Auxiliary charging method for mobile data collection in large-scale multi-task sensor network

Abstract

The application discloses an auxiliary charging method for mobile data collection in a large-scale multi-task sensing network, which solves the power shortage problem of mobile collectors by selectively deploying wireless chargers at data collection points in an economical way. First, the target problem is modeled into a wireless charger minimum cost deployment auxiliary charging problem based on a capacitance constraint, and the capacitance constraint means that the power of the mobile collector may overflow when charging. Then, it is proved that the problem is equivalent to a wireless charger minimum cost deployment auxiliary charging problem without a capacitance constraint, and the problem is modeled into a 0-1 integer programming model. Through redundant constraint pruning and monotone increasing submodular function construction, the 0-1 integer programming model is further converted into a minimum cost submodular cover optimization model, and a greedy strategy based on a submodular function is adopted to iteratively select the data collection point positions of the deployed wireless chargers, so that the final solution is obtained. Compared with other methods, the application has a significant advantage in cost control.

Description

Technical Field

[0001] This invention relates to an auxiliary charging method for mobile data collection in large-scale multi-task sensor networks, belonging to the field of wireless sensor network technology. Background Technology

[0002] Wireless Sensor Networks (WSNs), as a core technology of the Internet of Things (IoT), have made significant progress over time. For WSN applications focused on periodic data collection (such as environmental monitoring), efficiently acquiring data from sensor nodes is a crucial issue. Typically, each sensor node in a monitored area periodically senses local data and then forwards it to the base station via a single-hop or multi-hop path. However, this traditional data collection model can be inefficient in Large-Scale Multitask Wireless Sensor Networks (LSM-WSNs). In LSM-WSNs, a large number of sensor nodes are deployed over a large monitoring area, and each node is responsible for performing multiple tasks (such as temperature and humidity sensing, light intensity sensing, audio collection, image collection, and video collection). In this scenario, the traditional data collection model may consume more energy for data transmission. This is because large-scale multitask scenarios (i.e., large monitoring areas, numerous sensor nodes, and complex tasks with large data volumes) can lead to a higher probability of packet collisions, especially in harsh monitoring environments.

[0003] Many typical applications for periodic data collection (such as environmental monitoring) require sensor nodes to be deployed in harsh environments where equipment replacement is difficult, and demand long network lifecycles. Therefore, extending the network lifetime for such applications is crucial. Mobile data collection is an effective way to extend the lifecycle of LSM-WSNs. By having sensor nodes store sensing data locally and periodically scheduling mobile collectors (mobile vehicles or drones) to actively collect data from the network in a mobile-stationary manner, network lifetime can be significantly extended. This is because sensor nodes only need to locally transmit sensing data to the mobile collector, greatly reducing data transmission energy consumption. While mobile data collection may introduce increased data latency, this is tolerable for periodic data collection applications, which are typically insensitive to latency in practice. Nevertheless, in large-scale, multi-tasking scenarios, mobile collectors may experience power shortages due to their limited battery capacity. The limited battery capacity may not be sufficient to meet the energy demands of long-distance movement and large data volume collection, causing them to run out of power before the end of a data collection cycle. Deploying wireless chargers at data collection points to replenish the power of mobile collectors is a viable solution to the power shortage problem, but this approach incurs additional charger deployment costs. Therefore, optimizing these charger deployment costs while ensuring continuous power supply to multiple mobile collectors within the network is a significant and challenging issue. Summary of the Invention

[0004] The technical problem to be solved by this invention is to provide an auxiliary charging method for mobile data collection in a large-scale multi-task sensor network, which solves the power shortage problem of mobile collectors by selectively deploying wireless chargers at data collection points in an economical manner, and has a significant advantage in terms of charger deployment cost compared with other heuristic methods.

[0005] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0006] A method for assisted charging for mobile data collection in large-scale multi-task sensor networks is disclosed. This method provides assisted charging services to mobile collectors by deploying wireless chargers at candidate data collection points with minimal deployment cost. This enables each mobile collector in the network to continuously perform its periodic data collection tasks without power consumption. Each mobile collector has a limited battery capacity and collects task data at pre-determined data collection points during each data collection cycle. The method includes the following steps:

[0007] Step 1: To address the goal of minimizing the deployment cost of auxiliary charging for wireless chargers oriented towards mobile data collection in large-scale multi-task sensor networks, a 0-1 integer programming model based on capacitance constraints is constructed.

