Base station scheduling method, apparatus, device, storage medium and product

Abstract

The application discloses a base station scheduling method, device, equipment, storage medium and product, and multiple rounds of iterations are performed on a base station scheduling scheme. In each round of iteration, base station transportation distance, remaining amount of a parking point base station and bandwidth redundancy of the base station scheduling scheme are calculated, the base station scheduling scheme is adjusted, and the next round of iteration is entered until an iteration termination condition is met, and a final base station scheduling scheme is output. The bandwidth demand of each to-be-optimized cell, the performance of each portable base station and the distance between the portable base station and the to-be-optimized cell are fully considered, so that the to-be-optimized cell can be reasonably allocated to a suitable base station, and the accuracy of base station allocation is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of communication technology, in particular to a base station scheduling method, device, equipment, storage medium and product. BACKGROUND

[0002] The portable base station is a new type of base station with small size and mobility, which can be deployed on vehicles, has the characteristics of agile deployment, high mobility, self-provided power supply and radio frequency module, and can quickly access existing communication networks, etc., and can realize the rapid recovery of signal coverage in the wireless communication fault area. Therefore, the portable base station can flexibly cope with various temporary support events, including emergency support (such as natural disasters, meeting support), network optimization (such as temporary black spots, business surge), etc., and is one of the preferred network operation and maintenance solutions for operators.

[0003] However, in the optimal device analysis process before the device scheduling transportation, expert experience analysis is mainly adopted, that is, the current running state of the device is counted offline and the geographical position and site performance data are analyzed, and the schedulable device is selected and assigned through expert experience. This method has obvious limitations and cannot accurately schedule the base station. SUMMARY

[0004] Based on this, the present application provides a base station scheduling method, device, equipment, storage medium and product to solve the defect that the accuracy of base station scheduling by manual operation in the prior art is poor.

[0005] To achieve the above-mentioned purpose, the embodiment of the present application provides a base station scheduling method, comprising:

[0006] In the base station scheduling process, the base station scheduling scheme is iterated for multiple rounds;

[0007] In each round of iteration, the portable base station is scheduled from the base station parking point to the cell to be optimized according to the current base station scheduling scheme, and the base station transportation distance, the base station parking point remaining amount and the bandwidth redundancy of the base station scheduling scheme are calculated to adjust the base station scheduling scheme and enter the next round of iteration until the iteration termination condition is met, and output the final base station scheduling scheme; wherein the base station scheduling scheme refers to the scheme of scheduling the portable base station from each base station parking point to each cell to be optimized, and the bandwidth redundancy refers to the difference between the bandwidth of the cell to be optimized after the base station scheduling and the bandwidth demand of the cell to be optimized.

[0008] As an improvement of the above scheme, the base station scheduling scheme is iterated in multiple rounds in the base station scheduling process; in each round of iteration, the portable base stations are scheduled from the base station parking points to the to-be-optimized cells according to the current base station scheduling scheme, and the base station transportation distance, the base station parking point residual quantity and the bandwidth redundancy of the base station scheduling scheme are calculated to adjust the base station scheduling scheme and enter the next round of iteration until the iteration termination condition is met, and the final base station scheduling scheme is output, including:

[0009] A base station parking point is taken as an agent to construct multiple agents of reinforcement learning;

[0010] All agents are learned in multiple rounds;

[0011] In each round, all agents gradually perform actions on each to-be-optimized cell according to the current strategy to gradually schedule portable base stations from all base station parking points for each to-be-optimized cell;

[0012] The round reward is calculated after each round to adjust the strategy of the agent; wherein the round reward includes a first reward and a second reward, the first reward is calculated according to the bandwidth redundancy and the base station transportation distance, the bandwidth redundancy refers to the difference between the bandwidth of the to-be-optimized cell after base station scheduling and the bandwidth demand of the to-be-optimized cell, and the base station transportation distance refers to the sum of the transportation distances of all portable base stations scheduled by the to-be-optimized cell; the second reward has a positive correlation with the minimum base station residual quantity, and the minimum base station residual quantity is the minimum value of the number of residual portable base stations in all base station parking points after completing base station scheduling;

[0013] After the reinforcement learning is completed, the final base station scheduling scheme is output to schedule the portable base stations from the base station parking points for the to-be-optimized cells.

[0014] As an improvement of the above scheme, in each round, all agents gradually perform actions on each to-be-optimized cell according to the current strategy to gradually schedule portable base stations from all base station parking points for each to-be-optimized cell, including:

[0015] In the target step of the target round, all agents perform actions according to the strategy to schedule portable base stations from all base station parking points for the target to-be-optimized cell; wherein the target round is any one of all rounds, the target step is any one of multiple steps of the target round, and the target to-be-optimized cell is a to-be-optimized cell corresponding to the target step;

[0016] calculating a first reward for the target step, wherein the first reward of the target step is calculated according to a bandwidth redundancy of the target cell to be optimized and a base station transportation distance of the target cell to be optimized;

[0017] adjusting the strategy according to the first reward, and entering a next step of the target step until all the cells to be optimized are completed with base station scheduling, and ending the target round.

[0018] As an improvement of the above scheme, the adjusting the strategy according to the first reward comprises adjusting the strategy according to the first reward by using a Critic network.

[0019] As an improvement of the above scheme, the base station scheduling scheme is iterated for multiple rounds in the base station scheduling process; in each round of iteration, a portable base station is scheduled from a base station parking point to a cell to be optimized according to a current base station scheduling scheme, and a base station transportation distance, a parking point base station remaining amount and a bandwidth redundancy of the base station scheduling scheme are calculated to adjust the base station scheduling scheme and enter a next round of iteration until an iteration termination condition is met, and a final base station scheduling scheme is output, comprising:

[0020] obtaining initial information; the initial information comprises all cells to be optimized and initial bandwidths and bandwidth requirements thereof, all base station parking points and base station storage amounts thereof, distances between the cells to be optimized and the base station parking points;

[0021] constructing a cell to be optimized group, an initial bandwidth row vector, a bandwidth requirement row vector, a base station parking point group, a parking point base station storage amount row vector and a distance matrix for representing distances between the cells to be optimized and the base station parking points according to the initial information, so as to serve as basic information of a mathematical model;

[0022] constructing an objective function for the mathematical model, the objective function being a sum of transportation distances of all scheduled portable base stations, the transportation distance referring to a distance between a base station parking point where the portable base station is located and a cell to be optimized where the portable base station is scheduled to;

[0023] constructing a quantity constraint condition and a bandwidth constraint condition for the mathematical model, the quantity constraint condition being that a number of scheduled portable base stations in the base station parking point cannot exceed a number of portable base stations originally parked in the base station parking point, and the bandwidth constraint condition being that a bandwidth of the cell to be optimized after base station scheduling is greater than or equal to a bandwidth requirement of the cell to be optimized;

[0024] Solving the objective function under the quantity constraint and the bandwidth constraint obtains a final base station scheduling scheme, so as to schedule the portable base station at the base station parking point to the to-be-optimized cell according to the base station scheduling scheme.

