CN111356141B - Distributed resource allocation method for inter-operator spectrum sharing based on network slice - Google Patents

Distributed resource allocation method for inter-operator spectrum sharing based on network slice Download PDF

Info

Publication number
CN111356141B
CN111356141B CN202010151220.XA CN202010151220A CN111356141B CN 111356141 B CN111356141 B CN 111356141B CN 202010151220 A CN202010151220 A CN 202010151220A CN 111356141 B CN111356141 B CN 111356141B
Authority
CN
China
Prior art keywords
operator
iteration
spectrum
slice
value
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.)
Active
Application number
CN202010151220.XA
Other languages
Chinese (zh)
Other versions
CN111356141A (en
Inventor
陈翔
陈嘉嘉
邱继云
龚杰
陈晓春
田华
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Research Institute Tsinghua University
Original Assignee
Shenzhen Research Institute Tsinghua University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shenzhen Research Institute Tsinghua University filed Critical Shenzhen Research Institute Tsinghua University
Priority to CN202010151220.XA priority Critical patent/CN111356141B/en
Publication of CN111356141A publication Critical patent/CN111356141A/en
Application granted granted Critical
Publication of CN111356141B publication Critical patent/CN111356141B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • H04W16/14—Spectrum sharing arrangements between different networks
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W72/00—Local resource management
    • H04W72/04—Wireless resource allocation
    • H04W72/044—Wireless resource allocation based on the type of the allocated resource
    • H04W72/0453—Resources in frequency domain, e.g. a carrier in FDMA
    • 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

Abstract

The invention discloses a distributed resource allocation method for inter-operator spectrum sharing based on network slicing, which is suitable for a scene of spectrum resource shortage caused by diversified services of a large number of vertical industries in a fifth generation mobile communication technology (5G). The method establishes a corresponding communication model aiming at the scene, and proposes a solution for the model by using an ADMM algorithm. The method comprises the following steps: initializing a model variable; the operator determines whether to lease the spectrum resources; and the operator obtains the lease quantity and the distribution proportion coefficient of the spectrum resources by using an ADMM algorithm according to the actual spectrum resources. The method solves the problem of low utilization rate of frequency spectrum resources caused by non-sharing of frequency spectrum resources among operators, and improves the utilization rate of frequency spectrum and the network utility of the whole communication model. And by the network slicing technology, the user rate requirements of different service scenes in 5G are met.

Description

Distributed resource allocation method for spectrum sharing among operators based on network slices
Technical Field
The invention relates to the technical field of resource allocation of wireless communication, in particular to a distributed resource allocation method for inter-operator spectrum sharing based on network slicing.
Background
The fifth generation communication technology (5G) will support a large number of diverse business scenarios from the vertical industry, such as high definition video, smart home, and autonomous driving. However, the conventional mobile communication network is mainly used for serving a single mobile bandwidth service, and cannot adapt to future diversified service scenarios of 5G. Furthermore, the emergence of diversified service scenarios also means that the demand of spectrum resources will increase continuously, and thus the spectrum resources will face the problem of shortage. Network slicing is a key technology for meeting communication requirements of different services on the same physical network, and through a network slicing technology, an operator can segment the upper part of the same physical network infrastructure into a plurality of virtual networks according to the requirements of different users, so that a 5G multi-service scene is met.
Regarding the network slicing technology, a large amount of literature has been devoted to research, and most of research is also based on the spectrum resource allocation problem. However, in these studies, some system spectrum resources in the studies are allocated to each user in advance, but the spectrum resources required by each user in real life are not fixed. Further research does not take into account that users under different slice types have different rate requirements, which is also not in line with the actual requirements of network slices. Therefore, there is a need to develop a research related to a dynamic spectrum resource allocation method based on a network.
Disclosure of Invention
The invention aims to solve the problem of spectrum resource shortage in the prior art and meet the user rate requirements in different 5G service scenes, and provides a distributed resource allocation method for spectrum sharing among operators based on network slicing.
