CN112584386A - 5G C-RAN resource prediction and allocation method and system - Google Patents
5G C-RAN resource prediction and allocation method and system Download PDFInfo
- Publication number
- CN112584386A CN112584386A CN202011285025.2A CN202011285025A CN112584386A CN 112584386 A CN112584386 A CN 112584386A CN 202011285025 A CN202011285025 A CN 202011285025A CN 112584386 A CN112584386 A CN 112584386A
- Authority
- CN
- China
- Prior art keywords
- network
- energy consumption
- allocation
- chromosomes
- bbu
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000005265 energy consumption Methods 0.000 claims abstract description 81
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 75
- 238000013468 resource allocation Methods 0.000 claims abstract description 38
- 238000013508 migration Methods 0.000 claims abstract description 33
- 230000005012 migration Effects 0.000 claims abstract description 33
- 230000002068 genetic effect Effects 0.000 claims abstract description 26
- 238000004364 calculation method Methods 0.000 claims abstract description 23
- 238000004891 communication Methods 0.000 claims abstract description 17
- 230000008569 process Effects 0.000 claims abstract description 11
- 210000000349 chromosome Anatomy 0.000 claims description 88
- 238000005457 optimization Methods 0.000 claims description 36
- 108090000623 proteins and genes Proteins 0.000 claims description 24
- 238000012545 processing Methods 0.000 claims description 16
- 238000012549 training Methods 0.000 claims description 12
- 238000012937 correction Methods 0.000 claims description 10
- 238000012423 maintenance Methods 0.000 claims description 9
- 206010064571 Gene mutation Diseases 0.000 claims description 8
- 230000003044 adaptive effect Effects 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 4
- 238000010187 selection method Methods 0.000 claims description 3
- 230000003068 static effect Effects 0.000 claims description 3
- 239000000126 substance Substances 0.000 claims description 3
- 230000008859 change Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000004088 simulation Methods 0.000 description 4
- 238000009499 grossing Methods 0.000 description 3
- 238000013461 design Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 230000010355 oscillation Effects 0.000 description 2
- 238000005057 refrigeration Methods 0.000 description 2
- 101150052384 50 gene Proteins 0.000 description 1
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/22—Traffic simulation tools or models
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/02—Resource partitioning among network components, e.g. reuse partitioning
- H04W16/10—Dynamic resource partitioning
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/02—Arrangements for optimising operational condition
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention provides a 5G C-RAN resource prediction and allocation method and a 5G C-RAN resource prediction and allocation system, and belongs to the field of wireless communication. The method comprises the following steps: establishing a flow prediction model for a C-RAN system, and obtaining flow prediction data according to the flow prediction model; designing a network resource allocation algorithm for network resource allocation based on a genetic algorithm according to the flow prediction data; and obtaining an optimal allocation scheme of the network resources according to the network resource allocation algorithm, and dynamically allocating the network resources of the system according to the optimal allocation scheme. The operation energy consumption of the BBU is considered in the distribution process, and the calculation task migration energy consumption is also considered, so that the energy consumption of the system is saved while the service quality of the communication network is ensured to a certain extent.
Description
Technical Field
The invention relates to the field of wireless communication, in particular to a 5G C-RAN resource prediction and distribution method and a 5G C-RAN resource prediction and distribution system.
Background
With the gradual popularization and application of 5G, a radio access network C-RAN system, as a core technology of 5G, needs a more intelligent and flexible architecture and strategy to fully exploit the performance of the 5G network. Currently, the optimization of a baseband pool under a C-RAN architecture is less, and a BBU (Building Base band unit) in the baseband pool is in an active state whether idle or busy, and dynamic switching is not performed according to a load condition. BBUs that are active for long periods of time can cause significant waste of resources due to the highly unbalanced distribution of network resource demands over time and space.
Current research on methods of prediction and allocation of C-RAN resources is mainly centered on two aspects: firstly, the network load condition is predicted through a model, and secondly, the C-RAN resources are dynamically distributed according to predicted values or actual values. Most existing methods only pay attention to minimization of energy consumption, and no specific optimization scheme exists for extra energy consumption generated in a migration process of a computing task and possible interrupt risks of a communication task caused by dynamic resource allocation.
Disclosure of Invention
The invention aims to provide a 5G C-RAN resource prediction and allocation method, which at least solves the problems that most of the existing methods only concern about minimization of energy consumption, and no specific optimization scheme exists for extra energy consumption generated in a migration process of a computing task and possible interrupt risks of a communication task caused by dynamic allocation of resources.
In order to achieve the above object, a first aspect of the present invention provides a 5G C-RAN resource prediction and allocation method, the method comprising: establishing an optimization problem of comprehensive energy consumption optimization and network service quality maintenance; establishing a flow prediction model for a C-RAN system, and obtaining flow prediction data according to the flow prediction model; designing a network resource allocation algorithm for network resource allocation based on a genetic algorithm according to the flow prediction data; and obtaining an optimal allocation scheme of the network resources according to the network resource allocation algorithm, and dynamically allocating the network resources of the system according to the optimal allocation scheme.