[0025] As an improvement of the above scheme, the base station scheduling scheme is iterated for multiple rounds in the base station scheduling process. In each round of iteration, the portable base station is scheduled from the base station parking point to the to-be-optimized cell according to the current base station scheduling scheme, and the base station transportation distance, the parking point base station remaining quantity and the bandwidth redundancy of the base station scheduling scheme are calculated, so as to adjust the base station scheduling scheme and enter the next round of iteration, until the iteration termination condition is met, and the final base station scheduling scheme is output, including:

[0026] The initial information is used to construct a to-be-optimized cell group, an initial bandwidth row vector, a bandwidth demand row vector, a base station parking point group, a parking point base station storage quantity row vector and a distance matrix for representing the distance between the to-be-optimized cell and the base station parking point, so as to serve as the basic information of the mathematical model;

[0027] The first loss function, the second loss function and the third loss function are constructed for the mathematical model. The first loss function is equal to the difference between the parking point base station storage quantity row vector and the parking point base station scheduling quantity row vector of the current base station scheduling scheme. The parking point base station scheduling quantity row vector represents the quantity of the portable base station scheduled by each base station parking point. The second loss function is equal to the initial bandwidth row vector plus the bandwidth increment row vector of the current base station scheduling scheme minus the bandwidth demand row vector. The bandwidth increment row vector represents the bandwidth increment of each to-be-optimized cell after base station scheduling. The third loss function is equal to the inverse number of the sum of the transportation distances of each scheduled portable base station in the current base station scheduling scheme. The transportation distance refers to the distance between the base station parking point where the portable base station is located and the to-be-optimized cell to which the portable base station is scheduled.

[0028] The base station scheduling scheme is iterated, and the first loss function, the second loss function and the third loss function after each round of iteration are calculated.

[0029] The base station scheduling scheme of the iteration round that makes the total loss function maximum is selected as the final base station scheduling scheme for base station scheduling, wherein the total loss function is in a positive correlation with the first loss function, the second loss function and the third loss function respectively.

[0030] To achieve the above object, the embodiment of the present application further provides a base station scheduling device, which comprises:

[0031] The scheme iteration module is configured to iterate the base station scheduling scheme for multiple rounds in the base station scheduling process.

[0032] a base station scheduling module, configured to, in each iteration, schedule portable base stations from base station parking points to to-be-optimized cells according to a current base station scheduling scheme, and calculate base station transportation distances, base station parking point residual amounts and bandwidth redundancy amounts of the base station scheduling scheme, so as to adjust the base station scheduling scheme and enter a next iteration, until an iteration termination condition is met, and output a final base station scheduling scheme; wherein the base station scheduling scheme refers to a scheme of scheduling portable base stations from various base station parking points to various to-be-optimized cells, and the bandwidth redundancy amount refers to a difference between a bandwidth of the to-be-optimized cell after base station scheduling and a bandwidth demand of the to-be-optimized cell.

[0033] To achieve the above object, the embodiment of the present application further provides a base station scheduling device, comprising a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, and the processor implements the base station scheduling method as any of the above embodiments when executing the computer program.

[0034] To achieve the above object, the embodiment of the present application further provides a computer readable storage medium, comprising a stored computer program, wherein the computer readable storage medium controls a device where the computer readable storage medium is located to execute the base station scheduling method as any of the above embodiments when the computer program runs.

[0035] To achieve the above object, the embodiment of the present application further provides a computer program product, comprising computer programs / instructions, and the computer programs / instructions implement the base station scheduling method as any of the above embodiments when executed by a processor.

[0036] Compared with the prior art, the base station scheduling method, device, equipment, storage medium and product disclosed by the embodiment of the present application, in the base station scheduling process, perform multiple iterations on the base station scheduling scheme, in each iteration, calculate base station transportation distances, base station parking point residual amounts and bandwidth redundancy amounts of the base station scheduling scheme for adjusting the base station scheduling scheme and entering a next iteration, until an iteration termination condition is met, and output a final base station scheduling scheme; wherein the base station scheduling scheme refers to a scheme of scheduling portable base stations from various base station parking points to various to-be-optimized cells, and the bandwidth redundancy amount refers to a difference between a bandwidth of the to-be-optimized cell after base station scheduling and a bandwidth demand of the to-be-optimized cell, fully considering the bandwidth demand of each to-be-optimized cell, the performance of each portable base station and the distance between the portable base station and the to-be-optimized cell, so that the to-be-optimized cell can be reasonably allocated to a suitable base station, and the accuracy of base station allocation is improved. BRIEF DESCRIPTION OF DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the present application, the drawings required to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings described in the following embodiments are only some of the embodiments of the present application, and all other drawings obtained by those of ordinary skill in the art without creative effort based on these drawings also belong to the protection scope of the present application.

[0038] Figure 1 is a flowchart of a base station scheduling method provided by an embodiment of the present application;

[0039] Figure 2 is a structural diagram of a base station scheduling device provided by an embodiment of the present application;

[0040] Figure 3 is a structural diagram of a base station scheduling device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort belong to the protection scope of the present application.

[0042] Referring to Figure 1 , Figure 1 is a flowchart of a base station scheduling method provided by an embodiment of the present application. The base station scheduling method comprises steps S11-S12:

[0043] S11, in the base station scheduling process, performing multiple rounds of iteration on a base station scheduling scheme;

[0044] S12, in each round of iteration, scheduling portable base stations from base station parking points to to-be-optimized cells according to a current base station scheduling scheme, and calculating base station transportation distances, base station parking point residual amounts and bandwidth redundancy amounts of the base station scheduling scheme, so as to adjust the base station scheduling scheme and enter the next round of iteration, until an iteration termination condition is met, and outputting a final base station scheduling scheme; wherein the base station scheduling scheme refers to a scheme of scheduling portable base stations from various base station parking points to various to-be-optimized cells, and the bandwidth redundancy amount refers to a difference between a bandwidth of the to-be-optimized cell after base station scheduling and a bandwidth demand of the to-be-optimized cell.

