; where n represents the number of nodes to be charged, and u represents the number of nodes to be charged. i ∈U.

[0015] Furthermore, in S102, a charging position set X is determined using a K-means-based charging group generation KCPG algorithm, specifically including:

[0016] Initialization step: set the charging distance to the mobile charger to D and the number of clusters f to 1;

[0017] Clustering step: Use K-means clustering algorithm to cluster the nodes to be charged into f clusters and obtain the center point of each cluster;

[0018] Determine the charging location set step: determine whether the radius of the smallest circle corresponding to each cluster is smaller than the charging distance D of the mobile charger; if so, return the center points of these clusters and add them to the charging location set X = {x1, x2, ..., x w}, where w represents the number of charging positions, and charging position x r ∈X; if not, increase the number of clusters f by 1 and re-execute the clustering part;

[0019] Steps for optimizing the charging position set: Calculate the minimum covering circle of each cluster in the charging position set X, obtain the center point and diameter of the minimum covering circle, and update the charging position set X of each cluster.

[0020] Further, in S103, the optimal charging direction set is selected using the cMFRDS algorithm for creating a minimum functional representative direction set, which specifically includes:

[0021] Determine the charging direction step S1031: set the charging direction to the mobile charger to be any value in the continuous range [0, 2π), and set the charging distance and charging sector angle to the mobile charger to be constants;

[0022] Step S1032 of constructing a set of nodes to be charged: constructing a set of nodes to be charged N (x r ,D)={u i |u i ∈U(x r ,D),x r ∈X},U(x r ,D) indicates that the charging position is x r , the set of nodes to be charged whose charging distance is not greater than D;

[0023] Generate charging direction list step S1033: Build list L θ,a , the node set to be charged N(x r ,D) each node u to be charged i , starting from the center of the circle, with the horizontal right direction as the reference direction, counterclockwise at an angle θ i Sort in ascending order and add to list L θ,a ; Create a list Where L ψ Indicates the charging direction list, |N(x r ,D)| represents the set N(x r ,D) the number of elements in;

[0024] The step S1034 of optimizing the charging direction set is as follows:

[0025] S10341: Construct charging direction ψ i The set of all nodes to be charged N(x r ,D,ψ i ), where x r represents the charging position, ψ irepresents the charging direction, m is the variable of the control loop, and the optimal charging direction set S equivalent to the omnidirectional infinite direction set is dir =[(x r ,ψ1)], indicating that the charging position is x r When the charging direction is ψ1, m=1;

[0026] S10342: From i=2 to |N(x r ,D)|Loop: Determine N(x r ,D,ψ i ) is not included in N(x r ,D,ψ m ), if N(x,D,ψ i ) is not included in N(x r ,D,ψ m ), then execute S10343; otherwise, execute S10344; if the loop ends, enter S10345;

[0027] S10343: Get N(x r ,D,ψ i )'s maximum charging direction angle θ max and the minimum charging direction angle θ min , calculate ψ=(θ max +θ min ) / 2, update m=m+1, update S dir =[(x r ,ψ1,ψ)], where ψ represents the optimized charging direction, let i=i+1, and enter S10342;

[0028] S10344: If N(x,D,ψ i ) is included in N(x,D,ψ m ), then no update is performed, i=i+1 is set, and the process goes to S10342;

[0029] S10345: After the loop ends, judge: If N(x r ,D,ψ1) is included in N(x r ,D,ψ i ), then from S dir Remove the first element and finally return the optimal charging direction set S dir ,S dir ={(x r ,ψ p )|r∈{1,2,…,w} represents a charging position x r And all corresponding charging directions ψ p , where ψ1 is the list L ψ The first element of .

[0030] Furthermore, in S104, constructing asymmetric energy consumption adjacency matrices S and W specifically includes:

[0031] Asymmetric distance model steps: quantify and simulate the asymmetry of the path length between any two points in a wireless sensor network;

[0032] The steps of the asymmetric unit distance motion energy consumption model are as follows: quantify the energy consumption of a directional mobile charger when it moves from one location to another;

[0033] Determine the asymmetric energy consumption adjacency matrix step: calculate the energy consumption and movement time of the directional mobile charger on different paths, and optimize the path selection and scheduling.