The purpose of the invention can be achieved by adopting the following technical scheme:
a distributed resource allocation method based on inter-operator spectrum sharing of network slicing is suitable for a scene of spectrum resource shortage caused by diversified services of a large number of vertical industries in a fifth generation mobile communication technology (5G), and the specific operation steps of the distributed resource allocation method are as follows:
s1, initializing parameters of the communication model, determining the number of operators in the communication model as M, and determining the number of slices under any operator M as N m The number of users under slice n is U m，n The original spectrum resource quantity owned by the operator m is B m And the loan price per unit of spectrum resource is p m The spectrum resource cost of the operator m in the leasing process is C m The rate requirement of the user u of any slice type n under the operator m is L m，n The signal-to-noise ratio of the uplink channel is SNR m，n，u (ii) a Initializing method parameters, initializing relaxation variables x m 、y m，n，u 、z m 、u m，j 、v m，n，u 、w m，n Initializing a dual variable a to zero m 、d m，n，u 、h m 、k m，j 、l m，n，u 、t m，n Initializing a penalty factor rho when the value is zero, wherein the iteration number k is 0;
s2, calculating the spectrum utilization rate c of the user under each slice type according to the signal-to-noise ratio by the operator m，n，u ＝log 2 (1+SNR m，n，u )；
S3, starting iteration, and in the process of the (k + 1) th iteration, according to the total bandwidth B capable of being provided by the iteration m And the transmission rate L required to be satisfied by the user under different slice types m，n The operator determines whether the operator borrows the spectrum resources to other operators and which operator to borrow the spectrum resources;
s4, after determining whether to lease the spectrum and the target of the spectrum lease, the operator m obtains the amount of the spectrum to be leased based on the Alternating Direction Method of Multipliers (ADMM), thereby obtaining the final owned spectrum resource. According to the finally owned spectrum resources, the operator m obtains the spectrum resource distribution proportion coefficient f of the slice n m，n And the spectrum resource allocation proportion coefficient s of the slice n to the user u m，n，u Then, the operator m parallelly allocates the spectrum resources to the corresponding slices, and the slices provide corresponding services for the users according to the allocated spectrum resources;
s5, obtaining the relaxation variable x by gradient descent method according to the parameters, the relaxation variable and the value of the dual variable after the k-th iteration obtained in the step S4 m 、y m，n，u 、z m 、u m，j 、v m，n，u 、w m，n And dual variable a m 、d m，n，u 、h m 、k m，j 、l m，n，u 、t m，n And sequentially obtaining values of the relaxation variable and the dual variable after the (k + 1) th iteration according to the relational expression of the (k) th iteration and the (k + 1) th iteration, wherein the values are as follows:
Figure BDA0002402501650000031
Figure BDA0002402501650000032
Figure BDA0002402501650000033
Figure BDA0002402501650000034
Figure BDA0002402501650000035
Figure BDA0002402501650000036
Figure BDA0002402501650000037
Figure BDA0002402501650000038
Figure BDA0002402501650000039
Figure BDA00024025016500000310
Figure BDA00024025016500000311
Figure BDA0002402501650000041
wherein, in the formula (1.1)
Figure BDA0002402501650000042
Is x m After the (k + 1) th iteration,
Figure BDA0002402501650000043
is q m，j The value of the initial value of (a) after the (k + 1) th iteration,
Figure BDA0002402501650000044
is a m After the kth iteration, of the initial value of (2), equation (1)
Figure BDA0002402501650000045
Is y m，n，u The value of the initial value of (a) after the (k + 1) th iteration,
Figure BDA0002402501650000046
is d m，n，u After the kth iteration, of the initial value of (1.3)
Figure BDA0002402501650000047
Is z m The value after the (k + 1) th iteration,
Figure BDA0002402501650000048
is h m After the kth iteration, of the initial value of (1.4)
Figure BDA0002402501650000049
Is u m，j The value of the initial value of (a) after the (k + 1) th iteration,
Figure BDA00024025016500000410
is k m，j After the kth iteration, of the initial value of (a), equation (1.5)
Figure BDA00024025016500000411
Is v m，n，u The value of the initial value of (a) after the (k + 1) th iteration,
Figure BDA00024025016500000412
is s m，n，u After the (k + 1) th iteration,
Figure BDA00024025016500000413
is a m，n，u After the kth iteration, of the value of (1.6)
Figure BDA00024025016500000414
Is w m，n After the (k + 1) th iteration,
Figure BDA00024025016500000415
is f m，n The value of the initial value of (a) after the (k + 1) th iteration,
Figure BDA00024025016500000416
is t m，n After the kth iteration, of the initial value of (1.7)
Figure BDA00024025016500000417
Is a m After the (k + 1) th iteration, of the initial value of (1.8)
Figure BDA00024025016500000418
Is d m，n，u After the (k + 1) th iteration, of the initial value of (1.9)
Figure BDA00024025016500000419
Is h m After the (k + 1) th iteration, of the initial value of (1.10)
Figure BDA00024025016500000420
Is k m，j After the (k + 1) th iteration, of the initial value of (1.11)
Figure BDA00024025016500000421
Is a 1 m，n，u After the (k + 1) th iteration, of the initial value of (1.12)
Figure BDA00024025016500000422
Is t m，n The initial value of (a) after the (k + 1) th iteration;
s6, according to the distribution situation of the frequency spectrum resource, the transmission rate r of the user u of the slice n under the operator m is obtained m，n，u ＝b m f m，n s m，n，u c m，n，u The total network utility of the communication model is defined as
Figure BDA00024025016500000423
Figure BDA00024025016500000424
According to the transmission rate r of the user m，n，u The total network utility can be obtained;
and S7, judging whether the network utility is converged, if not, jumping back to the step S3 for next iteration, and if so, ending the distribution method.