[0045] Specifically, in a set area, a plurality of base station parking points are provided, containing a plurality of cells, when the network of a cell needs to be optimized (such as network failure), the cell can be used as a to-be-optimized cell, and the vehicle deployed with a portable base station is dispatched from the base station parking point to the to-be-optimized cell to optimize the network of the cell. The method calculates the base station transportation distance of all portable base stations dispatched by using different base station scheduling schemes, the number of portable base stations not dispatched at each parking point (the remaining amount of parking point base stations), and the difference between the bandwidth of each to-be-optimized cell after base station scheduling and the bandwidth demand of the cell itself (bandwidth redundancy). According to the base station transportation distance, the remaining amount of parking point base stations, and the bandwidth redundancy corresponding to the base station scheduling scheme, the performance of the base station scheduling scheme is analyzed, the base station scheduling scheme is adjusted, and the best base station scheduling scheme is obtained to perform base station scheduling. It can be understood that after the base station scheduling is completed, the fewer vehicles dispatched in each base station parking point, the larger the bandwidth of the to-be-optimized cell, and the shorter the driving distance of all vehicles dispatched from the base station parking point to the to-be-optimized cell, so as to realize accurate scheduling of the base station, that is, to save resources and improve the bandwidth of the to-be-optimized cell.

[0046] Compared with the prior art, the method significantly saves manual analysis time and improves the accuracy of base station scheduling, which is beneficial to effectively solve the problems of emergency communication support, black spot rectification, load optimization, and the like, can significantly improve the network satisfaction, and has good social benefits. In addition, through investigation, China Mobile has deployed more than 5000 portable base stations (communication cabinets) nationwide. Considering the temporary nature of equipment deployment, online management and intelligent scheduling analysis of equipment based on automatic means can significantly reduce the workload of frontline maintenance personnel. With a single device reducing the workload of maintenance personnel by 0.1 person-year per year, and the labor cost being 50,000 yuan per year, it is estimated that the labor cost will be reduced by about 25 million yuan, which has obvious economic benefits.

[0047] In an embodiment, the base station scheduling scheme is iterated for multiple rounds in the base station scheduling process; in each round of iteration, portable base stations are dispatched from base station parking points to to-be-optimized cells according to the current base station scheduling scheme, and the base station transportation distance, the remaining amount of parking point base stations, and the bandwidth redundancy of the base station scheduling scheme are calculated to adjust the base station scheduling scheme and enter the next round of iteration until the iteration termination condition is met, and the final base station scheduling scheme is output, including:

[0048] Obtaining initial information; the initial information includes all to-be-optimized cells and their initial bandwidth and bandwidth demand, all base station parking points and their base station storage quantity, and the distance between the to-be-optimized cells and the base station parking points;

[0049] According to the initial information, a to-be-optimized cell group, an initial bandwidth row vector, a bandwidth demand row vector, a base station parking point group, a parking point base station storage quantity row vector, and a distance matrix used for representing distances between the to-be-optimized cells and the base station parking points are constructed as basic information of a mathematical model;

[0050] A target function is constructed for the mathematical model, the target function being a sum of transportation distances of each scheduled portable base station, the transportation distance referring to a distance between a base station parking point where the portable base station is located and a to-be-optimized cell to which the portable base station is scheduled;

[0051] A quantity constraint condition and a bandwidth constraint condition are constructed for the mathematical model, the quantity constraint condition being that a quantity of scheduled portable base stations in the base station parking point cannot exceed a quantity of portable base stations originally parked in the base station parking point, and the bandwidth constraint condition being that a bandwidth of the to-be-optimized cell after base station scheduling is greater than or equal to a bandwidth demand of the to-be-optimized cell;

[0052] Under the quantity constraint condition and the bandwidth constraint condition, the target function is solved to obtain a final base station scheduling scheme, so as to schedule the portable base stations of the base station parking point to the to-be-optimized cell according to the base station scheduling scheme.

[0053] Specifically, the base station scheduling scheme is determined by mathematical modeling, and the mathematical modeling process is as follows:

[0054] M cells that need to be repaired are recorded as A={a1, a2,..., a m} (that is, a to-be-optimized cell group). After an emergency, an initial bandwidth of the jth cell is recorded as b i , a demand bandwidth to be reached is recorded as q j , initial bandwidths of the m cells form a row vector B=[b1, b2,..., b m ], and if a cell has no signal, the initial bandwidth is 0; the demand bandwidth to be reached forms a row vector Q=[q1, q2,..., q m ].

[0055] Meanwhile, n base station parking points that can be mobilized in the whole area are recorded as P={p1, p2,..., p n} (that is, a base station parking point group), quantities of portable base stations of the n base station parking points are recorded as K=[k1, k2,..., k n} (that is, a parking point base station storage quantity row vector); meanwhile, a driving distance of the jth cell to the ith parking point is recorded as d ij , distances of the jth cell to all parking points form a column vector matrix D j =[d 1j , d 2j..., d nj ] T The distance from each cell to different parking points forms a matrix D = [D1, D2,..., D m ](i.e., a distance matrix).

[0056] In order to make the emergency communication of the parking points meet the bandwidth demand of all cells, portable base stations are dispatched from P to A, and the following constraints need to be met:

[0057] (1) Vehicle number constraint, i.e., the number of vehicles dispatched by each parking point should be less than or equal to the number of vehicles of the parking point; it is worth noting that the vehicle number constraint is a quantity constraint condition, one vehicle corresponds to one portable base station, and the number of portable base stations dispatched in the parking point cannot exceed the number of portable base stations originally parked in the parking point.

[0058] (2) Bandwidth demand constraint, i.e., the actual bandwidth of each cell after scheduling should be higher than the demand bandwidth;

[0059] Let the number of portable base stations dispatched from the ith parking point to the jth cell be c ij , and the portable base stations dispatched from all parking points to all cells are denoted as a matrix M, which is represented as:

[0060]

[0061] The present application provides that, in a matrix X and Y with the same number of rows and columns, if each element in X is greater than or equal to the element at the corresponding position in Y, it is denoted as X ≥ Y. At the same time, it is provided that the bandwidth of any portable base station in the present application is equal and is b, and a row vector matrix containing n 1s is denoted as I.

[0062] Therefore, according to requirement (1), the constraint can be represented as:

[0063] M·I T ≤K

[0064] According to requirement (2), the constraint can be represented as:

[0065] B + S·M ≥ Q

[0066] Where S represents a row vector matrix containing n b (in this embodiment, it is assumed that the bandwidth of any portable base station is equal and is b), on the basis of meeting the two basic service constraints, considering the need for speed in response to sudden situations to be as fast as possible, the total distance of each dispatched portable base station in this scheduling is calculated:

[0067]

[0068] Let d xMinimizing the objective function, under the quantity constraint and the bandwidth constraint, to obtain the final base station scheduling scheme.