[0034] Further, in S104, the improved LKH algorithm for solving the ASTP problem is used to obtain a set of mobile plan items. Specifically include:

[0035] S1041: Set the set of all locations passed when dispatching to the mobile charger r Π =[x0,x1,…,x r ,x0], which consists of r+2 points, where x0 represents the initial position of the base station, x1,…,x r Indicates a certain charging position. First determine an initial solution, that is, randomly generate a traversal of all r Π The path of the internal node and calculate the path;

[0036] S1042: repeatedly applying k-pot sub-circuit exchange to adjust the circuit to reduce the total circuit length, exchanging k edges, replacing k path segments in the circuit with k new path segments, and searching for the best k-pot sub-circuit exchange in each iteration to most effectively reduce the total circuit length, wherein k represents the number of path segments used to replace the circuit in each iteration;

[0037] S1043: Dynamically adjust the value of k through variable depth search so that it can explore a larger solution space and escape the local optimum, and dynamically adjust the type of k-pot sub-loop exchange for instance-specific adaptation;

[0038] S1044: After multiple iterations of optimization, when the solution is no longer improved, the final minimum energy traversal sequence U is obtained Result , based on U Result , combined with the moving speed v of the directional mobile charger, the moving time length of the directional mobile charger on each path is obtained This determines all mobile scheduling Wherein, d(x0,x1) represents the asymmetric distance of the path from charging position x0 to charging position x1.

[0039] Furthermore, in S105, constructing a linear programming problem specifically includes:

[0040] Set the objective function to min t 1 1×K t, the constraint condition is C2:t≥0, Where t=[t1,t2,…,t K ] T represents the energy transfer time vector, K represents the number of all corresponding charging directions related to the charging position; constraint C1 ensures that the final energy of all nodes to be charged meets the demand; constraint C2 stipulates that the time is non-negative; constraint C3 represents the energy loss equation caused by energy transfer during the charging process, represents the total energy loss for completing the energy transfer time list t, represents the total energy loss of the directional mobile charger to meet a charging plan s, e B0 represents the initial energy of the directional mobile charger, e F0 It represents the final energy after the directional mobile charger completes a charging schedule. Where K represents all S dir Number of pairs; represents the total energy received by all nodes, where e F =min{e B +e R ,e C}, C is the energy transfer coefficient matrix, C = [c(r,p)] represents S dir Medium charging position x r Along the charging direction ψ p The energy transfer coefficient is but otherwise Where β represents the decay exponent, which controls the decay rate of the distance on the transfer coefficient; α represents a distance correction parameter used to affect the influence of distance d on the transfer coefficient; δ represents a calibration coefficient used to adjust the basic transfer level of the model; d represents the distance between the charging position and the node; the charging duration of each charging direction is set as a variable, and the problem of determining the shortest charging duration list is modeled as a linear programming problem.

[0041] Further, in S105, the CPlex solver is used to calculate a list of charging time lengths in each direction. Specifically include:

[0042] Find the time vector that minimizes the total energy transfer time and build a model suitable for CPlex solution, including the objective function and constraints. By inputting the relevant data of the problem into the CPlex model, use the solve function to solve the built model to obtain a list of charging time lengths for directional mobile chargers in each direction.

[0043] The present invention adopts the above technical solution and has the following beneficial effects:

[0044] (1) The present invention selects the optimal charging direction set through the cMFRDS algorithm (Create Minimum Functional Representative Direction Set algorithm), ensuring that the wireless charging vehicle can concentrate energy to charge in the most effective direction, thereby significantly improving the charging efficiency and reducing energy loss;

[0045] (2) The present invention uses the KCPG algorithm (K-means charging position Generation, a charging group generation algorithm based on K-means) to cluster the nodes, determine the charging position set of each cluster, and then use the improved LKH algorithm (Lin-Kernighan heuristic algorithm) to fully consider the asymmetric factors (such as terrain and wind force) in the actual environment to find the mobile path with the lowest energy consumption, thereby reducing the mobile energy consumption of the wireless charging vehicle;