Optionally, the establishing an optimization problem of comprehensive energy consumption optimization and maintaining network service quality includes: respectively calculating the total operating energy consumption, the total task migration energy consumption and the instantaneous power of the power communication task of the BBU; the BBU operation total energy consumption calculation formula is as follows:
wherein n is the total number of BBUs in the BBU pool; pbi(t) is the energy consumption of the ith BBU in time t, and the calculation formula is as follows:
wherein h is the total number of tasks;
is a Boolean variable; sz(t) is the task volume of the z-th task; gamma is the correction weight of the running energy consumption; pbbasic(t) is the base energy consumption when the BBU is turned on.
Optionally, the task migration total energy consumption βtotalThe formula for calculation of (t) is:
wherein h is the total number of tasks;
is a Boolean variable; s'z(t) indicates the number of tasks in performing data migrationThe size of the data volume; delta is the correction weight of the task migration energy consumption; the electric power communication task instantaneous power
The calculation formula of (2) is as follows:
wherein the content of the first and second substances,
is the processing rate of the mth task in the ith BBU; α is the power correction weight.
Optionally, a modeling relation taking the comprehensive energy consumption optimization and the maintenance of the network service quality as an optimization problem is as follows:
Min Ptotal(t)＝αtotal(t)+βtotal(t)+Pstatic(t)
wherein, Pstatic(t) is the sum of static energy consumption of the C-RAN system;
taking the comprehensive energy consumption optimization and the maintenance of the network service quality as optimization problems to meet the following constraint conditions:
γ,δ,α>0
wherein, PmaxIs the rated maximum power of the BBU; m represents that the task is the mth task in the BBU; m is the task total number of the BBU at the moment;
the downlink network rate of the mth task in the ith BBU is obtained; cMaxIs the rated maximum value of the BBU downlink network speed.
Optionally, the establishing a traffic prediction model for the C-RAN system includes: acquiring flow data; preprocessing the flow data according to an arithmetic moving average algorithm, and taking the data average value of the position of a sliding window as the numerical value of the position; wherein the arithmetic moving average algorithm comprises: setting the size of a sliding window; calculating the total data volume in the sliding window; obtaining the data average value according to the total data amount and the sliding window size; wherein, the calculation formula is:
wherein MA is the average value obtained finally; t isiIs flow data; j is the starting position of the window; w is the window width.
Optionally, the establishing a traffic prediction model for the C-RAN system further includes: constructing an LSTM flow prediction model and finishing the training of the LSTM flow prediction model; predicting according to the LSTM flow prediction model which is trained and the obtained average value, and obtaining a network flow data prediction value; wherein, the loss function of the LSTM flow pre-model is MAE average absolute error, and the calculation formula is as follows:
wherein, yiIs a predicted value; x is the number ofiIs the actual value.
Optionally, the network resource allocation algorithm based on the genetic algorithm is a group evolution process simulation algorithm; wherein, one chromosome represents the migration scheme of one RRH, and the genes on the chromosome represent BBUs to be migrated next time by the RRH at the corresponding position.
Optionally, the obtaining an optimal allocation scheme of network resources according to the network resource allocation algorithm, and performing dynamic allocation of system network resources according to the optimal allocation scheme includes: 1) completing chromosome coding according to a genetic algorithm, and generating a certain number of chromosomes after the coding is completed, wherein genes in the chromosomes are randomly generated, and the chromosomes are used as a first generation population to begin to evolve; 2) screening the adaptive performance of the chromosomes in the first generation population according to a tournament selection method, randomly selecting two chromosomes in the first generation population each time for adaptive comparison, selecting one chromosome with stronger adaptability as a progeny population until all the chromosomes are screened, and reserving half of the chromosomes in the original population with stronger adaptability as the progeny population; 3) selecting two chromosomes in the offspring population each time, performing gene random part exchange of the two chromosomes until all the chromosomes are crossed, and selecting all the crossed chromosomes into the offspring population, wherein the offspring population comprises the crossed chromosomes and the chromosomes subjected to adaptive selection; 4) presetting the gene mutation probability, expanding the number of chromosomes of the offspring population through gene mutation, selecting the migration scheme of the RRH represented by the chromosome with the strongest adaptability in the offspring population after the number expansion as the optimal allocation scheme of the network resources, and dynamically allocating the network resources according to the optimal allocation scheme.
A second aspect of the present invention provides a 5G C-RAN resource prediction and allocation system, the system comprising: the acquisition unit is used for acquiring flow data; the processing unit is used for establishing an optimization problem of comprehensive energy consumption optimization and network service quality maintenance; the prediction unit is used for establishing a flow prediction model aiming at the C-RAN system and obtaining flow prediction data according to the flow prediction model; an allocation unit for designing a network resource allocation algorithm based on a genetic algorithm for network resource allocation according to the traffic prediction data; and obtaining an optimal allocation scheme of the network resources according to the network resource allocation algorithm, and dynamically allocating the network resources of the system according to the optimal allocation scheme.
In another aspect, the present invention provides a computer readable storage medium having instructions stored thereon, which when executed on a computer, cause the computer to perform the 5G C-RAN resource prediction and allocation method described above.