[0069] In one embodiment, the base station scheduling scheme is iterated for multiple rounds in the base station scheduling process; in each round of iteration, the portable base stations are scheduled from the base station parking points to the to-be-optimized cells according to the current base station scheduling scheme, and the base station transportation distance, the base station parking point remaining amount and the bandwidth redundancy of the base station scheduling scheme are calculated to adjust the base station scheduling scheme and enter the next round of iteration until the iteration termination condition is met, and the final base station scheduling scheme is output, including:

[0070] The to-be-optimized cell group, the initial bandwidth row vector, the bandwidth demand row vector, the base station parking point group, the base station parking point storage quantity row vector and the distance matrix for representing the distance between the to-be-optimized cells and the base station parking points are constructed according to the initial information, to serve as the basic information of the mathematical model;

[0071] The first loss function, the second loss function and the third loss function are constructed for the mathematical model; wherein the first loss function is equal to the difference between the base station parking point storage quantity row vector and the parking point base station scheduling quantity row vector of the current base station scheduling scheme, the parking point base station scheduling quantity row vector representing the number of portable base stations scheduled by each base station parking point; the second loss function is equal to the initial bandwidth row vector plus the bandwidth increment row vector of the current base station scheduling scheme minus the bandwidth demand row vector, the bandwidth increment row vector representing the bandwidth increment of each to-be-optimized cell after base station scheduling; the third loss function is equal to the negative of the sum of the transportation distances of each scheduled portable base station in the current base station scheduling scheme, the transportation distance referring to the distance between the base station parking point where the portable base station is located and the to-be-optimized cell where the portable base station is scheduled to;

[0072] The base station scheduling scheme is iterated, and the first loss function, the second loss function and the third loss function after each round of iteration are calculated;

[0073] The base station scheduling scheme of the iteration round that maximizes the total loss function is selected as the final base station scheduling scheme for base station scheduling, wherein the total loss function is in a positive correlation with the first loss function, the second loss function and the third loss function respectively.

[0074] Specifically, first, the basic information of the mathematical model is constructed (for specific examples, refer to the above embodiments, which are not repeated here), and then the loss function is established for the mathematical model, and the loss function construction process is as follows:

[0075] From the modeling process in the last embodiment, it can be seen that the scheduling process needs to perfect two constraint satisfaction problems and an optimization problem, which constitute multi-objective tasks, and each task can give a loss function according to mathematical modeling.

[0076] For the vehicle number constraint, a first loss function can be given:

[0077] L c (M) = K - M I T

[0078] Where L c (M) needs to be maximized, and the minimum value needs to be greater than or equal to 0.

[0079] For the bandwidth demand constraint, a second loss function can be given:

[0080] L b (M) = B + βI M - Q

[0081] Where β represents the bandwidth that each portable base station can provide; L b (M) needs to be maximized, and the minimum value needs to be greater than or equal to 0.

[0082] For the total distance optimization, a third loss function can be given:

[0083] L d (M) = -d x

[0084] Where L d (M) needs to be maximized.

[0085] In the model training process, the first loss function, the second loss function and the third loss function are comprehensively adjusted to obtain the optimal base station scheduling scheme.

[0086] It is worth noting that by converting the mathematical model of each sub-problem into a loss function, the key includes the following:

[0087] (1) Vehicle number constraint loss function (first loss function): based on the characteristics of the mathematical model, the model is converted into a function that can be compared with 0. This function gives a clear convergence direction for the vehicle number constraint.

[0088] (2) Bandwidth demand constraint loss function (second loss function): based on the characteristics of the mathematical model, the model is converted into a function that can be compared with 0. This function gives a clear convergence direction for the bandwidth demand constraint.

[0089] (3) Distance target loss function (third loss function): considering that the distance target has no upper and lower limit, the distance target is taken as the opposite, as the loss function of the distance target. The scheduling spends the least distance, which represents the maximization of the function, and gives the distance target a clear convergence direction.

[0090] In an embodiment, the base station scheduling scheme is iterated for multiple rounds in the base station scheduling process; in each round of iteration, the portable base station is scheduled from the base station parking point to the cell to be optimized according to the current base station scheduling scheme, and the base station transportation distance, the base station parking point remaining amount and the bandwidth redundancy of the base station scheduling scheme are calculated to adjust the base station scheduling scheme and enter the next round of iteration until the iteration termination condition is met, and the final base station scheduling scheme is output, including:

[0091] A base station parking point is taken as an agent to construct multiple agents of reinforcement learning;

[0092] All agents are learned for multiple rounds;

[0093] In each round, all agents gradually perform actions on each cell to be optimized according to the current strategy to gradually schedule portable base stations from all base station parking points for each cell to be optimized;

[0094] The round reward is calculated after each round to adjust the strategy of the agent; wherein the round reward includes a first reward and a second reward, the first reward is calculated according to the bandwidth redundancy and the base station transportation distance, the bandwidth redundancy refers to the difference between the bandwidth of the cell to be optimized after base station scheduling and the bandwidth demand of the cell to be optimized, and the base station transportation distance refers to the sum of the transportation distances of all portable base stations scheduled by the cell to be optimized; the second reward has a positive correlation with the minimum base station remaining amount, and the minimum base station remaining amount is the minimum value of the number of remaining portable base stations in all base station parking points after completing base station scheduling;

[0095] After the reinforcement learning ends, the final base station scheduling scheme is output to schedule the portable base station from the base station parking point for the cell to be optimized.

[0096] Specifically, from the modeling process of the above embodiment, it can be known that the matrix M composed of independent variables (i.e. the base station scheduling scheme) is not completely independent of each other in each column and each row, so it is difficult to simplify the solving process by discussing the entire matrix. Therefore, in this embodiment, the calculation is divided into multiple agents, each agent represents a base station parking point, and each cell to be optimized is gradually discussed, which not only simplifies the calculation process by deconstructing the matrix, but also ensures the recursive solving characteristics of each row and each column.

[0097] Further, in each round, all agents perform actions on each to-be-optimized cell according to the current strategy to gradually schedule portable base stations for each to-be-optimized cell from all base station parking points, comprising:

[0098] In a target step of a target round, all agents perform actions according to the strategy to schedule portable base stations for a target to-be-optimized cell from all base station parking points; wherein the target round is any one of all rounds, the target step is any one of multiple steps of the target round, and the target to-be-optimized cell is a to-be-optimized cell corresponding to the target step.

[0099] A first reward is calculated for the target step, wherein the first reward of the target step is calculated according to a bandwidth redundancy of the target to-be-optimized cell and a base station transportation distance of the target to-be-optimized cell.

[0100] The strategy is adjusted according to the first reward, and the next step of the target step is entered until base station scheduling for all to-be-optimized cells is completed, and the target round is ended.

[0101] Further, the adjusting of the strategy according to the first reward comprises adjusting the strategy according to the first reward by using a Critic network (value function network).