[0046] (3) The present invention determines the optimal charging time in each direction by constructing a linear programming model and using the CPlex solver, thereby forming a complete scheduling list to ensure that the wireless charging vehicle can perform the charging task according to the most optimized solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a flow chart of a method for scheduling directional mobile chargers in WRSNs according to an embodiment of the present invention;

[0048] Figure 2 A schematic diagram of a wireless rechargeable sensor network according to an embodiment of the present invention;

[0049] Figure 3 A schematic diagram of a fan-shaped charging area of ​​a mobile charger according to an embodiment of the present invention;

[0050] Figure 4 This is a schematic diagram of charging of a directional mobile charger in an asymmetric environment according to an embodiment of the present invention. DETAILED DESCRIPTION

[0051] The present invention is further described in detail below in conjunction with embodiments and drawings, but the embodiments of the present invention are not limited thereto.

[0052] like Figure 1 As shown, the present invention provides a scheduling method for directional mobile chargers in WRSNs, comprising the following steps:

[0053] S101, obtaining nodes to be charged, directional mobile chargers and basic charging information.

[0054] Specifically, the basic charging information includes:

[0055] The set of nodes to be charged U = {u1,u2,…,u n Energy expectation of Initial Energy List Storage capacity list Where n represents the number of nodes to be charged, and u represents the number of nodes to be charged. i ∈U; the initial position x0 of the base station, the transmission sector angle φ of the directional mobile charger, the transmission power p0 and the moving speed v.

[0056] The node to be charged is equipped with an energy receiving device, and the directional mobile charger is equipped with a wireless energy transmission device.

[0057] like Figure 2 As shown, the directional mobile charger replenishes its energy at the base station and can set out at any time to provide energy for all nodes to be charged. The directional mobile charger is fully charged when it leaves the base station and follows the predetermined charging scheduling tasks, starting from the base station, completing all charging tasks in sequence to meet the energy needs of all nodes, and finally returning to the base station. During the charging process, the directional mobile charger does not change the charging plan of the newly added nodes.

[0058] S102, using a K-means-based charging group generation KCPG algorithm to determine a charging location set X, specifically including:

[0059] Initialization step: set the charging distance to the mobile charger to D and the number of clusters f to 1;

[0060] Clustering step: Use K-means clustering algorithm to cluster the nodes to be charged into f clusters and obtain the center point of each cluster;

[0061] Determine the charging location set step: determine whether the radius of the smallest circle corresponding to each cluster is smaller than the charging distance D of the mobile charger; if so, return the center points of these clusters and add them to the charging location set X = {x1, x2, ..., x w}, where w represents the number of charging positions, and charging position x r ∈X; if not, increase the number of clusters f by 1 and re-execute the clustering part;

[0062] Optimization of the charging position set step: Calculate the minimum covering circle of each cluster in the charging position set X, obtain the center point and diameter of the minimum covering circle, and update the charging position set X of each cluster.

[0063] Specifically, calculating the minimum covering circle of each cluster in the charging position set X includes the following steps:

[0064] Step Ⅰ: In the charging position set X, select an unselected cluster a, and randomize all the nodes to be charged in cluster a;

[0065] Step Ⅱ: Select two unselected nodes to be charged u1 and u2 from cluster a, and take the center of the line connecting the node to be charged u1 and the node to be charged u2 as the center of the circle to obtain the minimum covering circle A that can cover the two nodes to be charged at the same time;

[0066] Step Ⅲ: Select an unselected node to be charged u i (i is a positive integer) from cluster a, and judge whether the node to be charged u i is outside the minimum covering circle A; if so, go to Step IV; if not, the node to be charged u i exists in the minimum covering circle A, and there is no need to update the minimum covering circle A. Repeat Step Ⅲ; if there are no unselected nodes to be charged u i , go to Step VIII;

[0067] Step IV: Take the position of the node to be charged u i as the center of the circle, let the radius be 0, and make the minimum covering circle B, and enter Step V;

[0068] Step V: Traverse the previous u i-1 nodes to be charged and add them to the minimum covering circle B. During the addition process, the node to be charged u i is always on the boundary of the minimum covering circle B. Let the added node to be charged be u j , and judge whether the node to be charged u j (1 ≤ j < i) is outside the minimum covering circle B; if so, go to Step VI; if not, the node to be charged u j exists in the minimum covering circle B, and there is no need to update the minimum covering circle B. Repeat Step V; if the previous u i-1 nodes to be charged are traversed, then let the minimum covering circle A = the minimum covering circle B, and enter Step Ⅲ;