[0008] Step 2: Equivalently convert the 0-1 integer programming model based on capacitance constraints into a 0-1 integer programming model without capacitance constraints;

[0009] Step 3: Perform redundant constraint pruning on the 0-1 integer programming model without capacitance constraints;

[0010] Step 4: Construct non-negative and monotonically non-decreasing submodule functions to transform the 0-1 integer programming model after redundant constraint pruning into an equivalent minimum cost submodule coverage optimization model.

[0011] Step 5: Based on the sub-module function constructed in Step 4, a greedy strategy is used to iteratively select the data collection point locations for deploying the wireless charger, thereby obtaining the final solution.

[0012] As a preferred embodiment of the present invention, the specific process of step 1 is as follows:

[0013] Step 11, given the set of candidate data collection points S = {S1, S2, ..., S...} in the network. n} and the set of user terminals U = {u1, u2, ..., u m}, set each moving collector MC i From the corresponding user terminal u i Starting from [location], it periodically collects task data by passing through a series of pre-determined data collection points, following the path [path]. in and k i These represent the moving collector MC. i The j-th data collection point visited within a data collection cycle and the number of data collection points visited within a data collection cycle, j∈{1,...,k} i}, i∈{1,...,m}; use τ i and These represent the moving collector MC. i Any data collection point it passes through The dwell time and data collection energy consumption rate, and assuming that at the data collection point Deploying a wireless charger will enable the mobile collector MC i exist During the dwell time, it will receive a fixed charging power P. r Charge;

[0014] Step 12, Define E i (l p ,lq ) is the mobile collector MC i From position l p Move to position l q Energy consumption, and use and These represent the moving collector MC. i Arrival at the data collection point within each data collection cycle Remaining energy at time and departure from data collection point Remaining energy at time; for any moving collector MC i Its battery capacity is defined as C. i It is initially fully charged, and it is assumed that it will return to the user terminal after each data collection cycle. i Once the battery is fully charged, it will be ready for the next data collection cycle. Expressed as: in Indicates whether at the data collection point Deploy a 0-1 binary decision variable for a wireless charger. Indicates deployment, Indicates no deployment, and

[0015]

[0016] Step 13: To ensure that each mobile collector in a large-scale multi-task sensor network can continuously perform its periodic data collection task without power interruption, it is necessary to ensure that the remaining energy of each mobile collector when it leaves the data collection point it has passed is sufficient to support its movement to the next data collection point. That is, for any mobile collector MC i Must meet and For any data collection point s∈S, let c(s) denote the cost of deploying a wireless charger at s. The objective problem is how to determine the 0-1 binary decision variables x(S1), x(S2), ..., x(S) n To minimize charger deployment costs This ensures that each mobile collector in the network can continuously perform its periodic data collection task without power interruption; therefore, the objective problem is modeled as a 0-1 integer programming model based on capacitance constraints:

[0017]

[0018]

[0019] in,

[0020] As a preferred embodiment of the present invention, the specific process of step 2 is as follows:

[0021] Step 21, Define For MC i In its subpath The set of sequence numbers of data collection points where battery capacity overflow may occur, i.e. And MC i It may be at any data collection point it passes through Battery capacity overflow occurred. For each MC i , Obtain it through the following method: First, initially set... Then, assuming that a charger is deployed at each candidate data collection point, the MC is determined sequentially. i At each data collection point along its path Does the battery capacity overflow at any time, j∈{1,...,k}? i If the sequence overflows, add the sequence number j to the set. middle;

[0022] Step 22, Define For the moving collector MC under the assumption of no capacitance constraint i From leaving It's time to leave The total energy gain during the period of time, i.e. The 0-1 integer programming model based on capacitance constraints is equivalently transformed into the following 0-1 integer programming model without capacitance constraints:

[0023]

[0024] As a preferred embodiment of the present invention, the specific process of step 3 is as follows:

[0025] Step 31, remove the constraints from the 0-1 integer programming model without capacitance constraints. Equivalents can be obtained through expansion. Form, due to arbitrary Will Equivalent expression is in Let represent a known binary coefficient, and If and only if there exists f∈{k+1,k+2,...,j} such that... Define integer coefficients Therefore, the constraints in the 0-1 integer programming model without capacitance constraints will be... Transform it into the following form:

[0026] Step 32: Based on the transformed constraint form in Step 31, identify the set of redundant constraints in the 0-1 integer programming model without capacitance constraints. Any constraint... A redundant constraint is a constraint if and only if any one of the following conditions is satisfied: 1) β in the constraint ijk ≤0; 2) There exists at least one other constraint. So that And β i′j′k′ ≥β ijk ,in

[0027] Step 33: Remove the redundant constraint set identified in step 32 from all constraints. The 0-1 integer programming model without capacitance constraints is further expressed as:

[0028]

[0029]

[0030] Where, N c This represents the total number of constraints after redundant constraint pruning, which is arbitrary. Let β represent a known 0-1 binary coefficient. r Represents a known positive integer coefficient.

[0031] As a preferred embodiment of the present invention, the specific process of step 4 is as follows:

[0032] Step 41, define a finite set E = {1, 2, ..., n}, for any set Define a non-negative and monotonically non-decreasing submodule function F(·):2 E →N is as follows:

[0033]

[0034] Step 42: Based on the submodule function F(·), the 0-1 integer programming model after redundant constraint pruning is transformed into an equivalent minimum cost submodule coverage optimization model as follows:

[0035]

[0036] As a preferred embodiment of the present invention, the specific process of step 5 is as follows:

[0037] Step 51, Initial Setup

[0038] Step 52, from set E\T* Greedily select the optimal element x * , making The value reaches its maximum, then the set T is updated. * =T * ∪{x *};

[0039] Step 53, repeat step 52 until for all elements x∈E\T * Both have F(T) * ∪{x})=F(T * Until it is established;

[0040] Step 54, for any t∈E, if t∈T * Then set x(S) t ) = 1; if t ∈ E\T * Then set x(S) t ) = 0, meaning the final set of data collection points for deploying wireless chargers is S. * ={S t |t∈T *}

[0041] Compared with the prior art, the present invention, employing the above technical solution, has the following technical effects:

[0042] 1. This invention can be widely applied to application scenarios in large-scale multi-task sensor networks where multiple mobile data collectors periodically perform data collection tasks. By deploying wireless chargers at data collection points at minimal cost, it provides power to multiple mobile collectors with data collection tasks, enabling the mobile data collectors to operate continuously, wherein each mobile data collector periodically travels along a fixed route.

[0043] 2. This invention demonstrates that it can achieve an effective approximation ratio of theoretical performance, and extensive simulation experiments show that this invention has a significant advantage in charger deployment cost compared to other heuristic methods.

[0044] 3. This invention has a certain degree of applicability and can be applied to general scenarios such as heterogeneous mobile collectors and heterogeneous data collection point deployment costs, while also having a certain degree of scalability. Attached Figure Description

[0045] Figure 1 This is a flowchart of the present invention;

[0046] Figure 2 This is a schematic diagram of the network model in an embodiment of the present invention. Detailed Implementation

[0047] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0048] This invention provides an auxiliary charging method for mobile data collection in a large-scale multi-task sensor network, the flowchart of which is shown below. Figure 1 As shown, this invention is applied to a wireless sensor network consisting of multiple data collection points and multiple user terminals, where each user terminal utilizes a mobile collector to perform periodic data collection tasks along a mobile path passing through a pre-determined data collection point. The invention considers how to deploy chargers at data collection points at minimal cost so that all mobile collectors can operate continuously, specifically including the following steps:

[0049] (1) To address the minimum-cost deployment of auxiliary charging for wireless chargers oriented towards mobile data collection in large-scale multi-task sensor networks, a 0-1 integer programming model based on capacitance constraints is constructed; specifically, the following steps are included:

[0050] (11): Given a set of candidate data collection points S = {S1, S2, ..., S...} in the network n} and the set of user terminals U = {u1, u2, ..., u m}, assuming each mobile collector (mobile car or drone) has MC i From the corresponding user terminal u i Starting from [location], it periodically collects task data by passing through a series of pre-determined data collection points, following the path [path]. in and k i These represent the moving collector MC. i The j-th data collection point visited within a data collection cycle and the number of data collection points visited within a data collection cycle. Let τ be the criterion. i and These represent the moving collector MC. i Any data collection point it passes through Dwell time and data collection energy consumption rate (here, due to each mobile collector MC) i It collects the same type of task data at all the data collection points it passes through, therefore MC i The dwell time for data collection at each data collection point it passes through can be set to the same τ. i ), and assume that if at the data collection point Deploying a wireless charger will enable the mobile collector MC i exist During the dwell time, it will receive a fixed charging power P. r Charge it.