According to the technical scheme, firstly, the problem of saving system energy consumption while ensuring the service quality of the communication network is analyzed and modeled, and all factors needing to be considered in the process of optimizing and maintaining the network service quality, including the maximum power of the BBU, the relation between the processing rate of a calculation task and the corresponding power, the relation between the downlink rate of each task and the processing rate and the like, are integrated; then designing a flow prediction model aiming at the C-RAN system, and taking prediction data obtained according to the model as a basis for resource allocation; a network resource allocation algorithm based on a genetic algorithm for network resource allocation is designed, and data obtained by prediction and system network resources are dynamically allocated. The operation energy consumption of the BBU is considered in the distribution process, and the calculation task migration energy consumption is also considered, so that the energy consumption of the system is saved while the service quality of the communication network is ensured to a certain extent.
Additional features and advantages of embodiments of the invention will be set forth in the detailed description which follows.
Drawings
The accompanying drawings, which are included to provide a further understanding of the embodiments of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the embodiments of the invention without limiting the embodiments of the invention. In the drawings:
fig. 1 is a flowchart of a method for 5G C-RAN resource prediction and allocation provided by an embodiment of the present invention;
FIG. 2 is a flow chart of a method for modeling an optimization problem according to an embodiment of the present invention
Fig. 3 is a flowchart of a method for modeling a traffic prediction model of a C-RAN system according to an embodiment of the present invention;
FIG. 4 is a flow chart of a method for obtaining an optimization scheme provided by one embodiment of the present invention;
fig. 5 is a system block diagram of a 5G C-RAN resource prediction and allocation system provided in one embodiment of the present invention;
FIG. 6 is a graph of network traffic prediction model learning training provided by an embodiment of the present invention;
FIG. 7 is a graph comparing a predicted curve and an actual network traffic curve provided by one embodiment of the present invention;
FIG. 8 is an iterative graph of network resource allocation algorithm population fitness for a genetic algorithm provided in accordance with an embodiment of the present invention;
FIG. 9 is a comparison graph of BBU active numbers provided by one embodiment of the present invention;
FIG. 10 is a graph comparing UE migration times according to an embodiment of the present invention;
fig. 11 is a comparison diagram of total energy consumption of a C-RAN system provided in an embodiment of the present invention.
Description of the reference numerals
10-an acquisition unit; 20-a processing unit; 30-a prediction unit; 40-allocation unit.
Detailed Description
The following detailed description of embodiments of the invention refers to the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating the present invention, are given by way of illustration and explanation only, not limitation.
Fig. 5 is a system architecture diagram of a 5G C-RAN resource prediction and allocation system according to an embodiment of the present invention, the system including: the acquisition unit 10 is used for acquiring flow data; a processing unit 20, configured to model the comprehensive energy consumption optimization and the maintenance of network service quality as an optimization problem; a prediction unit 30 configured to design a traffic prediction model for the C-RAN system, and obtain traffic prediction data according to the traffic prediction model; an allocation unit 40 that designs a network resource allocation algorithm based on a genetic algorithm for network resource allocation from the traffic prediction data; and solving the optimization problem modeling according to the network resource allocation algorithm, and performing dynamic allocation of system network resources.
Fig. 1 is a flowchart of a method for predicting and allocating 5G C-RAN resources according to an embodiment of the present invention. As shown in fig. 1, an embodiment of the present invention provides a 5G C-RAN resource prediction and allocation method, where the method includes:
step S10: and modeling by taking the comprehensive energy consumption optimization and the maintenance of the network service quality as optimization problems.
Specifically, in order to realize the process of resource allocation, the operation energy consumption of the BBU is considered, and the calculation task migration energy consumption is also considered, so that the communication network service quality is guaranteed to a certain extent, the system energy consumption is saved, and various factors including the maximum power of the BBU, the relationship between the processing rate of the calculation task and the corresponding power, the relationship between the downlink rate of each task and the processing rate and the like are considered in the process of comprehensively optimizing the energy consumption and maintaining the network service quality. Therefore, when modeling optimization problems, the total energy consumption of BBU operation, the total energy consumption of task migration and the instantaneous power of power communication tasks need to be comprehensively considered. Specifically, as shown in fig. 2, the method includes the following steps:
step S101: calculating total energy consumption alpha of BBU operationtotal(t)。
Specifically, the calculation formula is as follows:
wherein n is the total number of BBUs in the BBU pool; pbi(t) is the energy consumption of the ith BBU over time t, which is calculated as:
wherein h is the total number of tasks;
in order to be a boolean variable, the method comprises the following steps,
it is stated that at time t, the z-th task is assigned to the i-th BBU; sz(t) is the task volume of the z-th task; gamma is the correction weight of the running energy consumption; pbbasic(t) is the base energy consumption when the BBU is turned on, and the part of energy consumption is fixed and is not generated when the BBU is turned off.
Step S102: computing task migration Total energy consumption βtotal(t)。
Specifically, the calculation formula is as follows:
wherein h is the total number of tasks;
in order to be a boolean variable, the method comprises the following steps,
indicating that the z task is migrated from the ith BBU to the jth BBU at time t; s'z(t) represents the data size of the task when data migration is performed; and delta is the correction weight of the task migration energy consumption.
Step S103: calculating instantaneous power of electric power communication task
Specifically, the calculation formula is as follows:
wherein the content of the first and second substances,
for the processing rate of the mth task in the ith BBU, according to the relevant reference, the instantaneous power of the task and the corresponding processing rate are in a linear relation; α is the power correction weight.
Step S104: an optimization objective is obtained.