Further, in this embodiment, the process of step S3 is as follows:
s31, defining the frequency spectrum resource q borrowed by the operator m whose frequency spectrum resource can not satisfy the transmission rate of the user under the slice to the operator j whose frequency spectrum resource satisfies the transmission rate of the user under the slice m，j And q is m，j Is zero;
s32, according to the condition that the operator owns the spectrum resource, determining that the spectrum resource lent to the operator j by the operator m is q m，j To meet the user transmission rates of different slice types and calculate the lease fee sigma generated by the operator m in the lease process m≠j q m，j p j ；
S33 spectrum resource q if operator m borrows from operator j m，j Greater than the original spectrum resource B of operator j j I.e. q m，j ≥B j If the lease process is over, or if the cost of the operator m for the spectrum resource is higher than the self-regulationCost Cm, i.e. Σ m≠j q m，j p j ≥C m The lease process ends or q m，j If the result is less than 0, the lease process is ended.
S34, each operator broadcasts the lease condition of the frequency spectrum resource to all operators, and the operators update the actual frequency spectrum amount b m Wherein, the spectrum resource can not satisfy the actual spectrum amount b of the operator of the transmission rate of the user under the slice after the spectrum leasing process is finished m ＝B m +∑ j≠m q m，j The actual spectrum amount of operator j with spectrum resources meeting the transmission rate of users under the slice after the lease process is finished is b j ＝B j -∑ j≠m q m，j 。
Further, in this embodiment, the process of step S4 is as follows:
s41, initializing the spectrum resource allocation proportion coefficient f allocated to the slice n by the operator m m，n And the spectrum resource allocation proportion coefficient S of the slice n to the user u m，n，u Is zero;
s42, obtaining the spectrum lease q according to the step S3 m，j And obtaining that the frequency spectrum resource actually owned by the operator m is Bm-sigma i≠m q i，m +∑ j≠m q m，j Then the actual transmission rate of user u of slice n under operator m is r m，n，u ＝b m f m，n s m，n， u c m，n，u Defining the total network utility of the communication model as
Figure BDA0002402501650000051
Establishing a communication model with the maximized network total utility as an optimization objective function, wherein the model can be abstracted into a convex problem, and the convex problem has the following limitations: (1) transmission rate r of user u of operator m down-slice n m，n，u Must not be less than the minimum transmission rate requirement L m，n I.e. r m，n，u ≥L m，n (2) Resource allocation scaling factor f m，n And s m，n，u The value range of (A) is 0 to 1, the sum of the two should not exceed 1, f is less than or equal to m，n ≤1，
Figure BDA0002402501650000061
Figure BDA0002402501650000062
And 0. ltoreq.s m，n，u ≤1，
Figure BDA0002402501650000063
S43 using the relaxation variable x m 、y m，n，u 、z m 、u m，j 、v m，n，u 、w m，n Converting the inequality constraints in steps S33 and S42 into equality constraints, wherein x m Q is to be m，j ≥B j Conversion to q m，j -B j +x m ＝0，y m，n，u Will r is m，n，u ≥L m，n Is converted into r m，n，u -L m，n -y m，n，u ＝0，z m Will be sigma m≠j q m，j p j ≥C m Conversion to sigma m≠j q m，j p j -C m +z m ＝0，u m，j Q is to be m，j Conversion to q of ≥ 0 m，j -u m，j ＝0，v m，n，u Will s is m，n，u Conversion of ≥ 0 into S m，n，u -v m，n，u ＝0，w m，n Will f is mixed m，n Conversion of ≥ 0 into f m，n -w m，n Converting the non-normalized convex optimization problem into a normalized convex optimization problem suitable for the ADMM algorithm;
s44, converting the normalized convex optimization problem in the step S43 into a Lagrange function form by adding a penalty factor rho;
s45, using dual variable a m 、d m，n，u 、h m 、k m，j 、l m，n，u 、t m，n Adding the equality constraint of step S43 to the function of step S44, thereby eliminating the equality constraint in the communication model and the dual problem in the form of the Lagrangian function in deriving step S44, wherein a m Constraint q of corresponding equation m，j -B j +x m ＝0，d m，n，u Corresponding equation r m，n，u -L m，n -y m，n，u ＝0，h m Constraint of equation m≠j q m，j p j -C m +z m ＝0，k m，j Constraint q of corresponding equation m，j -u m，j ＝0，l m，n，u Constraint S of corresponding equation m，n，u -v m，n，u ＝0，t m，n Constraint f of corresponding equation m，n -w m，n ＝0；
S46, in the process of the kth iteration, the dual problem of the step S45 is solved by using a gradient descent method at each operator, and therefore the resource proportion distribution coefficient f is obtained m，n And s m，n，u The convex optimization tool is utilized to obtain the resource proportion coefficient f distributed to the slice n by the operator m m，n And the spectrum resource allocation ratio s of slice n to user u m，n，u 。
Compared with the prior art, the invention has the following advantages and effects:
1. the invention improves the utilization rate of frequency spectrum resources. The spectrum resources are shared among operators, so that operators with deficient spectrum resources can rent the spectrum resources from operators with abundant spectrum resources, and the utilization rate of the spectrum resources and the network utility of the whole communication model are improved.