[0051] (12): Define E i (l p ,l q ) is the mobile collector MC i From position l p Move to position l q Energy consumption, and use and These represent the moving collector MC. i Arrival at the data collection point within each data collection cycle Remaining energy at time and departure from data collection point The remaining energy at that time. For any moving collector MC i Its battery capacity is defined as C. i It is initially fully charged, and it is assumed that it will return to the user terminal u after each data collection cycle. i After fully charging, it will prepare for the next data collection cycle. Because the mobile collector's limited battery capacity may cause energy overflow after charging at any data collection point where a charger is deployed, therefore... This can be expressed as: in Indicates whether at the data collection point Deploying a wireless charger: 0-1 binary decision variables ( Indicates deployment, (Indicates no deployment), and

[0052]

[0053] (13): In order to ensure that each mobile collector in a large-scale multi-task network can continuously perform its periodic data collection tasks without power consumption, it is necessary to ensure that the remaining energy of each mobile collector when it leaves the data collection point it has passed is sufficient to support its movement to the next data collection point. That is, for any mobile collector MC i Must meet and For any data collection point s∈S, let c(s) denote the cost of deploying a wireless charger at s. The objective is to determine how to determine the 0-1 binary decision variables x(S1), x(S2), ..., x(S) n To minimize charger deployment costs This ensures that each mobile collector in the network can continuously perform its periodic data collection task without power interruption. Therefore, the objective problem can be modeled as a 0-1 integer programming model based on capacitance constraints:

[0054]

[0055]

[0056] Among them, the definition

[0057] (2) The 0-1 integer programming model based on capacitance constraints is equivalently converted into a 0-1 integer programming model without capacitance constraints; specifically, the following steps are included:

[0058] (21): Definition For MC i In its subpath The set of sequences of data collection points where battery capacity overflow may occur; in other words, the set of sequences of data collection points where battery capacity overflow may occur. And MC i It may be at any data collection point it passes through Battery capacity overflow occurs. For each MC i (i∈{1,...,m}), This can be obtained through the following method: First, initial setup. Then, assuming that a charger is deployed at each candidate data collection point, the MC is determined sequentially. i At each data collection point along its path If the battery capacity overflows, then add the sequence number j to the set. middle.

[0059] (22): Definition For the moving collector MC under the assumption of no capacitance constraint i From leaving It's time to leave The total energy gain during the period of time, i.e. Model the following 0-1 integer programming model without capacitance constraints:

[0060]

[0061]

[0062] (23): By proving that the 0-1 integer programming model based on capacitance constraints in step (13) and the 0-1 integer programming model without capacitance constraints in step (22) have the same range of feasible solution space, it can be proved that the 0-1 integer programming model based on capacitance constraints in step (13) is essentially equivalent to the 0-1 integer programming model without capacitance constraints in step (22).

[0063] (3) Redundancy constraint pruning is performed on the 0-1 integer programming model after the equivalent transformation; specifically, the following steps are included:

[0064] (31) The constraints in the 0-1 integer programming model without capacitance constraints Equivalents can be obtained through expansion. Form. Due to arbitrary therefore It can be equivalently expressed as in Let represent a known binary coefficient, and If and only if there exists f∈{k+1,k+2,...,j} such that... For ease of expression, integer coefficients are defined. Therefore, arbitrary constraints in a 0-1 integer programming model without capacitance constraints It can be transformed into the following form:

[0065] (32): Based on the transformed constraint form in step (31), identify the set of redundant constraints in the capacitor-free 0-1 integer programming model. Here, any constraint A redundant constraint is a constraint if and only if any one of the following conditions is satisfied: 1) β in the constraint ijk ≤0; 2) There exists at least one other constraint. So that And β i′j′k′ ≥β ijk ,in

[0066] (33): Remove the redundant constraint set identified in step (32) from all constraints. The 0-1 integer programming model without capacitance constraints can be further expressed as:

[0067]

[0068]

[0069] Where N c This represents the total number of constraints after redundant constraint pruning, which is arbitrary. Let β represent a known 0-1 binary coefficient. r Represents a known positive integer coefficient.