[0102] Specifically, the matrix M composed of independent variables is not completely independent in each column and each row, so it is difficult to simplify the solving process by discussing the entire matrix. Therefore, the present embodiment divides the calculation into multiple-step decisions of multiple agents, each decision is discussed separately, which not only simplifies the calculation process by deconstructing the matrix, but also ensures the recursive solving characteristics of each row and each column.

[0103] Wherein, the matrix formed by the portable base stations scheduled by all parking points for all cells is:

[0104]

[0105] In the matrix M, c ij represents the number of portable base stations scheduled by the i-th base station parking point for the j-th to-be-optimized cell, one row represents the number of base stations scheduled by one base station parking point for each to-be-optimized cell, and one column represents the number of base stations scheduled by one to-be-optimized cell from each base station parking point.

[0106] The training process of multi-agent reinforcement learning is composed of multiple episodes, and each episode contains multiple steps, and all agents will make a decision in each step. The application proposes that multiple agents complete the scheduling process of one of the multiple to-be-optimized cells in the same step, that is, one column of data of the matrix M is improved in the same step.

[0107] Specifically, the multi-agent reinforcement learning algorithm is implemented using the multi-agent proximal policy optimization algorithm (MAPPO) to solve the multi-objective optimization problem in the scheduling process of the portable base station. The process of multi-agent multi-objective reinforcement learning is as follows:

[0108] First, consider the multi-agent reinforcement learning process, and split each loss function into the form of the reward function of each agent. On the one hand, the loss function that is not explicit or difficult to calculate can automatically consider the optimization of the target through interaction with the environment; on the other hand, thanks to the characteristics of MAPPO, the reward function can be more intuitive, updated more frequently, and converged faster than sequence learning. Therefore, we have:

[0109]

[0110] Where i represents a certain row, i is greater than or equal to 1 and less than or equal to m, j represents a certain column, j is greater than or equal to 1 and less than or equal to n, u = 1, 2, …, m, v = 1, 2, …, n. All Constitute the vehicle number constraint R c .

[0111] Second, consider the multi-objective learning process, and propose to use r d as the main detection, r b as the auxiliary detection for each step, and R c as the auxiliary detection for each episode, and give the following reward function:

[0112]

[0113] Based on the reward function, each reinforcement learning model can give the next scheduling strategy according to the current strategy, that is, predict the most likely result in the x+1th column according to the filling result of the xth column of the matrix M. Specifically, when the scheduling has not started, the decision will be made from the preset strategy s0, and a random decision a0 is executed to constitute the filling result of the 1st column, and the reward of the decision a0 is calculated. After that, the x+1th decision a x+1 will be made according to the xth strategy s x and the xth decision ax is calculated and the policy s x is determined x The final policy s x is formed after the completion of the decision a x+1 . And the evaluation value Q(s x , a x ) under the policy s x , a x is calculated by the Critic network of MAPPO, which is used to guide the prediction of the next policy in the Actor network of MAPPO. After each prediction, s x , a x , the reward r x calculated according to formula (1) is stored in the scheduling experience pool E. It can be understood that here "s" represents the number of emergency communication vehicles "obtained" by all cells, which is accurate to the number of emergency communication vehicles obtained from a certain parking point, and can be understood as the "scheduling scheme" at the present stage; and "a" represents the number of emergency communication vehicles that will be obtained by a certain cell from other parking points, which is accurate to the number of emergency communication vehicles obtained from a certain parking point, and can be understood as the "decision" of the "optimal adjustment" to the present stage scheduling scheme.

[0114] When all agents complete the scheduling of all cells, according to the satisfaction of the current vehicle number constraint, an additional feedback is given to all scheduling records stored in E in this scheduling:

[0115] r' x = r x + min{R c}

[0116] Where min{R c} represents the minimum value in R c , if the vehicle number constraint is met, that is, the number of vehicles dispatched from any base station parking point is less than the number of vehicles parked in the parking point, then min{R c}>0, r' x represents that positive feedback is obtained from the vehicle number constraint; if not, negative feedback is obtained.

[0117] At the same time, each base station parking point is assigned a separate Actor network for each agent, and each agent will give a scheduling for a cell in a step. After an episode ends, the final reward R' is calculated, and the Critic network is used to perform centralized learning on all decisions, that is, to minimize the gap between the current action policy and the next action policy:

[0118] The Critic network in the MAPPO model will sample from E during the learning process, and will perform centralized learning according to all the schedules s x , the scheduling scheme a x , and the reward r x . The loss function of the Critic network during centralized learning can be expressed as

[0119]

[0120] where L c (θ) represents the calculation to be performed using the current policy s x , the current decision a x , the next policy s x+1 , and the next decision a x+1 during learning. ∈ is a constant, and the clip function is used to limit the gradient generated by learning within the range of 1-∈ and 1+∈, thereby avoiding the huge impact of the problems of gradient explosion and gradient disappearance.

[0121] In this embodiment, 1, about multi-agent reinforcement learning. Each base station parking point is regarded as an agent, and the scheduling arrangement of all cells is given through the distributed decision of each agent. On the one hand, through the bounded rationality theory, each parking point can actively seek a solution, and determine the most likely solution through probability; on the other hand, according to the probability, the scheduling scheme is tried, which prevents falling into local optimum, and finally obtains a global optimal solution with a larger probability. At the same time, the centralized training is carried out for the arrangement, and an explicit judgment is given for each scheduling of each agent. Through the iteration process of multiple judgments, Nash equilibrium is reached between each agent, and the overall scheduling scheme reaches Pareto optimality. 2, about reward function. According to the first loss function, the second loss function and the third loss function, the scheduling problem is converted into a multi-step decision problem according to the characteristics of the multi-agent reinforcement learning loss function, and the decoupled sub-problem is further decomposed into a multi-step solution process. At the same time, the iteration process of the loss function is decomposed into each scheduling in the form of the reward function, the next scheduling can be judged according to the current scheduling, and the design and optimization of the loss function in the iteration process are simplified. Finally, the normalized reward function can prevent the reward from increasing sharply, thereby bringing devastating effects to the learning process. 3, about step feedback and episode feedback. Considering the coordination between multi-objective problems, according to the business requirements, the constraints are placed in step and episode respectively to play a feedback role. Among them, the step feedback can give an auxiliary constraint check for each iteration, if the constraint is not satisfied, the constraint loss is continuously increased on the distance loss; and if the constraint is satisfied, the constraint reward is increased on the distance loss. The episode feedback can be checked after the scheduling is completed, and then a guiding check process is given for an episode, which enriches the check mode of multiple constraints.