[0069] Step VI: Update the center of the minimum covering circle B to the midpoint of the line connecting the node to be charged u i and the node to be charged u j so that the node to be charged u i and the node to be charged uj At the boundary of the minimum covering circle, enter Step VII;

[0070] Step VII: Traverse the first u j-1 nodes to be charged and add them to the minimum covering circle B. During the addition process, the node to be charged u i and the node to be charged u j are always on the boundary of the minimum covering circle B. Let the added node to be charged be u k , and determine whether the added node to be charged u k (1 ≤ k < j) is outside the minimum covering circle B. If so, find u i , u j , u k 's shared center point and update the minimum covering circle B; if not, the node to be charged k exists within the minimum covering circle B and there is no need to update the minimum covering circle B. Repeat Step VII; if all the first u j-1 nodes to be charged have been traversed, then set the minimum covering circle A = the minimum covering circle B and enter Step V;

[0071] Step VIII: Determine whether there are still clusters that have not been selected; if so, enter Step I; if not, update the charging position set X according to the center points and diameters of the minimum covering circles corresponding to each cluster.

[0072] S103. Based on the basic charging information and the charging position set X, use the cMFRDS algorithm to select the optimal charging direction set S dir .

[0073] As Figure 3 shown, the energy transmission signal of the directional mobile charger covers a fan-shaped area centered on the directional mobile charger. The central axis of this fan represents the charging direction, and the radius of the fan corresponds to the charging distance, including the following key parameters: ψ (charging direction angle), (fan angle), θ (node direction angle), u (node), and D (charging distance of the DMC). The three parameters θ and ψ are measured in degrees within the range of [0°, 360°). Both ψ and θ are measured relative to a common reference direction, and the angles increase counterclockwise. Assume that the fan angle of the directional mobile charger remains unchanged, and ψ is called the corresponding charging direction. Assume that ψ of the directional mobile charger can be freely adjusted.

[0074] Specifically, use the create minimum functional representative direction set cMFRDS algorithm to select the optimal charging direction set, which specifically includes:

[0075] Determine the charging direction step S1031: set the charging direction to the mobile charger to be any value in the continuous range [0, 2π), and set the charging distance and charging sector angle to the mobile charger to be constants;

[0076] Step S1032 of constructing a set of nodes to be charged: constructing a set of nodes to be charged N (x r ,D)={u i |u i ∈U(x r ,D),x r ∈X},U(x r ,D) indicates that the charging position is x r , the set of nodes to be charged whose charging distance is not greater than D;

[0077] Generate charging direction list step S1033: Build list L θ,a , the node set to be charged N(x r ,D) each node u to be charged i , starting from the center of the circle, with the horizontal right direction as the reference direction, counterclockwise at an angle θ i Sort in ascending order and add to list L θ,a ; Create a list Where L ψ Indicates the charging direction list, |N(x r ,D)| represents the set N(x r ,D) the number of elements in;

[0078] The step S1034 of optimizing the charging direction set is as follows:

[0079] S10341: Construct charging direction ψ i The set of all nodes to be charged N(x r ,D,ψ i ), where x r represents the charging position, ψ i represents the charging direction, m is the variable of the control loop, and the optimal charging direction set S equivalent to the omnidirectional infinite direction set is dir =[(x r ,ψ1)], indicating that the charging position is x r When the charging direction is ψ1, m=1;

[0080] S10342: From i=2 to |N(x r ,D)|Loop: Determine N(x r ,D,ψ i ) is not included in N(x r ,D,ψ m ), if N(x,D,ψ i) is not included in N(x r ,D,ψ m ), then execute S10343; otherwise, execute S10344; if the loop ends, enter S10345;

[0081] S10343: Get N(x r ,D,ψ i )'s maximum charging direction angle θ max and the minimum charging direction angle θ min , calculate ψ=(θ max +θ min ) / 2, update m=m+1, update S dir =[(x r ,ψ1,ψ)], where ψ represents the optimized charging direction, let i=i+1, and enter S10342;

[0082] S10344: If N(x,D,ψ i ) is included in N(x,D,ψ m ), then no update is performed, i=i+1 is set, and the process goes to S10342;

[0083] S10345: After the loop ends, judge: If N(x r ,D,ψ1) is included in N(x r ,D,ψ i ), then from S dir Remove the first element and finally return the optimal charging direction set S dir ,S dir ={(x r ,ψ p )|r∈{1,2,…,w} represents a charging position x r And all corresponding charging directions ψ p , where ψ1 is the list L ψ The first element of .