Specifically, modeling is performed according to the problem of generating the optimization problem by calculating the total energy consumption of BBU operation, the total energy consumption of task migration and the instantaneous power of the power communication task, and the relational expression is as follows:
Min Ptotal(t)＝αtotal(t)+βtotal(t)+Pstatic(t)
wherein, PstaticAnd (t) is the sum of static energy consumption of the system, such as air conditioner refrigeration energy consumption, fronthaul link energy consumption and the like. The above relation satisfies partial constraint conditions, including:
γ,δ,α>0 (1)
wherein, the formula (1) is a constraint on the correction weight, that is, γ, δ, α are all positive values.
The formula (2) is the constraint to the BBU instantaneous power, which means that in the ith BBU, the sum of the power consumed by all tasks is not more than the rated maximum power of the BBU, M means that the task is the mth task in the BBU, M is the total number of the tasks at this moment of the BBU, PmaxIs the nominal maximum power of the BBU.
The formula (3) is the constraint to the BBU downlink rate, which means that in the ith BBU, the sum of the downlink rates of all tasks is not greater than the rated maximum downlink rate of the BBU, M means that the task is the mth task in the BBU, M is the total number of the tasks of the BBU at the moment,
is the m task in the ith BBULine network rate, CMaxIs the rated maximum value of the BBU downlink network speed.
Equation (4) is a constraint on the processing speed of the mth task in the ith BBU and the downstream network speed, and the two values should be approximately equal, i.e. the two values should be approximately equal
To satisfy the quality of service of the user in the downlink data reception.
Step S20: and designing a flow prediction model aiming at the C-RAN system, and obtaining flow prediction data according to the flow prediction model.
Specifically, in order to meet the demand for predicting the flow of the power communication network of the C-RAN architecture, the patent provides an access station flow prediction algorithm based on an LSTM model. The method learns the change rule of the time sequence network flow data by constructing the LSTM model, inputs the real-time sequence data of the network after the model training is finished, and obtains the prediction data as the basis of the network resource allocation. Specifically, as shown in fig. 3, the method includes the following steps:
step S201: and preprocessing the network flow data.
Specifically, the BBU often includes a plurality of key resources, such as baseband usage, memory resource usage, network bandwidth usage, and the like, and the patent uses network traffic as a load index for prediction. The load index data of the base station can be seen as a time sequence, has complex nonlinear characteristics, comprises large-span oscillation with change comparison rule and small-span oscillation with random change and burst, and has certain self-similarity characteristic. Because the traffic data has a strong burstiness, it will exhibit irregular fluctuation in a short time interval, and therefore, to accurately reflect the actual situation of the network traffic data, the original network data needs to be preprocessed first. One important item in the preprocessing of network data is data smoothing, and the smoothing of data has many different ways, and the different ways also have advantages and disadvantages. Preferably, a simpler and commonly used arithmetic moving average is adopted, and the algorithm adds the data in the sliding window to average by setting the size of the sliding window, and replaces the value of the position with the average value so as to achieve the purpose of data smoothing. The formula of the algorithm is as follows:
wherein MA is the average value obtained finally; t isiIs flow data; j is the starting position of the window; w is the window width.
Step S202: and constructing an LSTM flow prediction model.
Specifically, preferably, the LSTM layer time step of the LSTM model-based network traffic prediction model is set to 40, the LSTM layer is connected through the full-connection layer MLP, Adam is used as an optimizer, and the loss function uses mae (mean Absolute error) mean Absolute error. The MAE can better reflect the actual situation of the error of the predicted value, and the calculation formula is as follows:
wherein, yiIs a predicted value; x is the number ofiIs the actual value.
Step S203: and training the constructed flow prediction model.
Specifically, in one possible implementation, the data set includes a total of 1345 pieces of network traffic data, the time interval between each collected data point is 5 minutes, and each piece of data includes a total of 40 pieces of time-series network traffic data. The data set is divided into two parts, 1145 pieces of data are used in the training data set, 200 pieces of data are used in the testing data set, the training parameter batch size is set to be 64, the training turn is set to be 120, and 15% of data in the training set are used as a verification set in the training process.
And step S30, designing a network resource allocation algorithm based on a genetic algorithm for network resource allocation according to the flow prediction data.
Specifically, after the prediction data of the C-RAN network traffic is acquired through the access site traffic prediction algorithm based on the LSTM model, the problem of network resource allocation needs to be solved, so the application provides a network resource allocation algorithm based on a genetic algorithm on this basis. The algorithm simulates the evolution process of a population, a solution is represented by a chromosome, genes on the chromosome represent BBUs to be migrated next time by the RRH at the corresponding position, and if the genes are equal to the current value, the migration operation is not performed. Before the stopping condition is reached, the operations of selection, crossing and mutation are repeatedly carried out to obtain the chromosome individuals with better fitness. The obtained individual with the best fitness is the optimal solution of the migration.
Step S40: and solving the optimization problem modeling according to the network resource allocation algorithm, and performing dynamic allocation of system network resources. Specifically, as shown in fig. 4, the method includes the following steps:
step S401: carrying out chromosome coding.