[0122] Compared with the prior art, the embodiment studies the characteristics of the to-be-optimized cell and the portable base station, proposes a multi-agent multi-objective scheduling mode, and according to the actual problems that need to be considered in the actual scheduling process, the actual problems are converted into constraints through the mathematical modeling method. At the same time, in view of the problem that it is difficult to consider comprehensively in a large-scale scheduling scene, through the multi-agent reinforcement learning process of centralized learning and distributed execution, the optimization process is simplified in the form of trial and error. Finally, on the basis of multiple constraints, multiple optimization targets are constructed, and the multi-objective and multi-agent are fused through the reward function, so that the model can comprehensively consider the optimization process of multiple targets, and give a scheduling result that satisfies the constraints.

[0123] Reference Figure 2 , Figure 2 is a structural schematic diagram of a base station scheduling device provided by the embodiment of the application, the device comprises:

[0124] a scheme iteration module 21, configured to perform multiple rounds of iteration on a base station scheduling scheme in a base station scheduling process;

[0125] a base station scheduling module 22, configured to, in each round of iteration, schedule portable base stations from base station parking points to to-be-optimized cells according to a current base station scheduling scheme, and calculate base station transportation distances, base station parking point residual amounts and bandwidth redundancy amounts of the base station scheduling scheme, so as to adjust the base station scheduling scheme and enter a next round of iteration, until an iteration termination condition is met, and output a final base station scheduling scheme; wherein the base station scheduling scheme refers to a scheme of scheduling portable base stations from various base station parking points to various to-be-optimized cells, and the bandwidth redundancy amount refers to a difference between a bandwidth of the to-be-optimized cell after base station scheduling and a bandwidth demand of the to-be-optimized cell.

[0126] In an embodiment, the scheme iteration module 21 is specifically configured to:

[0127] construct multiple agents of reinforcement learning with one base station parking point as one agent;

[0128] perform multiple rounds of learning on all the agents;

[0129] the base station scheduling module 22 is specifically configured to, in each round, all the agents gradually perform actions on each to-be-optimized cell according to a current strategy, so as to gradually schedule portable base stations from all the base station parking points for each to-be-optimized cell;

[0130] calculate a round reward after each round, so as to adjust the strategy of the agent; wherein the round reward includes a first reward and a second reward, the first reward is calculated according to a bandwidth redundancy amount and a base station transportation distance, the bandwidth redundancy amount refers to a difference between a bandwidth of the to-be-optimized cell after base station scheduling and a bandwidth demand of the to-be-optimized cell, and the base station transportation distance refers to a sum of transportation distances of all the portable base stations scheduled by the to-be-optimized cell; the second reward is in a positive correlation with a minimum base station residual amount, and the minimum base station residual amount is a minimum value of a number of residual portable base stations of all the base station parking points after base station scheduling is completed;

[0131] after the reinforcement learning is completed, output a final base station scheduling scheme, so as to schedule the portable base stations for the to-be-optimized cells from the base station parking points.

[0132] In an embodiment, the, in each round, all the agents gradually perform actions on each to-be-optimized cell according to a current strategy, so as to gradually schedule portable base stations from all the base station parking points for each to-be-optimized cell, includes:

[0133] In a target step of a target round, all agents perform actions according to a strategy to schedule portable base stations from all base station parking points to a target to-be-optimized cell; wherein the target round is any one of all rounds, the target step is any one of multiple steps of the target round, and the target to-be-optimized cell is a to-be-optimized cell corresponding to the target step;

[0134] A first reward is calculated for the target step, wherein the first reward of the target step is calculated according to a bandwidth redundancy of the target to-be-optimized cell and a base station transportation distance of the target to-be-optimized cell;

[0135] The strategy is adjusted according to the first reward, and a next step of the target step is entered until base station scheduling of all to-be-optimized cells is completed, and the target round is ended.

[0136] In an implementation, the adjusting of the strategy according to the first reward comprises adjusting the strategy according to the first reward by using a Critic network.

[0137] In an implementation, in the base station scheduling process, a base station scheduling scheme is iterated for multiple rounds; in each round of iteration, portable base stations are scheduled from base station parking points to to-be-optimized cells according to a current base station scheduling scheme, and a base station transportation distance, a parking point base station remaining amount, and a bandwidth redundancy of the base station scheduling scheme are calculated to adjust the base station scheduling scheme and enter a next round of iteration until an iteration termination condition is met, and a final base station scheduling scheme is output, comprising:

[0138] Initial information is obtained; the initial information comprises all to-be-optimized cells and initial bandwidths and bandwidth demands thereof, all base station parking points and base station storage amounts thereof, and distances between the to-be-optimized cells and the base station parking points;

[0139] A to-be-optimized cell group, an initial bandwidth row vector, a bandwidth demand row vector, a base station parking point group, a parking point base station storage amount row vector, and a distance matrix for representing distances between the to-be-optimized cells and the base station parking points are constructed according to the initial information to serve as basic information of a mathematical model;

[0140] A target function is constructed for the mathematical model, the target function being a sum of transportation distances of all scheduled portable base stations, the transportation distance being a distance between a base station parking point where the portable base station is located and a to-be-optimized cell to which the portable base station is scheduled to be dispatched;

[0141] construct a quantity constraint condition and a bandwidth constraint condition for the mathematical model, the quantity constraint condition being that a number of scheduled portable base stations in the base station parking point cannot exceed a number of portable base stations originally parked in the base station parking point, and the bandwidth constraint condition being that a bandwidth of the to-be-optimized cell after base station scheduling is greater than or equal to a bandwidth demand of the to-be-optimized cell;

[0142] solving the objective function under the quantity constraint condition and the bandwidth constraint condition to obtain a final base station scheduling scheme, and scheduling the portable base stations of the base station parking point to the to-be-optimized cell according to the base station scheduling scheme.