[0070] (4) By constructing monotonically increasing submodule functions, the 0-1 integer programming model after redundant constraint pruning is transformed into an equivalent minimum cost submodule coverage optimization model; specifically, the following steps are included:

[0071] (41): Define a finite set E = {1, 2, ..., n}, for any set Define a function F(·):2 E →N is as follows:

[0072]

[0073] It can be proven that F(·) is a non-negative and monotonically non-decreasing submodule function.

[0074] (42): Based on the submodular function F(·), the following minimum cost submodular coverage optimization model is constructed:

[0075]

[0076]

[0077] (43): It can be proven that the capacitor-free 0-1 integer programming model in step (33) is equivalent to the minimum cost submodule coverage optimization model in step (42).

[0078] (5) Based on the constructed sub-modulus function, a greedy strategy is used to iteratively select the data collection point locations for deploying wireless chargers, thereby obtaining the final solution; specifically, the following steps are included:

[0079] (51): Initial setup

[0080] (52): From set E\T * Greedily select the optimal element x * , making The value reaches its maximum, then the set T is updated. * =T * ∪{x *}

[0081] (53): Repeat step (52) until all elements x∈E\T * Both have F(T) * ∪{x})=F(T * Until it is established.

[0082] (54): For any t∈E, if t∈T * Then set x(S) t ) = 1; if t ∈ E\T * Then set x(S) t ) = 0. In other words, the final set of data collection points for deploying wireless chargers is S. * ={S t |t∈T *}

[0083] like Figure 2As shown, the network has four candidate data collection points S1, S2, S3, and S4, and two user terminals u1 and u2. Two mobile collectors MC1 and MC2 start from user terminals u1 and u2 respectively. The movement route of MC1 in each data collection cycle is as follows:<u1,S1,S2,S3,u1> The movement route of MC2 within each data collection cycle is as follows:<u2,S2,S3,S4,u2> ,Right now Assume the mobile collector battery capacity is C1 = C2 = C = 4, the dwell time of mobile collector MC1 at each data collection point is τ1 = 2, the dwell time of mobile collector MC2 at each data collection point is τ2 = 1, and the charging received power P of the mobile collector at each data collection point is... r =2. It's not hard to see that the moving collector MC i At the data collection point The amount of electricity that can be charged during the stay time is in, It is a binary decision variable. If and only if at the data collection point Deploy the charger. Assume any moving collector MC i At the data collection point Energy consumption rate during data collection Then move collector MC i At any data collection point Data collection energy consumption is For simplicity and without loss of generality, assume an arbitrary moving collector MC. i From any data collection point Move to the next data collection point The energy consumption for each mobile collector is a fixed value e, and e = 1. To ensure that each mobile collector can continuously perform periodic data collection within the network, the remaining energy of each mobile collector when leaving each data collection point on its route must be sufficient to move to the next data collection point; that is, the energy remaining for each mobile collector MC is equal to the energy required to move to the next data collection point. i Leaving the data collection point Energy of time Greater than or equal to the mobile energy consumption e. Further, assume that at each candidate data collection point S... i The cost of deploying chargers c(S) i All values ​​are fixed at 1. Based on steps (11) to (13), the following 0-1 integer programming model based on capacitance constraints can be constructed:

[0084] min c(S1)x(S1)+c(S2)x(S2)+c(S3)x(S3)+c(S4)x(S4)

[0085] st

[0086]

[0087]

[0088]

[0089]

[0090]

[0091]

[0092] According to the method in step (21), the following can be calculated: Next, according to step (22), the above 0-1 integer programming model based on capacitance constraints can be transformed into the following equivalent 0-1 integer programming model without capacitance constraints:

[0093] min c(S1)x(S1)+c(S2)x(S2)+c(S3)x(S3)+c(S4)x(S4)