[0084] S104, construct asymmetric energy consumption adjacency matrices S and W, and use the improved LKH algorithm to obtain a set of mobile plan items

[0085] like Figure 4 As shown in the figure, the path is in an asymmetric environment. In order to simulate this asymmetry, an asymmetric energy consumption coefficient matrix is ​​constructed according to the asymmetric environment, and the corresponding path is adjusted accordingly. In this model, the final charging position is determined as A, 7, B, C, and D in the figure, and the DMC (directional mobile charger) is specified to start from the BS (base station). Therefore, the movement process of the DMC includes path planning based on the environment, finding the path with the least energy consumption, visiting each location in turn for charging, and finally returning to the BS.

[0086] Specifically, the construction of asymmetric energy consumption adjacency matrices S and W specifically includes:

[0087] Asymmetric distance model steps: quantify and simulate the asymmetry of the path length between any two points in a wireless sensor network;

[0088] Asymmetric unit distance motion energy consumption model steps: Quantify the energy consumption of a directional mobile charger when it moves from one location to another.

[0089] Determine the asymmetric energy consumption adjacency matrix step: calculate the energy consumption and movement time of the directional mobile charger on different paths, and optimize the path selection and scheduling.

[0090] The part of constructing the asymmetric distance model: defining the routing-asymmetric (RA) distance coefficient Simulate from x i to x j The asymmetric effect of the path segment length is determined by the factors mRA, hRA, and fRA, where mRA represents the combined effect of other miscellaneous factors caused by topography and topographic heterogeneity, and hRA represents the x i With x j The height distance between them, fRA represents the influence of MC from x i To x j The cumulative external force of movement, for any two positions x in the WRSN area i ,x j , let d Ideal (x i ,x j ) is the ideal Euclidean distance between them, let d(x i ,x j ) is the actual asymmetric distance of the path, and the path segment is i to x j The RA distance is modeled as

[0091] The asymmetric unit distance motion energy consumption model part: define the RA motion energy consumption coefficient Change the path segment from x i to x j The RA motion energy consumption is modeled as and Where w0 is the energy consumption of the directional mobile charger moving a unit distance, which can be set as a constant value;

[0092] The part of determining the asymmetric energy consumption adjacency matrix: All the energy consumed during the movement of the directional mobile charger is: Move time length list During the experiment Set to a random value between 0.5 and 1.5. Set to 1, collect the RA distances between all position pairs into the matrix S = [d(x i ,x j )]{x i ,x j ∈X}, the RA motion energy consumption rates between all position pairs are collected into the matrix W = [w(x i ,x j )]{x i ,x j ∈X}.

[0093] Specifically, the improved LKH algorithm for solving the ASTP problem is used to obtain a set of mobile plan items: Specifically include:

[0094] S1041: Set the set of all locations passed when dispatching to the mobile charger r Π =[x0,x1,…,x r ,x0], which consists of r+2 points, where x0 represents the initial position of the base station, x1,…,x r Indicates a certain charging position. First determine an initial solution, that is, randomly generate a traversal of all r Π The path of the internal node and calculate the path;

[0095] S1042: repeatedly applying k-pot sub-circuit exchange to adjust the circuit to reduce the total circuit length, exchanging k edges, replacing k path segments in the circuit with k new path segments, and searching for the best k-pot sub-circuit exchange in each iteration to most effectively reduce the total circuit length, wherein k represents the number of path segments used to replace the circuit in each iteration;

[0096] S1043: Dynamically adjust the value of k through variable depth search so that it can explore a larger solution space and escape the local optimum, and dynamically adjust the type of k-pot sub-loop exchange for instance-specific adaptation;

[0097] S1044: After multiple iterations of optimization, when the solution is no longer improved, the final minimum energy traversal sequence U is obtained Result , based on U Result , combined with the moving speed v of the directional mobile charger, the moving time length of the directional mobile charger on each path is obtained Thus, the set of all mobile plan items is determined Wherein, d(x0,x1) represents the asymmetric distance of the path from charging position x0 to charging position x1.