Specifically, in the C-RAN architecture, the RRHs can continuously migrate among different BBUs, so that the loads of the BBUs are in a more reasonable state, and a shutdown operation can be performed on a BBU with a lower load, so as to save the overall energy consumption of the BBU pool. The genetic algorithm is used for firstly completing the coding of a chromosome, assuming that the number of RRHs in the system is m, the number of BBUs is n, namely, the chromosome has m genes, each gene has n different choices, the genes at different positions represent different RRHs, the numerical value of the gene represents the BBU to which the RRH is about to migrate, and each chromosome is a scheme for migrating one RRH. After the encoding is completed, a certain number of chromosomes are generated, genes in the chromosomes are all generated randomly, and the chromosomes begin to evolve as a first generation population.
Step S402: and (5) carrying out chromosome adaptability screening.
Specifically, the total energy consumption of each chromosome, that is, the fitness of the scheme, may be calculated by the optimization problem model of step S10, including BBU baseband processing energy consumption of the dynamic part, task migration energy consumption, and BBU pool supporting facility energy consumption of the fixed part, such as energy consumption of refrigeration equipment such as air conditioners. A chromosome with higher total energy consumption will have a lower fitness, i.e. will be less suitable for the external environment. In the selection step of the genetic algorithm, the selection strategy adopted is a championship selection method, two chromosomes are randomly selected from a population each time, the fitness of the chromosomes is compared, wherein, the chromosome with better environmental fitness, namely the chromosome with lower total energy consumption, is selected to enter the offspring population, the chromosome with poorer fitness is eliminated, the operations are repeated until each individual in the population is selected, at the moment, the population quantity is half of the original population quantity, for example, 100 chromosomes are obtained after coding, two chromosomes are randomly selected each time for adaptability comparison, one chromosome with better adaptability is screened out and put into a progeny population, the other chromosome is directly eliminated, screening is carried out for 50 times according to the rule, 50 chromosomes with better adaptability are selected as offspring populations, and the other 50 chromosomes are eliminated.
Step S403: chromosome crossing is performed.
Specifically, chromosome crossing is a part of genes of two chromosomes which are mutually exchanged, and because chromosomes in the crossing stage are subjected to selection operation, the fitness of each chromosome is higher, and the corresponding total energy consumption is smaller, so that chromosome crossing can possibly obtain a better solution. In this stage, two chromosomes are randomly selected, a random number is generated, and the genes behind the random numbers of the two chromosomes are all exchanged. Because the population quantity is reduced to half of the original population quantity in the selection stage, not only the chromosomes after crossing are added into the offspring population, but also the chromosomes before crossing are added into the offspring population, namely, the chromosomes of the parent generation are reserved, and at the moment, the total quantity of the population is restored to the original scale. For example, the offspring population includes 50 chromosomes, two chromosomes are selected for gene exchange to generate a random number, for example, 5, then all genes behind the gene No. 5 of the two chromosomes are exchanged to obtain two gene exchanged chromosomes, and according to the rule, 25 times of gene exchange are performed to obtain 50 gene exchanged genes. The 50 genes after gene exchange are brought into the offspring population together, and the number of chromosomes of the offspring population returns to 100.
Step S404: carrying out chromosome variation.
Specifically, after the chromosomes are crossed, gene mutation occurs with a certain probability, whether the chromosome is mutated or not is determined by setting the probability of the gene mutation, and if the chromosome is mutated, two random numbers are generated at the moment to respectively determine the specific position where the gene mutation occurs and the value after the gene mutation occurs. Through variation operation, the situation that the population converges on the local optimal solution can be avoided, the diversity of population genes is maintained to a certain extent, and the algorithm can fully search the optimal solution. And acquiring a chromosome with the best adaptability, taking the migration scheme of the RRH represented by the chromosome as an optimal scheme, and dynamically allocating network resources according to the optimal scheme.
In a possible implementation manner, the data set provided in step S203 is used to perform network traffic prediction to perform a simulation experiment, a Keras framework is used to build a neural network model, training is performed on the neural network model by using training set data, and an obtained learning curve is shown in fig. 6. The model is used for predicting network traffic data, an obtained prediction curve is compared with an actual network traffic curve, the comparison result is shown in figure 7, and the comparison result is analyzed, so that the prediction curve obtained by the access site traffic prediction algorithm based on the LSTM model has better fitting degree with the actual network traffic curve, and the actual network traffic condition and the change trend can be reflected to a certain degree.
In another possible implementation, a Python language is used to perform a simulation experiment of a network resource allocation algorithm based on a genetic algorithm, and compared with the existing algorithm, and the simulation parameters used are shown in table 1.