[0094] st

[0095] Ce-τ1+P r τ1x(S1)≥e

[0096] Ce-τ1+P r τ1x(S1)-e-τ1+P r τ1x(S2)≥e

[0097] Ce-τ1+P r τ1x(S2)≥e

[0098] Ce-τ1+P r τ1x(S1)-e-τ1+P r τ1x(S2)-e-τ1+P r τ1x(S3)≥e

[0099] Ce-τ1+P r τ1x(S2)-e-τ1+P r τ1x(S3)≥e

[0100] Ce-τ1+P r τ1x(S3)≥e

[0101] Ce-τ2+P r τ2x(S2)≥e

[0102] Ce-τ2+Pr τ2x(S2)-e-τ2+P r τ2x(S3)≥e

[0103] Ce-τ2+P r τ2x(S2)-e-τ2+P r τ2x(S3)-e-τ2+P r τ2x(S4)≥e

[0104] Substituting the above 0-1 integer programming model without capacitance constraints into specific values, and according to steps (31) to (33), the following model after redundancy constraint pruning can be obtained:

[0105] min x(S1)+x(S2)+x(S3)+x(S4)

[0106] st

[0107] x(S1)+x(S2)≥1

[0108] x(S1)+x(S2)+x(S3)≥2

[0109] x(S2)+x(S3)≥1

[0110] x(S2)+x(S3)+x(S4)≥2

[0111] Given a finite set E = {1, 2, 3, 4}, for any set According to step (41), define the function F(·):2 E →N is as follows:

[0112]

[0113] It can be proven that this is a non-negative and monotonically non-decreasing submodule function. Based on this submodule function, a minimum cost submodule coverage optimization model as shown in step (42) can be constructed.

[0114] According to step (51), the initial settings are as follows:

[0115] Based on steps (52) and (53), first from set E\T * Greedily select the optimal element x from {1,2,3,4} * So that The value of x reaches its maximum, and x is obtained through calculation and comparison. * =2, then update set T * =T * ∪{x *} = {2}; Next, continue from set E\T *Greedily select the optimal element x from {1,3,4} * So that The value of x reaches its maximum, and x is obtained through calculation and comparison. * =3, then update set T * =T * ∪{x *}={2,3}. At this point, it is found that for E\T * =All elements in {1,4} satisfy F(T) * ∪{1})=F(T * ) = 6 and F(T) * ∪{4})=F(T * Since ) = 6, the iteration process terminates, and we finally obtain T. * ={2,3}.

[0116] According to step (54), we can obtain x(S2) = x(S3) = 1, x(S1) = x(S4) = 0, that is, S * = {S2, S3}. Therefore, the final solution of this embodiment is to deploy wireless chargers at data collection points S2 and S3.

[0117] The above embodiments are merely illustrative of the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solutions based on the technical concept proposed in this invention shall fall within the scope of protection of this invention.