2. The invention meets the user speed requirement under 5G different service scenes. According to the characteristic that the speed requirements of users under different slice types are different, a general model is established for the user speeds of different slice types, so that the user speed requirements under 5G different service scenes are met.
3. Based on the ADMM algorithm, the invention provides a distributed resource allocation method, in which an operator can allocate own spectrum resources in parallel. Compared with a centralized resource allocation method, the distributed resource allocation method provided by the invention is more consistent with the actual situation and more general.
Drawings
Fig. 1 is a flowchart of a distributed resource allocation method for inter-operator spectrum sharing based on network slicing according to the present invention;
fig. 2 is a flow chart of inter-operator spectrum resource leasing in the present invention;
FIG. 3 is a flow chart of the distributed resource allocation by the operator using the ADMM algorithm in the present invention;
FIG. 4 is a graph comparing the overall network utility of the system with and without shared spectrum resources among operators in accordance with the present invention;
fig. 5 is a graph of transmission rate versus user for different slice types for an operator.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example one
The embodiment provides a distributed resource allocation method based on inter-operator spectrum sharing of network slices, which specifically comprises the following operation steps:
s1, initializing parameters of the communication model, determining the number of operators in the communication model as M, and determining the number of slices under any operator M as N m The number of users under slice n is U m，n The original spectrum resource quantity owned by the operator m is B m And the loan price per unit of spectrum resource is p m The spectrum resource cost of the operator m in the leasing process is C m The user u rate requirement of any slice type n under the operator m is L m，n The signal-to-noise ratio of the uplink channel is SNR m，n，u (ii) a Initializing method parameters, initializing relaxation variables x m 、y m，n，u 、z m 、u m，j 、v m，n，u 、w m，n Initializing a dual variable a to zero m 、d m，n，u 、h m 、k m，j 、l m，n，u 、t m，n And initializing a penalty factor rho when the number of iterations k is zero, wherein the number of iterations k is 0.
S2, calculating the spectrum utilization rate c of the user under each slice type according to the signal-to-noise ratio by the operator m，n，u ＝log 2 (1+SNR m，n，u )。
S3, starting iteration, and in the process of the (k + 1) th iteration, according to the total bandwidth B capable of being provided by the iteration m And the transmission rate L required to be satisfied by the user under different slice types m，n The operator determines whether the operator borrows the spectrum resources to other operators and which operator to borrow the spectrum resources;
in this embodiment, the process of step S3 is as follows:
s31, defining the frequency spectrum resource q borrowed by the operator m whose frequency spectrum resource can not meet the transmission rate of the user under the slice to the operator j whose frequency spectrum resource meets the transmission rate of the user under the slice m，j And q is m，j Is zero;
s32, determining the frequency spectrum resource lent from the operator m to the operator j as q according to the condition that the operator owns the frequency spectrum resource m，j To meet the user transmission rates of different slice types and calculate the lease fee sigma generated by the operator m in the lease process m≠j q m，j p j ；
S33 spectrum resource q if operator m borrows from operator j m，j Greater than the original spectrum resource B of operator j j I.e. q m，j ≥B j The lease process is over or if the operator m has a higher cost of borrowing spectrum resources than the cost C specified by the operator m m I.e. Σ m≠j q m，j p j ≥C m If yes, the leasing process is ended;
s34, each operator broadcasts the lease condition of the frequency spectrum resource to all operators, and the operators update the actual frequency spectrum amount b m Wherein, the spectrum resource can not satisfy the actual spectrum amount b of the operator of the transmission rate of the user under the slice after the spectrum leasing process is finished m ＝B m +∑ j≠m q m，j The actual spectrum amount of the operator j with spectrum resources meeting the transmission rate of the users under the slice after the lease process is finished is b j ＝B j -∑ j≠m q m，j ；
S4, after determining whether to lease the spectrum and the target of the spectrum lease, the operator m obtains the amount of the spectrum to be leased based on the Alternating Direction Method of Multipliers (ADMM), thereby obtaining the final owned spectrum resource. According to the finally owned spectrum resources, the operator m obtains the spectrum resource distribution proportion coefficient f of the slice n m，n And the spectrum resource allocation proportion coefficient s of the slice n to the user u m，n，u Then, the operator m parallelly allocates the spectrum resources to the corresponding slices, and the slices provide corresponding services for the users according to the allocated spectrum resources;
in this embodiment, the procedure of step S4 is as follows:
s41, initializing the spectrum resource allocation proportion coefficient f allocated to the slice n by the operator m m，n And the spectrum resource allocation proportion coefficient s of the slice n to the user u m，n，u Is zero;
s42, obtaining the lease q of the frequency spectrum according to the step S3 m，j And obtaining the frequency spectrum resource actually owned by the operator m as b m ＝B m -∑ i≠m q i，m +∑ j≠m q m，j Then the actual transmission rate of user u of slice n under operator m is r m，n，u ＝b m f m，n s m，n， u C m，n，u Defining the total network utility of the communication model as
Figure BDA0002402501650000091
Establishing a communication model with the maximized network total utility as an optimization objective function, wherein the model can be abstracted into a convex problem, and the convex problem has the following limitations: (1) transmission rate r of user u of operator m down-slice n m，n，u Must not be less than the minimum transmission rate requirement L m，n I.e. r m，n，u ≥L m，n (2) Resource partitioningCoefficient of proportionality f m，n And s m，n，u The value range of (a) is 0 to 1, the sum of the two should not exceed 1, f is less than or equal to m，n ≤1，
Figure BDA0002402501650000092
Figure BDA0002402501650000093
And 0. ltoreq.s m，n，u ≤1，
Figure BDA0002402501650000094
S43 using the relaxation variable x m 、y m，n，u 、z m 、u m，j 、v m，n，u 、w m，n Converting the inequality constraints in steps S33 and S42 into equality constraints, wherein x m Q is to be m，j ≥B j Conversion to q m，j -B j +x m ＝0，y m，n，u Will r is m，n，u ≥L m，n Is converted into r m，n，u -L m，n -y m，n，u ＝0，z m Will be sigma m≠j q m，j p j ≥C m Conversion to sigma m≠j q m，j p j -C m +z m ＝0，u m，j Q is to be m，j Conversion to q of ≥ 0 m，j -u m，j ＝0，v m，n，u Will s m，n，u Conversion of ≧ 0 to s m，n，u -v m，n，u ＝0，w m，n Will f is m，n Conversion of ≥ 0 into f m，n -w m，n The non-normalized convex optimization problem is converted into a normalized convex optimization problem suitable for the ADMM algorithm, wherein the value of the normalized convex optimization problem is 0;
s44, converting the normalized convex optimization problem in the step S44 into a Lagrangian function form by adding a penalty factor rho;
s45, using dual variable a m 、d m，n，u 、h m 、k m，j 、l m，n，u 、t m，n Coupling the constraints of the equation in step S44 into an optimization objective function of the communication model,thereby eliminating the equality constraint in the communication model and the dual problem in the form of the lagrange function in the deriving step S44, where a m Constraining equation to q m，j -B j +x m 0 to an optimization objective function, d m，n，u Will be equation r m，n，u -L m，n -y m，n，u Coupled to the objective function, h m Constrain equation to Sigma m≠j q m，j p j -C m +z m Coupled to the objective function, k m，j Constraining equation to q m，j -u m，j Coupled to the objective function, < 0 >, l m，n，u Constraining the equality to s m，n，u -v m，n，u Coupled to the objective function, t ═ 0 m，n Constraining equation to f m，n -w m，n 0 is coupled into the objective function;
s46, in the process of the k-th iteration, solving the dual problem of the step S45 by using a gradient descent method at each operator, thereby obtaining a resource proportion distribution coefficient f m，n And s m，n，u The convex optimization tool is utilized to obtain the resource proportion coefficient f distributed to the slice n by the operator m m，n And the spectrum resource allocation ratio s of slice n to user u m，n，u ；
S5, obtaining a slack variable x by a gradient descent method according to the parameters obtained in the step S4 and the values of the dual variable and the slack variable after the kth iteration m 、y m，n，u 、z m 、u m，j 、v m，n，u 、w m，n And dual variable a m 、d m，n，u 、h m 、k m，j 、l m，n，u 、t m，n And sequentially obtaining values of the relaxation variable and the dual variable after the (k + 1) th iteration according to the relational expression of the (k) th iteration and the (k + 1) th iteration, wherein the values are as follows:
Figure BDA0002402501650000111
Figure BDA0002402501650000112
Figure BDA0002402501650000113
Figure BDA0002402501650000114
Figure BDA0002402501650000115
Figure BDA0002402501650000116
Figure BDA0002402501650000117
Figure BDA0002402501650000118
Figure BDA0002402501650000119
Figure BDA00024025016500001110
Figure BDA00024025016500001111
Figure BDA00024025016500001112
wherein, in the formula (1.1)
Figure BDA00024025016500001113
Is x m The value of the initial value of (a) after the (k + 1) th iteration,
Figure BDA00024025016500001114
is q m，j After the (k + 1) th iteration,
Figure BDA00024025016500001115
is a m After the kth iteration, of the value of (1.2)
Figure BDA00024025016500001116
Is y m，n，u The value of the initial value of (a) after the (k + 1) th iteration,
Figure BDA00024025016500001117
is d m，n，u After the kth iteration, of the initial value of (1.3)
Figure BDA0002402501650000121
Is z m The value after the (k + 1) th iteration,
Figure BDA0002402501650000122
is h m After the kth iteration, of the initial value of (1.4)
Figure BDA0002402501650000123
Is u m，j After the (k + 1) th iteration,
Figure BDA0002402501650000124
is k m，j After the kth iteration, of the initial value of (a), equation (1.5)
Figure BDA0002402501650000125
Is v m，n，u At the (k + 1) th iterationThe value of the generation after the generation,
Figure BDA0002402501650000126
is s m，n，u The value of the initial value of (a) after the (k + 1) th iteration,
Figure BDA0002402501650000127