[0098] S105, construct a linear programming problem and use the CPlex solver to calculate the length of the charging time list

[0099] Specifically, the construction of the linear programming problem includes:

[0100] Set the objective function to min t 1 1×K t, the constraint condition is C2:t≥0, Where t=[t1,t2,…,t K ] T represents the energy transfer time vector, K represents the number of all corresponding charging directions related to the charging position; constraint C1 ensures that the final energy of all nodes to be charged meets the demand; constraint C2 stipulates that the time is non-negative; constraint C3 represents the energy loss equation caused by energy transfer during the charging process, represents the total energy loss for completing the energy transfer time list t, represents the total energy loss of the directional mobile charger to meet a charging plan s, e B0 represents the initial energy of the directional mobile charger, e F0 It represents the final energy after the directional mobile charger completes a charging schedule. Where K represents all S dir Number of pairs; represents the total energy received by all nodes, where e F =min{e B +e R ,e C}, C is the energy transfer coefficient matrix, C = [c(r,p)] represents S dir Medium charging position x r Along the charging direction ψ p The energy transfer coefficient is but otherwise Where β represents the decay exponent, which controls the decay rate of the distance on the transfer coefficient; α represents a distance correction parameter used to affect the influence of distance d on the transfer coefficient; δ represents a calibration coefficient used to adjust the basic transfer level of the model; d represents the distance between the charging position and the node; the charging duration of each charging direction is set as a variable, and the problem of determining the shortest charging duration list is modeled as a linear programming problem.

[0101] Specifically, the CPlex solver uses a solution idea of ​​a solver, that is, using constraint propagation technology to transfer constraint information between variables in an iterative manner. When the value or value range of a variable changes, the constraints related to it will cause the value ranges of other variables to be updated accordingly. Try to assign different values ​​to the variables until a solution that satisfies all constraints is found, or it is determined that there is no solution. In the solution process, try to reduce the value range of the variable to reduce the search space and speed up the solution process. Find the time vector that minimizes the total energy transfer time to construct a model suitable for CPlex solution, including objective functions and constraints. By inputting relevant data of the problem into the CPlex model, use the solution function to solve the constructed model to obtain a list of charging time lengths of the directional mobile charger in each direction.

[0102] S106, based on S dir and Build Will and Combined, the complete dispatch list S of directional mobile chargers is calculated MC .

[0103] The complete dispatch list of directional mobile chargers is calculated as follows: Where m1 represents the operation schedule list of the directional mobile charger with m1 items, where each item is a tuple state takes values ​​between 0 and 1, x r represents the charging position, ψ represents the optimized charging direction, t represents the duration, and if state=0, it means If state=1, it means

[0104] The items are divided into two categories: the first category is used to arrange the movement plan items between charging locations, and the second category is used to arrange the transmission plan items for transmitting energy along a specific location; if state = 0, it means that the directional mobile charger moves to the location x within a certain period of time. r Move, the moving time is t, and the field ψ is not used; if state = 1, it means that the directional mobile charger should move along the direction of ψ at position x r The energy is transferred for a duration of t. S MC The items in are executed in order to complete the schedule, but express The start time of express The end time of .

[0105] Although the present invention has been specifically shown and described in conjunction with the preferred embodiments, it should be understood by those skilled in the art that various changes may be made to the present invention in form and details without departing from the spirit and scope of the present invention as defined by the appended claims, all of which are within the scope of protection of the present invention.