Claims

1. A method for assisted charging for mobile data collection in large scale multi-task sensor networks, characterized in that, The method is aimed at large-scale multi-task sensor network, and provides auxiliary charging service for mobile collectors by deploying wireless chargers at candidate data collection points with minimum charging deployment cost, so that each mobile collector in the network can continuously perform its own periodic data collection task, wherein each mobile collector has limited battery capacity and collects task data at each data collection point within each data collection period. Step 1, for the minimum cost deployment of auxiliary charging target problem of wireless chargers for mobile data collection in large-scale multi-task sensor network, a 0-1 integer programming model based on capacitance constraint is constructed; the specific process is as follows: Step 11, given the set of candidate data collection points S = {S1, S2, ..., S...} in the network. n } and the set of user terminals U = {u1, u2, ..., u m }, set each moving collector MC i From the corresponding user terminal u i Starting from [location], it periodically collects task data by passing through a series of pre-determined data collection points, following the path [path]. in and k i These represent the moving collector MC. i The j-th data collection point visited within a data collection cycle and the number of data collection points visited within a data collection cycle, j∈{1,...,k} i }, i∈{1,...,m}; use τ i and These represent the moving collector MC. i Any data collection point it passes through The dwell time and data collection energy consumption rate, and assuming that at the data collection point Deploying a wireless charger will enable the mobile collector MC i exist During the dwell time, it will receive a fixed charging power P. r Charge; Step 12, define E i (l p ,l q ) as the mobile collector MC i moving from location l p to location l q , and let and denote the remaining energy of the mobile collector MC i when it arrives at the data collection point and when it leaves the data collection point in each data collection period, respectively; for any mobile collector MC i , let its battery capacity be C i and initially be in full energy state, and assume that it will return to the user terminal u i after each data collection period to be recharged to full energy before preparing for the next period of data collection; then is expressed as: where is a 0-1 binary decision variable indicating whether or not to deploy a wireless charger at the data collection point , denotes deployment, denotes no deployment, and Step 13, to ensure each mobile collector in the large-scale multi-task sensor network can continuously perform its periodic data collection task, it is required to satisfy that the residual energy of each mobile collector when it leaves a passed data collection point can support it to move to the next data collection point, i.e. for any mobile collector MC i It is required to satisfy and For any data collection point s∈S, let c(s) denote the cost of deploying a wireless charger at s, then the target problem is how to determine 0-1 binary decision variables x(S1), x(S2),..., x(S n ) to minimize the charger deployment cost so as to ensure each mobile collector in the network can continuously perform its periodic data collection task; therefore, the target problem is modeled into the following 0-1 integer programming model based on the capacitive constraint: wherein Step 2, the 0-1 integer programming model based on capacitance constraint is equivalent to the 0-1 integer programming model without capacitance constraint; the specific process is as follows: Step 21, define for MC i in its sub-path , i.e. and MC i may overflow at any data collection point it passes through, for each MC i , by the following method: first, initially set then, in the case of assuming that each candidate data collection point is deployed with a charger, sequentially judge whether the battery capacity of MC i overflows at each data collection point on its path, j∈{1,...,k i -1}, if it overflows, then add the sequence number j to the set , respectively;​ Step 22, define For the mobile collector MC under the assumption of no capacitance constraint i From the time of departure The total energy gain within the period from the time of departure The total energy gain within the period from the time of departure The 0-1 integer programming model based on the capacitance constraint will be equivalent to the following 0-1 integer programming model without capacitance constraint: Step 3, the 0-1 integer programming model without capacitance constraint is pruned by redundant constraints; Step 4, a non-negative and monotonically non-decreasing submodular function is constructed, and the 0-1 integer programming model after pruning redundant constraints is converted into an equivalent minimum cost submodular cover optimization model; Step 5, based on the submodular function constructed in step 4, the data collection point positions of deploying wireless chargers are selected iteratively by using greedy strategy, so as to obtain the final solution.

2. The method of claim 1, wherein, The specific process of step 3 is as follows: Step 31, transforming the constraints in the 0-1 integer programming model without capacitated constraints by the expansion transformation into the equivalent form, since any is equivalent to is equivalent to where denotes a known binary coefficient, and if and only if there exists f e {k+1, k+2,..., j} such that define the integer coefficient Thus, the constraints in the 0-1 integer programming model without capacitated constraints are transformed into the form: Step 32, based on the transformed constraint form in step 31, identify a redundant constraint set in the 0-1 integer programming model without capacitated constraints, any one constraint is a redundant constraint if and only if any one of the following conditions is satisfied: 1) β ijk ≤ 0 in the constraint; 2) there exists at least one other constraint such that and β i′j′k′ ≥ β ijk where Step 33, the redundant constraint set identified in step 32 is removed from all constraints, and the 0-1 integer programming model without capacitance constraint is further expressed as: where N c represents the number of all constraints after pruning by redundancy constraint, arbitrary represents a known 0-1 binary coefficient, arbitrary β r represents a known positive integer coefficient.

3. The method of claim 2, wherein, The specific process of step 4 is as follows: Step 41, define a finite set E = {1, 2,..., n}, for any set Define a non-negative and monotone non-decreasing sub-module function F(·): 2 E → N as follows: Step 42, based on the submodular function F(·), the 0-1 integer programming model after pruning redundant constraints is converted into an equivalent minimum cost submodular cover optimization model as follows:

4. The method of claim 3, wherein, The specific process of step 5 is as follows: Step 51, initially set Step 52, from set E\T * Greedily select the optimal element x * , so that (F(T) * ∪{x * })-F(T * )) / c(S x* The value of ) reaches its maximum, then the set T is updated. * =T * ∪{x * }; Step 53, repeat step 52 until for all elements x E T * F(T * ∪{x}) = F(T * ) holds; Step 54, for any t E, if t E T * , set x(S t ) = 1; if t E E\T * , set x(S t ) = 0, that is, the final deployment of wireless charger data collection point set S * = {S t |t E T *}.

Images