is a m，n，u After the kth iteration, of the initial value of (1.6)
Figure BDA0002402501650000128
Is w m，n The value of the initial value of (a) after the (k + 1) th iteration,
Figure BDA0002402501650000129
is f m，n After the (k + 1) th iteration,
Figure BDA00024025016500001210
is t m，n After the kth iteration, of the initial value of (1.7)
Figure BDA00024025016500001211
Is a m After the (k + 1) th iteration, of the initial value of (1.8)
Figure BDA00024025016500001212
Is d m，n，u After the (k + 1) th iteration, of the initial value of (1.9)
Figure BDA00024025016500001213
Is h m After the (k + 1) th iteration, of the initial value of (1.10)
Figure BDA00024025016500001214
Is k m，j After the (k + 1) th iteration, of the initial value of (1.11)
Figure BDA00024025016500001215
Is a m，n，u At an initial value ofValue after the k +1 th iteration, of equation (1.12)
Figure BDA00024025016500001216
Is t m，n The initial value of (a) after the (k + 1) th iteration;
s6, according to the distribution situation of the frequency spectrum resource, the transmission rate r of the user u of the slice n under the operator m is obtained m，n，u ＝b m f m，n s m，n，u c m，n，u The total network utility of the system model is defined as
Figure BDA00024025016500001217
Figure BDA00024025016500001218
The total network utility can be obtained according to the transmission rate of the user.
And S7, judging whether the network utility is converged, if not, jumping back to the step S3 for next iteration, and if so, ending the distribution method.
Example two
The present embodiment will describe the method for allocating distributed resources based on spectrum sharing between operators by using a specific network slicing system with multiple operators and multiple types of slices in conjunction with fig. 1 to fig. 5.
Consider the system model as follows: assume that there are 4 operators in the system, each operator has 2 slice types, and different slice types have different transmission rate requirements. Corresponding services need to be provided for 2 users under each slice type. Specific simulation parameters are shown in table 1.
TABLE 1 simulation parameter Table
Operator 1 2 3 4
Spectrum resource 50Mhz 20Mhz 50Mhz 20Mhz
Price per unit of spectrum resource 15 30 15 20
Cost limitation 200 150 200 300
Transmission rate requirement for slice type 1 20Mhz 10Mhz 15Mhz 60Mhz
Transmission rate requirement for slice type 2 1Mhz 40Mhz 30Mhz 1Mhz
Suppose that a spectrum resource borrowed from an operator j by a certain operator m is q m，j The resource proportion coefficient allocated to its slice n by the operator m is f m，n The resource proportion coefficient allocated to the user u by the slice n under the operator m is s m，n，u . The resource allocation problem of the system is constructed into a decomposable convex optimization problem, and then a distributed algorithm suitable for the decomposable convex problem, namely an alternating direction multiplier method, is used for solving. The method adopts a decomposition-coordination process form to coordinate the solution of the small local subproblem, and finds the solution of the largest global problem. Through the ADMM method, each operator can solve q in parallel m，j ，f m，n And s and m，n，u thereby obtaining the self frequency spectrum resource allocation scheme.
Fig. 4 shows a comparison of the total network utility of the system with or without shared spectrum resources between operators. When spectrum resource borrowing is allowed in the system as shown in fig. 4, the optimal network utility of the system varies from 72 to 79 as the signal-to-noise ratio increases. When spectrum resource borrowing is not allowed in the system, the optimal network utility of the system changes from 68 to 74 as the signal-to-noise ratio increases. This proves that when the borrowing of the spectrum resources is allowed in the system, the utilization rate of the spectrum resources and the network utility of the system can be effectively improved.
Fig. 5 shows a transmission rate comparison of users in different slice types of the operator. As shown in fig. 5, in contrast to the transmission rate requirements of the slices in table 1, the transmission rate of the user under each operator can reach the rate requirements specified under different slice types. In order to achieve the maximum network utility, operators 1 and 3 equally allocate spectrum resources to different types of slices on the premise of meeting the transmission rate requirements of the slices. And operator 2 and operator 4 are scarce in spectrum resources, they borrow spectrum resources from operator 1 and operator 3 to meet the transmission rate requirements for the slice.
In summary, the distributed resource allocation method for inter-operator spectrum sharing based on network slicing provided by this embodiment can effectively improve the utilization rate of spectrum resources and the network utility of the system, and meet the rate requirements under different slicing types.
The above embodiments are preferred embodiments of the present invention, but the present invention is not limited to the above embodiments, and any other changes, modifications, substitutions, combinations, and simplifications which do not depart from the spirit and principle of the present invention should be construed as equivalents thereof, and all such changes, modifications, substitutions, combinations, and simplifications are intended to be included in the scope of the present invention.