TABLE 1C-RAN System simulation parameters Table
Fig. 8 is a population fitness iteration diagram of the present algorithm, where the fitness of each chromosome is represented by the total power consumption of the system, and the higher the total power consumption is, the lower the fitness of the corresponding chromosome is, the more easily the chromosome is eliminated, so that chromosomes with lower total power consumption are left, that is, the solution of the problem is obtained through stepwise iteration. As shown in fig. 9, for a comparison graph of BBU active numbers, the network resource allocation algorithm based on the genetic algorithm and the existing algorithm proposed herein are used respectively to optimize the same network environment, and it can be seen that, when the total traffic is low, the number of BBUs used by the genetic algorithm is higher than that of the existing algorithm, but as the traffic increases, the number of BBUs used by the genetic algorithm and the existing algorithm is closer to each other. The method is characterized in that the existing algorithm does not consider whether the load is at a reasonable level or not and also does not consider calculation task migration energy consumption, and the genetic algorithm can intelligently restrict the loads of all BBUs within a reasonable interval, so that the phenomenon that when the traffic is increased, the BBUs are restarted for many times, the network service quality is influenced, and the communication task is interrupted is avoided. As shown in fig. 10, a graph comparing the number of times of UE (user equipment) migration is shown, where the UE is a user terminal equipment, is accessed through an RRH, and is finally processed in a BBU. It can be seen that the UE migration times optimized by using the genetic algorithm are small and change smoothly, while the migration times of the UE in the existing algorithm are large and fluctuate sharply. This is because the existing algorithm only considers the processing rate of the BBU baseband minimization, but does not consider the energy consumption caused by task migration, and does not consider the problems of network service quality degradation and the like that may be caused by frequent computation task migration, and the network resource allocation algorithm based on the genetic algorithm proposed herein well considers this point. As shown in fig. 11, which is a comparison graph of total energy consumption of a C-RAN system, the total energy consumption of the system includes BBU baseband processing energy consumption and task migration energy consumption, and it can be seen from the graph that two curves are closer when the total energy consumption is lower, and when the total energy consumption of the system increases due to an increase in calculation tasks, the total energy consumption of the genetic algorithm has a smaller variation range, while the total energy consumption of the existing algorithm has a larger variation range. The total energy consumption curve of the existing algorithm is generally positioned on the curve of the genetic algorithm, so that a conclusion can be drawn, and compared with the existing algorithm, the network resource allocation algorithm based on the genetic algorithm, which is provided by the invention, can reduce the energy consumption of a C-RAN system while maintaining the network service quality.
Embodiments of the present invention also provide a computer-readable storage medium having instructions stored thereon, which when executed on a computer, cause the computer to perform the 5G C-RAN resource prediction and allocation method described above.
Those skilled in the art will appreciate that all or part of the steps in the method for implementing the above embodiments may be implemented by a program, which is stored in a storage medium and includes several instructions to enable a single chip, a chip, or a processor (processor) to execute all or part of the steps in the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
While the embodiments of the present invention have been described in detail with reference to the accompanying drawings, the embodiments of the present invention are not limited to the details of the above embodiments, and various simple modifications can be made to the technical solution of the embodiments of the present invention within the technical idea of the embodiments of the present invention, and the simple modifications are within the scope of the embodiments of the present invention. It should be noted that the various features described in the above embodiments may be combined in any suitable manner without departing from the scope of the invention. In order to avoid unnecessary repetition, the embodiments of the present invention will not be described separately for the various possible combinations.
In addition, any combination of the various embodiments of the present invention is also possible, and the same should be considered as disclosed in the embodiments of the present invention as long as it does not depart from the spirit of the embodiments of the present invention.
Claims (10)
1. A 5G C-RAN resource prediction and allocation method, the method comprising:
establishing an optimization problem of comprehensive energy consumption optimization and network service quality maintenance;
establishing a flow prediction model for a C-RAN system, and obtaining flow prediction data according to the flow prediction model;
designing a network resource allocation algorithm for network resource allocation based on a genetic algorithm according to the flow prediction data;
and solving the optimization problem according to the network resource allocation algorithm to obtain an optimal allocation scheme of the network resources, and dynamically allocating the system network resources according to the optimal allocation scheme.
2. The 5G C-RAN resource prediction and allocation method of claim 1, wherein establishing an optimization problem that integrates energy consumption optimization and maintaining network quality of service comprises:
respectively calculating the total operating energy consumption, the total task migration energy consumption and the instantaneous power of the power communication task of the BBU; wherein, the BBU runs the total energy consumption alphatotal(t) the calculation formula is:
wherein n is the total number of BBUs in the BBU pool;
Pbi(t) is the energy consumption of the ith BBU in time t, and the calculation formula is as follows:
wherein h is the total number of tasks;
Sz(t) is the task volume of the z-th task;
gamma is the correction weight of the running energy consumption;
Pbbasic(t) is the base energy consumption when the BBU is turned on.
3. The 5G C-RAN resource prediction and allocation method of claim 2, wherein the task migration total energy consumption βtotalThe formula for calculation of (t) is:
wherein h is the total number of tasks;
S′z(t) represents the data size of the task when data migration is performed;
delta is the correction weight of the task migration energy consumption;
the electric power communication task instantaneous power
The calculation formula of (2) is as follows:
wherein the content of the first and second substances,
is the processing rate of the mth task in the ith BBU;
α is the power correction weight.
4. The 5G C-RAN resource prediction and allocation method of claim 3, wherein the relation between the integrated energy consumption optimization and the maintaining network quality of service optimization problem is:
Min Ptotal(t)＝αtotal(t)+βtotal(t)+Pstatic(t)
wherein, Pstatic(t) is the sum of static energy consumption of the C-RAN system;
taking the comprehensive energy consumption optimization and the maintenance of the network service quality as optimization problems to meet the following constraint conditions:
γ,δ,α>0
wherein, PmaxIs the rated maximum power of the BBU;
m represents that the task is the mth task in the BBU;
m is the task total number of the BBU at the moment;
CMaxis the rated maximum value of the BBU downlink network speed.