Claims

1. A method for scheduling directional mobile chargers in WRSNs, characterized in that: The following steps are involved: S101, obtaining a node to be charged, a directional mobile charger and basic charging information; S102, using a K-means-based charging group generation KCPG algorithm to cluster the acquired nodes to be charged, and determine a charging location set X for each cluster; S103, based on the basic charging information and the charging position set X, use the minimum functional representative direction set cMFRDS algorithm to select the optimal charging direction set S equivalent to the omnidirectional infinite direction set dir ; S104, construct asymmetric energy consumption adjacency matrices S and W, and use the improved LKH algorithm to obtain a set of mobile plan items S105, construct a linear programming problem and use the CPlex solver to calculate the charging time length list S106, based on the optimal charging direction set S dir and charging time length list Building a collection of charging scheduling projects A collection of charging scheduling items and a collection of mobile plan items Combined, the complete dispatch list S of directional mobile chargers is calculated MC .

2. The method for scheduling directional mobile chargers in WRSNs according to claim 1, characterized in that: In S101, the basic charging information includes: The set of nodes to be charged U = {u1,u2,…,u n Energy expectation of Initial Energy List Storage capacity list The initial position x0 of the base station and the transmission sector angle of the directional mobile charger Transmission power p0 and moving speed v; where n represents the number of nodes to be charged, and u represents the number of nodes to be charged. i ∈U.

3. The method for scheduling directional mobile chargers in WRSNs according to claim 1, characterized in that: In S102, a charging position set X is determined using a K-means-based charging group generation KCPG algorithm, which specifically includes: Initialization step: set the charging distance to the mobile charger to D and the number of clusters f to 1; Clustering step: Use K-means clustering algorithm to cluster the nodes to be charged into f clusters and obtain the center point of each cluster; Determine the charging location set step: determine whether the radius of the smallest circle corresponding to each cluster is smaller than the charging distance D of the mobile charger; if so, return the center points of these clusters and add them to the charging location set X = {x1, x2, ..., x w }, where w represents the number of charging positions, and charging position x r ∈X; if not, increase the number of clusters f by 1 and re-execute the clustering part; Steps for optimizing the charging position set: Calculate the minimum covering circle of each cluster in the charging position set X, obtain the center point and diameter of the minimum covering circle, and update the charging position set X of each cluster.

4. The method for scheduling directional mobile chargers in WRSNs according to claim 1, characterized in that: In S103, the optimal charging direction set is selected by using the cMFRDS algorithm for creating a minimum functional representative direction set, which specifically includes: Determine the charging direction step S1031: set the charging direction to the mobile charger to be any value in the continuous range [0, 2π), and set the charging distance and charging sector angle to the mobile charger to be constants; Step S1032 of constructing a set of nodes to be charged: constructing a set of nodes to be charged N (x r , D)={u i |u i ∈U(x r , D), x r ∈X},U(x r , D) indicates that the charging position is x r , the set of nodes to be charged whose charging distance is not greater than D; Generate charging direction list step S1033: Build list L θ，a , the node set to be charged N(x r , each node u to be charged in D) i , starting from the center of the circle, with the horizontal right direction as the reference direction, counterclockwise at an angle θ i Sort in ascending order and add to list L θ,a ; Create a list Where L ψ Indicates the charging direction list, |N(x r , D)| represents the set N(x r , the number of elements in D); The step S1034 of optimizing the charging direction set is as follows: S10341: Construct charging direction ψ i The set of all nodes to be charged N(x r , D, ψ i ), where x r represents the charging position, ψ i represents the charging direction, m is the variable of the control loop, and the optimal charging direction set S equivalent to the omnidirectional infinite direction set is dir =[(x r , ψ1)], indicating that the charging position is x r When the charging direction is ψ1, m=1; S10342: From i=2 to |N(x r , D)|loop: judge N(x r , D, ψ i ) is not included in N(x r , D, ψ m ), if N(x, D, ψ i ) is not included in N(x r , D, ψ m ), then execute S10343; otherwise, execute S10344; if the loop ends, enter S10345; S10343: Get N(x r , D, ψ i )'s maximum charging direction angle θ max and the minimum charging direction angle θ min , calculate ψ=(θ max +θ min ) / 2, update m=m+1, update S dir =[(x r , ψ1, ψ)], where ψ represents the optimized charging direction, let i=i+1, and enter S10342; S10344: If N(x, D, ψ i ) is included in N(x, D, ψ m ), then no update is performed, i=i+1 is set, and the process goes to S10342; S10345: After the loop ends, judge: If N(x r , D, ψ1) is included in N(x r , D, ψ i ), then from S dir Remove the first element and finally return the optimal charging direction set S dir , S dir ={(x r , ψ p )|r∈{1, 2, ..., w} represents a charging position x r And all corresponding charging directions ψ p , where ψ1 is the list L ψ The first element of .