[0143] In an implementation, in the base station scheduling process, the base station scheduling scheme is iterated for multiple rounds; in each round of iteration, the portable base stations are scheduled from the base station parking point to the to-be-optimized cell according to a current base station scheduling scheme, and a base station transportation distance, a parking point base station remaining amount, and a bandwidth redundancy amount of the base station scheduling scheme are calculated to adjust the base station scheduling scheme and enter a next round of iteration, until an iteration termination condition is met, and a final base station scheduling scheme is output, including:

[0144] constructing a to-be-optimized cell group, an initial bandwidth row vector, a bandwidth demand row vector, a base station parking point group, a parking point base station storage quantity row vector, and a distance matrix for representing distances between the to-be-optimized cells and the base station parking points according to the initial information, as basic information of the mathematical model;

[0145] constructing a first loss function, a second loss function, and a third loss function for the mathematical model; the first loss function is equal to a difference between the parking point base station storage quantity row vector and a parking point base station scheduling quantity row vector of a current base station scheduling scheme, the parking point base station scheduling quantity row vector representing a number of scheduled portable base stations of each base station parking point; the second loss function is equal to the initial bandwidth row vector plus a bandwidth increment row vector of the current base station scheduling scheme minus the bandwidth demand row vector, the bandwidth increment row vector representing bandwidth increments of each to-be-optimized cell after base station scheduling; and the third loss function is equal to a sum of transportation distances of each scheduled portable base station in the current base station scheduling scheme in reverse, the transportation distance referring to a distance between a base station parking point where the portable base station is located and a to-be-optimized cell where the portable base station is scheduled to;

[0146] iterating the base station scheduling scheme, and calculating the first loss function, the second loss function, and the third loss function after each round of iteration;

[0147] The base station scheduling scheme of an iteration round that maximizes the total loss function is selected as a final base station scheduling scheme for base station scheduling, wherein the total loss function is in positive correlation with the first loss function, the second loss function and the third loss function respectively.

[0148] It should be noted that the working principle of the device provided in the above embodiments can refer to the working process of the base station scheduling method provided in any of the above embodiments, which will not be repeated here.

[0149] Compared with the prior art, the device provided in the embodiments of the present application performs multiple rounds of iteration on the base station scheduling scheme in the base station scheduling process, calculates the base station transportation distance, the remaining amount of the base station parking point and the bandwidth redundancy of the base station scheduling scheme in each round of iteration for adjusting the base station scheduling scheme and entering the next round of iteration until the iteration termination condition is met, and outputs the final base station scheduling scheme; wherein the base station scheduling scheme refers to the scheme of scheduling portable base stations from various base station parking points to various to-be-optimized cells, and the bandwidth redundancy refers to the difference between the bandwidth of the to-be-optimized cell after base station scheduling and the bandwidth demand of the to-be-optimized cell, which fully considers the bandwidth demand of each to-be-optimized cell, the performance of each portable base station and the distance between the portable base station and the to-be-optimized cell, so that the to-be-optimized cell can be reasonably allocated to a suitable base station, and the accuracy of base station allocation is improved.

[0150] Referring to Figure 3 , the embodiments of the present application also provide a base station scheduling device, which comprises a processor 31, a memory 32, and a computer program stored in the memory 32 and configured to be executed by the processor 31, wherein the processor 31 implements the steps in the above base station scheduling method embodiments, such as S11-S12 in Figure 1 , when executing the computer program; or the processor 31 implements the functions of the modules in the above device embodiments when executing the computer program.

[0151] For example, the computer program can be divided into one or more modules, which are stored in the memory 32 and executed by the processor 31 to complete the present application. The one or more modules can be a series of computer program instruction segments that can complete a specific function, which are used to describe the execution process of the computer program in the base station scheduling device. For example, the computer program can be divided into multiple modules, and each module is used to execute a specific step in the method described in any of the above embodiments.

[0152] The base station scheduling device can be a desktop computer, a notebook, a palm computer, a cloud server, and the like. The base station scheduling device can include, but is not limited to, a processor 31 and a memory 32. Those skilled in the art can understand that the base station scheduling device can also include an input / output device, a network access device, a bus, and the like.

[0153] The processor 31 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, and the like. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The processor 31 is the control center of the base station scheduling device, and is connected to various parts of the base station scheduling device through various interfaces and lines.

[0154] The memory 32 can be used to store computer programs and / or modules. The processor 31 realizes various functions of the base station scheduling device by running or executing computer programs and / or modules stored in the memory 32, and calling data stored in the memory 32. The memory 32 can mainly include a program storage area and a data storage area. The program storage area can store an operating system, at least one application program required for a function (such as an image playing function, etc.), and the like. The data storage area can store data created according to the use of the mobile phone, and the like. In addition, the memory 32 can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state memory device.

[0155] If the base station scheduling device integrated module is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer-readable storage medium. When the processor 31 executes the computer program, the steps of each method embodiment described above can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer-readable medium can include any entity or device that can carry the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0156] The embodiment of the present application also provides a computer program product, including computer programs / instructions, which are executed by a processor to realize the base station scheduling method as described in any of the above embodiments.

[0157] The above is the preferred embodiment of the present application. It should be noted that for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, which are also considered within the scope of protection of the present application.

Claims

1. A base station scheduling method, characterized by, The method comprises the following steps: In the base station scheduling process, a plurality of iterations of the base station scheduling scheme are performed; In each iteration, portable base stations are dispatched from base station parking points to the to-be-optimized cells according to the current base station scheduling scheme, and the base station transportation distance, the base station parking point residual quantity and the bandwidth redundancy of the base station scheduling scheme are calculated to adjust the base station scheduling scheme and enter the next iteration until the iteration termination condition is met, and the final base station scheduling scheme is output; wherein the base station scheduling scheme refers to the scheme of dispatching portable base stations from various base station parking points to various to-be-optimized cells, the bandwidth redundancy refers to the difference between the bandwidth of the to-be-optimized cell after base station scheduling and the bandwidth demand of the to-be-optimized cell, and the adjustment direction of the base station scheduling scheme is to increase the base station parking point residual quantity, increase the bandwidth redundancy and shorten the base station transportation distance.

2. The base station scheduling method of claim 1, wherein, The method comprises the following steps: In the base station scheduling process, a plurality of iterations of the base station scheduling scheme are performed; In each iteration, portable base stations are dispatched from base station parking points to the to-be-optimized cells according to the current base station scheduling scheme, and the base station transportation distance, the base station parking point residual quantity and the bandwidth redundancy of the base station scheduling scheme are calculated to adjust the base station scheduling scheme and enter the next iteration until the iteration termination condition is met, and the final base station scheduling scheme is output; wherein the base station scheduling scheme refers to the scheme of dispatching portable base stations from various base station parking points to various to-be-optimized cells, the bandwidth redundancy refers to the difference between the bandwidth of the to-be-optimized cell after base station scheduling and the bandwidth demand of the to-be-optimized cell, and the adjustment direction of the base station scheduling scheme is to increase the base station parking point residual quantity, increase the bandwidth redundancy and shorten the base station transportation distance. The method comprises the following steps: A base station parking point is taken as an agent, and a plurality of agents of reinforcement learning are constructed; All agents are learned for a plurality of rounds; 3. The base station scheduling method of claim 2, wherein, In each round, all agents perform actions on each to-be-optimized cell according to the current strategy to gradually dispatch portable base stations from all base station parking points to each to-be-optimized cell; After each round, a round reward is calculated to adjust the strategy of the agent; wherein the round reward comprises a first reward and a second reward, the first reward is calculated according to the bandwidth redundancy and the base station transportation distance, the bandwidth redundancy refers to the difference between the bandwidth of the to-be-optimized cell after base station scheduling and the bandwidth demand of the to-be-optimized cell, and the base station transportation distance refers to the sum of the transportation distances of all portable base stations dispatched by the to-be-optimized cell; the second reward has a positive correlation with the minimum base station residual quantity, and the minimum base station residual quantity is the minimum value of the number of residual portable base stations in all base station parking points after the base station scheduling is completed; After the reinforcement learning is completed, the final base station scheduling scheme is output to dispatch the portable base stations from the base station parking points to the to-be-optimized cells. The method comprises the following steps: In each round, all agents perform actions on each to-be-optimized cell according to the current strategy to gradually dispatch portable base stations from all base station parking points to each to-be-optimized cell; In the target step of the target round, all agents perform actions according to the strategy to dispatch portable base stations from all base station parking points to the target to-be-optimized cell; wherein the target round is any one of all rounds, the target step is any one of the plurality of steps of the target round, and the target to-be-optimized cell is the to-be-optimized cell corresponding to the target step. calculating a first reward for the target step, wherein the first reward of the target step is calculated according to a bandwidth redundancy of the target cell to be optimized and a base station transportation distance of the target cell to be optimized; adjusting the strategy according to the first reward, and entering a next step of the target step until base station scheduling of all the cells to be optimized is completed, and the target round is ended.