5. The 5G C-RAN resource prediction and allocation method of claim 1, wherein the establishing a traffic prediction model for a C-RAN system comprises:
acquiring flow data;
preprocessing the flow data according to an arithmetic moving average algorithm, and taking the data average value of the position of a sliding window as the numerical value of the position; wherein the arithmetic moving average algorithm comprises:
setting the size of a sliding window;
calculating the total data volume in the sliding window;
obtaining the data average value according to the total data amount and the sliding window size; wherein, the calculation formula is:
wherein MA is the average value obtained finally;
Tiis flow data;
j is the starting position of the window;
w is the window width.
6. The 5G C-RAN resource prediction and allocation method of claim 5, wherein the establishing a traffic prediction model for a C-RAN system further comprises:
constructing an LSTM flow prediction model and finishing the training of the LSTM flow prediction model;
predicting according to the LSTM flow prediction model which is trained and the obtained average value, and obtaining a network flow data prediction value;
wherein, the loss function of the LSTM flow pre-model is MAE average absolute error, and the calculation formula is as follows:
wherein, yiIs a predicted value;
xiis the actual value.
7. The 5G C-RAN resource prediction and allocation method of claim 1, wherein the genetic algorithm based network resource allocation algorithm is a population evolution process simulation algorithm; wherein, one chromosome represents the migration scheme of one RRH, and the genes on the chromosome represent BBUs to be migrated next time by the RRH at the corresponding position.
8. The 5G C-RAN resource prediction and allocation method of claim 7, wherein the obtaining an optimal allocation scheme of network resources according to the network resource allocation algorithm and performing dynamic allocation of system network resources according to the optimal allocation scheme comprises:
1) completing chromosome coding according to a genetic algorithm, and generating a certain number of chromosomes after the coding is completed, wherein genes in the chromosomes are randomly generated, and the chromosomes are used as a first generation population to begin to evolve;
2) screening the adaptive performance of the chromosomes in the first generation population according to a tournament selection method, randomly selecting two chromosomes in the first generation population each time for adaptive comparison, selecting one chromosome with stronger adaptability as a progeny population until all the chromosomes are screened, and reserving half of the chromosomes in the original population with stronger adaptability as the progeny population;
3) selecting two chromosomes in the offspring population each time, performing gene random part exchange of the two chromosomes until all the chromosomes are crossed, and selecting all the crossed chromosomes into the offspring population, wherein the offspring population comprises the crossed chromosomes and the chromosomes subjected to adaptive selection;
4) presetting the gene mutation probability, expanding the number of chromosomes of the offspring population through gene mutation, selecting the migration scheme of the RRH represented by the chromosome with the strongest adaptability in the offspring population after the number expansion as the optimal allocation scheme of the network resources, and dynamically allocating the network resources according to the optimal allocation scheme.
9. A 5G C-RAN resource prediction and allocation system, the system comprising:
the acquisition unit is used for acquiring flow data;
the processing unit is used for establishing an optimization problem of comprehensive energy consumption optimization and network service quality maintenance;
the prediction unit is used for establishing a flow prediction model aiming at the C-RAN system and obtaining flow prediction data according to the flow prediction model;
an allocation unit for designing a network resource allocation algorithm based on a genetic algorithm for network resource allocation according to the traffic prediction data; and obtaining an optimal allocation scheme of the network resources according to the network resource allocation algorithm, and dynamically allocating the network resources of the system according to the optimal allocation scheme.
10. A computer readable storage medium having stored thereon instructions which, when executed on a computer, cause the computer to perform the 5G C-RAN resource prediction and allocation method of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011285025.2A CN112584386A (en) | 2020-11-17 | 2020-11-17 | 5G C-RAN resource prediction and allocation method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011285025.2A CN112584386A (en) | 2020-11-17 | 2020-11-17 | 5G C-RAN resource prediction and allocation method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112584386A true CN112584386A (en) | 2021-03-30 |
Family
ID=75122741
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011285025.2A Pending CN112584386A (en) | 2020-11-17 | 2020-11-17 | 5G C-RAN resource prediction and allocation method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112584386A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114126019A (en) * | 2021-11-30 | 2022-03-01 | 北京邮电大学 | Forward-transmitting optical network dynamic resource mapping method and system based on energy efficiency optimization |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103024048A (en) * | 2012-12-17 | 2013-04-03 | 南京邮电大学 | Resources scheduling method under cloud environment |
| US20130308570A1 (en) * | 2012-05-17 | 2013-11-21 | Beijing University Of Posts And Telecommunications | Method for joint optimization of schedule and resource allocation based on the genetic algorithm |
| CN103945548A (en) * | 2014-04-29 | 2014-07-23 | 西安电子科技大学 | Resource distribution system and task/service scheduling method in C-RAN |