5. The method for scheduling directional mobile chargers in WRSNs according to claim 1, characterized in that: In S104, constructing asymmetric energy consumption adjacency matrices S and W specifically includes: Asymmetric distance model steps: quantify and simulate the asymmetry of the path length between any two points in a wireless sensor network; The steps of the asymmetric unit distance motion energy consumption model are as follows: quantify the energy consumption of a directional mobile charger when it moves from one location to another; Determine the asymmetric energy consumption adjacency matrix step: calculate the energy consumption and movement time of the directional mobile charger on different paths, and optimize the path selection and scheduling.

6. The method for scheduling directional mobile chargers in WRSNs according to claim 1, characterized in that: In S104, the improved LKH algorithm for solving the ASTP problem is used to obtain a set of mobile plan items: Specifically include: S1041: Set the set of all locations passed when dispatching to the mobile charger r Π =[x0, x1, ..., x r , x0], which consists of r+2 points, where x0 represents the initial position of the base station, x1,...,x r Indicates a certain charging position. First determine an initial solution, that is, randomly generate a traversal of all r Π The path of the internal node and calculate the path; S1042: repeatedly applying k-pot sub-circuit exchange to adjust the circuit to reduce the total circuit length, exchanging k edges, replacing k path segments in the circuit with k new path segments, and searching for the best k-pot sub-circuit exchange in each iteration to most effectively reduce the total circuit length, wherein k represents the number of path segments used to replace the circuit in each iteration; S1043: Dynamically adjust the value of k through variable depth search so that it can explore a larger solution space and escape the local optimum, and dynamically adjust the type of k-pot sub-loop exchange for instance-specific adaptation; S1044: After multiple iterations of optimization, when the solution is no longer improved, the final minimum energy traversal sequence U is obtained Result , based on U Resul t, combined with the moving speed v of the directional mobile charger, the moving time length of the directional mobile charger on each path is obtained. This determines all mobile scheduling Wherein, d(x0, x1) represents the asymmetric distance of the path from charging position x0 to charging position x1.

7. The method for scheduling directional mobile chargers in WRSNs according to claim 1, characterized in that: In S105, constructing a linear programming problem specifically includes: Set the objective function to The constraints are C2: t ≥ 0, Where t = [t1, t2, ..., t K ] T represents the energy transfer time vector, K represents the number of all corresponding charging directions related to the charging position; constraint C1 ensures that the final energy of all nodes to be charged meets the demand; constraint C2 stipulates that the time is non-negative; constraint C3 represents the energy loss equation caused by energy transfer during the charging process, represents the total energy loss for completing the energy transfer time list t, represents the total energy loss of the directional mobile charger to meet a charging plan s, e B0 represents the initial energy of the directional mobile charger, e F0 It represents the final energy after the directional mobile charger completes a charging schedule. Where K represents all S dir Number of pairs; represents the total energy received by all nodes, where e F =min{e B +e R , e C }, C is the energy transfer coefficient matrix, C = [c (r, p)] represents S dir Medium charging position x r Along the charging direction ψ p The energy transfer coefficient is but otherwise Where β represents the decay exponent, which controls the decay rate of the distance on the transfer coefficient, α represents a distance correction parameter used to affect the influence of distance d on the transfer coefficient, δ represents a calibration coefficient used to adjust the basic transfer level of the model, and d represents the distance between the charging position and the node; the charging duration of each charging direction is set as a variable, and the problem of determining the shortest charging duration list is modeled as a linear programming problem.

8. The method for scheduling directional mobile chargers in WRSNs according to claim 1, characterized in that: In S105, the CPlex solver is used to calculate the length of charging time in each direction. Specifically include: Find the time vector that minimizes the total energy transfer time and build a model suitable for CPlex solution, including the objective function and constraints. By inputting the relevant data of the problem into the CPlex model, use the solve function to solve the built model to obtain a list of charging time lengths for directional mobile chargers in each direction.

Images