4. The base station scheduling method of claim 3, characterized by The adjusting the strategy according to the first reward comprises adjusting the strategy according to the first reward by using a Critic network.

5. The base station scheduling method of claim 1, wherein, In the base station scheduling process, the base station scheduling scheme is iterated for multiple rounds. In each round of iteration, portable base stations are scheduled from base station parking points to the cells to be optimized according to a current base station scheduling scheme, and a base station transportation distance, a parking point base station remaining amount and a bandwidth redundancy of the base station scheduling scheme are calculated to adjust the base station scheduling scheme and enter a next round of iteration until an iteration termination condition is met, and a final base station scheduling scheme is output, comprising: obtaining initial information; the initial information comprises all the cells to be optimized and initial bandwidths and bandwidth demands thereof, all the base station parking points and base station storage amounts thereof, and distances between the cells to be optimized and the base station parking points; constructing a cell to be optimized group, an initial bandwidth row vector, a bandwidth demand row vector, a base station parking point group, a parking point base station storage amount row vector and a distance matrix for representing distances between the cells to be optimized and the base station parking points according to the initial information, to serve as basic information of a mathematical model; constructing an objective function for the mathematical model, the objective function being to minimize a sum of transportation distances of all the scheduled portable base stations, the transportation distance referring to a distance between a base station parking point where the portable base station is located and a cell to be optimized where the portable base station is scheduled to; constructing a quantity constraint condition and a bandwidth constraint condition for the mathematical model, the quantity constraint condition being that a number of the scheduled portable base stations in the base station parking point cannot exceed a number of the portable base stations originally parked in the base station parking point, and the bandwidth constraint condition being that a bandwidth of the cell to be optimized after base station scheduling is greater than or equal to a bandwidth demand of the cell to be optimized; under the quantity constraint condition and the bandwidth constraint condition, solving the objective function to obtain a final base station scheduling scheme, so as to schedule the portable base stations of the base station parking points to the cells to be optimized according to the base station scheduling scheme.

6. The base station scheduling method of claim 1, wherein, In the base station scheduling process, the base station scheduling scheme is iterated for multiple rounds. In each round of iteration, portable base stations are scheduled from base station parking points to the cells to be optimized according to a current base station scheduling scheme, and a base station transportation distance, a parking point base station remaining amount and a bandwidth redundancy of the base station scheduling scheme are calculated to adjust the base station scheduling scheme and enter a next round of iteration until an iteration termination condition is met, and a final base station scheduling scheme is output, comprising: constructing, according to the initial information, a to-be-optimized cell group, an initial bandwidth row vector, a bandwidth demand row vector, a base station parking point group, a parking point base station storage quantity row vector, and a distance matrix used to represent distances between the to-be-optimized cells and the base station parking points, as basic information of a mathematical model; constructing, for the mathematical model, a first loss function, a second loss function, and a third loss function; the first loss function is equal to a difference between the parking point base station storage quantity row vector and a parking point base station scheduling quantity row vector of a current base station scheduling scheme, the parking point base station scheduling quantity row vector representing a quantity of portable base stations scheduled by each base station parking point; the second loss function is equal to a sum of the initial bandwidth row vector, a bandwidth increment row vector of the current base station scheduling scheme, and a subtraction result of the bandwidth demand row vector, the bandwidth increment row vector representing bandwidth increments of the to-be-optimized cells after base station scheduling; and the third loss function is equal to a reciprocal of a sum of transportation distances of the portable base stations in the current base station scheduling scheme, the transportation distance being a distance between a base station parking point where the portable base station is located and a to-be-optimized cell where the portable base station is scheduled to be located; iterating the base station scheduling scheme, and calculating the first loss function, the second loss function, and the third loss function after each iteration; selecting, as a final base station scheduling scheme, a base station scheduling scheme of an iteration round in which a total loss function is maximum, to perform base station scheduling, wherein the total loss function is in a positive correlation with the first loss function, the second loss function, and the third loss function.

7. A base station scheduling device, characterized in that: comprising: a scheme iteration module, configured to perform multiple iterations on a base station scheduling scheme in a base station scheduling process; a base station scheduling module, configured to, in each iteration, schedule a portable base station from a base station parking point to a to-be-optimized cell according to a current base station scheduling scheme, and calculate a base station transportation distance, a parking point base station remaining quantity, and a bandwidth redundancy of the base station scheduling scheme, to adjust the base station scheduling scheme and enter a next iteration, until an iteration termination condition is met, and output a final base station scheduling scheme; wherein the base station scheduling scheme refers to a scheme of scheduling portable base stations from each base station parking point to each to-be-optimized cell, the bandwidth redundancy refers to a difference between a bandwidth of the to-be-optimized cell after base station scheduling and a bandwidth demand of the to-be-optimized cell, and an adjustment direction of the base station scheduling scheme is to increase the parking point base station remaining quantity, increase the bandwidth redundancy, and shorten the base station transportation distance.

8. A base station scheduling apparatus, characterized by comprising: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, when the processor executes the computer program, a base station scheduling method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium comprises a stored computer program, wherein when the computer program runs, the device where the computer readable storage medium is located is controlled to perform the base station scheduling method according to any one of claims 1 to 6.

10. A computer program product comprising computer programs / instructions, characterized in that, The computer program / instructions, when executed by the processor, implement the base station scheduling method according to any one of claims 1 to 6.

Images