Claims (3)

1. A distributed resource allocation method based on inter-operator spectrum sharing of network slices is characterized in that the distributed resource allocation method comprises the following steps:
s1, initializing parameters of the communication model, determining the number of operators in the communication model as M, and determining the number of slices under any operator M as N m The number of users under slice n is U m,n The original spectrum resource quantity owned by the operator m is B m And the loan price per unit of spectrum resource is p m The spectrum resource cost of the operator m in the leasing process is C m The user u rate requirement of any slice type n under the operator m is L m,n The signal-to-noise ratio of the uplink channel is SNR m,n,u (ii) a Initializing method parameters, initializing relaxation variables x m 、y m,n,u 、z m 、u m,j 、v m,m,u 、w m,n Initializing a dual variable a to zero m 、d m,n,u 、h m 、k m,j 、l m,n,u 、t m,n Initializing a penalty factor rho when the value is zero, wherein the iteration number k is 0;
s2, calculating the spectrum utilization rate c of the user under each slice type according to the signal-to-noise ratio by the operator m,n,u ＝log 2 (1+SNR m,n,u )；
S3, starting iteration, and in the process of the (k + 1) th iteration, according to the total bandwidth B capable of being provided by the iteration m And the transmission rate L required to be satisfied by the user under different slice types m,n The operator determines whether the operator borrows the spectrum resources to other operators and which operator to borrow the spectrum resources;
s4, the operator m obtains the spectrum amount needing to be leased based on the ADMM algorithm to obtain the finally owned spectrum resource, and then the operator m obtains the spectrum resource distribution proportion coefficient f of the slice n according to the finally owned spectrum resource m,n And the spectrum resource allocation proportion coefficient s of the slice n to the user u m,n,u Then, the operator m allocates the spectrum resources to the corresponding slices in parallel, and the slices provide corresponding services for the users according to the allocated spectrum resources;
s5, obtaining the relaxation variable x by gradient descent method according to the parameters, the relaxation variable and the value of the dual variable after the k-th iteration obtained in the step S4 m 、y m,n,u 、z m 、u m,j 、v m,n,u 、w m,n And dual variable a m 、d m,n,u 、h m 、k m,j 、l m,n,u 、t m,n And sequentially obtaining values of the relaxation variable and the dual variable after the (k + 1) th iteration according to the relational expression of the (k) th iteration and the (k + 1) th iteration, wherein the values are as follows:
Figure FDA0003731106580000021
Figure FDA0003731106580000022
Figure FDA0003731106580000023
Figure FDA0003731106580000024
Figure FDA0003731106580000025
Figure FDA0003731106580000026
Figure FDA0003731106580000027
Figure FDA0003731106580000028
Figure FDA0003731106580000029
Figure FDA00037311065800000210
Figure FDA00037311065800000211
Figure FDA0003731106580000031
wherein, in the formula (1.1)
Figure FDA0003731106580000032
Is x m After the (k + 1) th iteration,
Figure FDA0003731106580000033
is q m,j The value of the initial value of (a) after the (k + 1) th iteration,
Figure FDA0003731106580000034
is a m After the kth iteration, of the initial value of (2), equation (1)
Figure FDA0003731106580000035
Is y m,n,u The value of the initial value of (a) after the (k + 1) th iteration,
Figure FDA0003731106580000036
is d m,n,u After the kth iteration, of the initial value of (1.3)
Figure FDA0003731106580000037
Is z m The value after the (k + 1) th iteration,
Figure FDA0003731106580000038
is h m After the kth iteration, of the value of (1.4)
Figure FDA0003731106580000039
Is u m,j The value of the initial value of (a) after the (k + 1) th iteration,
Figure FDA00037311065800000310
is k m,j After the kth iteration, of the initial value of (a), equation (1.5)
Figure FDA00037311065800000311
Is v m,n,u The value of the initial value of (a) after the (k + 1) th iteration,
Figure FDA00037311065800000312
is s m,n,u The value of the initial value of (a) after the (k + 1) th iteration,
Figure FDA00037311065800000313
is a 1 m,n,u After the kth iteration, of the value of (1.6)
Figure FDA00037311065800000314
Is w m,n The value of the initial value of (a) after the (k + 1) th iteration,
Figure FDA00037311065800000315
is f m,n The value of the initial value of (a) after the (k + 1) th iteration,
Figure FDA00037311065800000316
is t m,n After the kth iteration, of the initial value of (1.7)
Figure FDA00037311065800000317
Is a m After the (k + 1) th iteration, of the value of (1.8)
Figure FDA00037311065800000318
Is d m,n,u After the (k + 1) th iteration, of the initial value of (1.9)
Figure FDA00037311065800000319
Is h m After the (k + 1) th iteration, of the initial value of (1.10)
Figure FDA00037311065800000320
Is k m,j After the (k + 1) th iteration, of the initial value of (1.11)
Figure FDA00037311065800000321
Is a 1 m,n,u After the (k + 1) th iteration, of the initial value of (1.12)
Figure FDA00037311065800000322
Is t m,n The initial value of (a) after the (k + 1) th iteration;
s6, according to the distribution situation of the frequency spectrum resource, the transmission rate r of the user u of the slice n under the operator m is obtained m,n,u ＝b m f m,n s m,n,u c m,n,u The total network utility of the communication model is defined as
Figure FDA00037311065800000323
Figure FDA00037311065800000324
According to the transmission rate r of the user m,n,u The total network utility can be obtained;
and S7, judging whether the network utility is converged, if not, jumping back to the step S3 for next iteration, and if so, ending the distribution method.