| EP1952655B1 (en) * | 2005-11-23 | 2016-02-17 | Telefonaktiebolaget LM Ericsson (publ) | Method, computer programm and base station for prediction based allocation of processing ressources in a non-serving base station |
| CN105740051A (en) * | 2016-01-27 | 2016-07-06 | 北京工业大学 | Cloud computing resource scheduling realization method based on improved genetic algorithm |
| CN106102079A (en) * | 2016-06-08 | 2016-11-09 | 西安电子科技大学 | Based on the C RAN carrier wave emigration resource requirement Forecasting Methodology improving PSO |
| WO2018064856A1 (en) * | 2016-10-08 | 2018-04-12 | 北京邮电大学 | Resource allocation method and device |
| CN110099415A (en) * | 2019-04-29 | 2019-08-06 | 哈尔滨工业大学（深圳） | A kind of cloud wireless access network computational resource allocation method and system based on volume forecasting |
| CN110493826A (en) * | 2019-08-28 | 2019-11-22 | 重庆邮电大学 | A kind of isomery cloud radio access network resources distribution method based on deeply study |
-
2020
- 2020-11-17 CN CN202011285025.2A patent/CN112584386A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1952655B1 (en) * | 2005-11-23 | 2016-02-17 | Telefonaktiebolaget LM Ericsson (publ) | Method, computer programm and base station for prediction based allocation of processing ressources in a non-serving base station |
| US20130308570A1 (en) * | 2012-05-17 | 2013-11-21 | Beijing University Of Posts And Telecommunications | Method for joint optimization of schedule and resource allocation based on the genetic algorithm |
| CN103024048A (en) * | 2012-12-17 | 2013-04-03 | 南京邮电大学 | Resources scheduling method under cloud environment |
| CN103945548A (en) * | 2014-04-29 | 2014-07-23 | 西安电子科技大学 | Resource distribution system and task/service scheduling method in C-RAN |
| CN105740051A (en) * | 2016-01-27 | 2016-07-06 | 北京工业大学 | Cloud computing resource scheduling realization method based on improved genetic algorithm |
| CN106102079A (en) * | 2016-06-08 | 2016-11-09 | 西安电子科技大学 | Based on the C RAN carrier wave emigration resource requirement Forecasting Methodology improving PSO |
| WO2018064856A1 (en) * | 2016-10-08 | 2018-04-12 | 北京邮电大学 | Resource allocation method and device |
| CN110099415A (en) * | 2019-04-29 | 2019-08-06 | 哈尔滨工业大学（深圳） | A kind of cloud wireless access network computational resource allocation method and system based on volume forecasting |
| CN110493826A (en) * | 2019-08-28 | 2019-11-22 | 重庆邮电大学 | A kind of isomery cloud radio access network resources distribution method based on deeply study |
Non-Patent Citations (4)
| Title |
|---|
| WEI-CHE CHIEN: "Dynamic Resource Prediction and Allocation in C-RAN With Edge Artificial Intelligence", 《IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS》 * |
| 王志朋等: "基于需求预测的云无线接入网计算资源分配策略研究", 《物联网学报》 * |
| 王汝言等: "带有混合能量供电的TWDM-PON与C-RAN联合架构资源分配机制", 《通信学报》 * |
| 王诗蕊: "C-RAN系统中的资源管理策略研究", 《中国优秀硕士论文电子期刊网》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114126019A (en) * | 2021-11-30 | 2022-03-01 | 北京邮电大学 | Forward-transmitting optical network dynamic resource mapping method and system based on energy efficiency optimization |
| CN114126019B (en) * | 2021-11-30 | 2024-04-23 | 北京邮电大学 | Energy efficiency optimization-based dynamic resource mapping method and system for forward optical network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108848520B (en) | Base station dormancy method based on flow prediction and base station state | |
| CN113543284A (en) | Energy saving method, base station, control unit and storage medium | |
| CN112286677A (en) | Resource-constrained edge cloud-oriented Internet of things application optimization deployment method | |
| CN106095529B (en) | A kind of carrier wave emigration method under C-RAN framework | |
| CN115238599B (en) | Energy-saving method and model reinforcement learning training method and device for refrigerating system | |
| CN113573363A (en) | MEC calculation unloading and resource allocation method based on deep reinforcement learning | |
| Lan et al. | A spectrum prediction approach based on neural networks optimized by genetic algorithm in cognitive radio networks | |
| CN112036651A (en) | Electricity price prediction method based on quantum immune optimization BP neural network algorithm | |
| CN114662751B (en) | Garden multifunctional short-term load forecasting and optimizing method based on LSTM | |
| CN112202847A (en) | Server resource allocation method based on mobile edge calculation | |
| CN112584386A (en) | 5G C-RAN resource prediction and allocation method and system | |
| CN113950134B (en) | Dormancy prediction method, device, equipment and computer readable storage medium of base station | |
| CN114615705B (en) | Single-user resource allocation strategy method based on 5G network | |
| CN112752327A (en) | Power regulation method and access network equipment | |
| CN100556171C (en) | The rapid simulation method of 3-G (Generation Three mobile communication system) network layer performance | |
| CN116388291A (en) | Large power grid new energy consumption capability calculation method, system, device and medium | |
| Kong et al. | Energy saving strategy for task migration based on genetic algorithm | |
| CN114978259A (en) | Deep learning power distribution method for downlink of cellular-free system | |
| CN114493090A (en) | Intelligent control method and device for comprehensive energy of industrial park | |
| CN111290853B (en) | Cloud data center scheduling method based on self-adaptive improved genetic algorithm | |
| CN111047465B (en) | Power grid friendly load response method based on power big data | |
| CN114189904A (en) | LTE carrier frequency resource scheduling method and device | |
| CN102625342B (en) | Method and equipment for optimizing LAC | |
| CN115473821B (en) | Network capacity prediction method and device, electronic equipment and storage medium | |
| CN107423847A (en) | A kind of Methods of electric load forecasting of distributed intelligence power distribution network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210330 |