2. The method for distributed resource allocation based on inter-operator spectrum sharing of network slice according to claim 1, wherein the step S3 is performed as follows:
s31, defining that the spectrum resource can not meet the transmission rate L of the user under the slice m,n The spectrum resource borrowed by the operator j, the spectrum resource of which satisfies the transmission rate of the user under the slice from the operator m, is q m,j And q is m,j Is zero;
s32, determining the frequency spectrum resource lent from the operator m to the operator j as q according to the condition that the operator owns the frequency spectrum resource m,j To meet the user transmission rates of different slice types and calculate the lease fee sigma generated by the operator m in the lease process m≠j q m,j p j ；
S33, if the operator m borrows the frequency spectrum resource q from the operator j m,j Greater than the original spectrum resource B of operator j j I.e. q m,j ≥B j The lease process is over or if the operator m has a higher cost of borrowing spectrum resources than the cost C specified by the operator m m I.e. Σ m≠j q m,j p j ≥C m The lease process ends or q m,j <0, the lease process is finished;
s34, each operator broadcasts the lease condition of the frequency spectrum resource to all operators, and the operators update the actual frequency spectrum amount b m Wherein, the spectrum resource can not satisfy the actual spectrum amount b of the operator of the transmission rate of the user under the slice after the spectrum leasing process is finished m ＝B m +∑ j≠m q m,j The actual spectrum amount of operator j with spectrum resources meeting the transmission rate of users under the slice after the lease process is finished is b j ＝B j -∑ j≠m q m,j 。
3. The method for distributed resource allocation based on inter-operator spectrum sharing of network slice according to claim 1, wherein the step S4 is performed as follows:
s41, initializing the spectrum resource distribution proportion coefficient f distributed to the slice n by the operator m m,n And the spectrum resource allocation proportion coefficient s of the slice n to the user u m,n,u Is zero;
s42, obtaining the spectrum lease q according to the step S3 m,j And obtaining the frequency spectrum resource actually owned by the operator m as b m ＝B m +∑ j≠m q m,j Then the actual transmission rate of user u of slice n under operator m is r m,n,u ＝b m f m,n s m,n,u c m,n,u Defining the total network utility of the communication model as
Figure FDA0003731106580000041
Figure FDA0003731106580000051
Establishing a communication model with the maximized network total utility as an optimization objective function, wherein the model can be abstracted into a convex problem, and the convex problem has the following limitations: (1) transmission rate r of user u of operator m down-slice n m,n,u Must not be less than the minimum transmission rate requirement L m,n I.e. r m,n,u ≥L m,n (ii) a (2) Resource allocation scaling factor f m,n And s m,n,u The value range of (1) is 0 to 1, the sum of the two should not exceed 1, f is not less than 0 m,n ≤1，
Figure FDA0003731106580000052
And 0. ltoreq.s m,n,u ≤1，
Figure FDA0003731106580000053
S43 using the relaxation variable x m 、y m,n,u 、z m 、u m,j 、v m,n,u 、w m,n Converting the inequality constraints in steps S33 and S42 into equality constraints, wherein x m Q is to be m,j ≥B j Conversion to q m,j -B j +x m ＝0，y m,n,u Will r is m,n,u ≥L m,n Is converted into r m,n,u -L m,n -y m,n,u ＝0，z m Will be sigma m≠j q m,j p j ≥C m Conversion to sigma m≠j q m,j p j -C m +z m ＝0，u m,j Q is to be m,j Conversion to q of ≥ 0 m,j -u m,j ＝0，v m,n,u Will s m,n,u Conversion of ≧ 0 to s m,n,u -v m,n,u ＝0，w m,n Will f is m,n Conversion of ≥ 0 into f m,n -w m,n The non-normalized convex optimization problem is converted into a normalized convex optimization problem suitable for the ADMM algorithm, wherein the value of the normalized convex optimization problem is 0;
s44, converting the normalized convex optimization problem in the step S42 into a Lagrange function form by adding a penalty factor rho;
s45, using dual variable a m 、d m,n,u 、h m 、k m,j 、l m,n,u 、t m,n Coupling the equality constraint of step S44 to the optimization objective function of the communication model, thereby eliminating the equality constraint of the communication model and deriving the dual problem in the form of the Lagrangian function of step S44, wherein a m Constraining equation to q m,j -B j +x m Coupled with an optimization objective function, d m,n,u Will be equation r m,n,u -L m,n -y m,n,u 0 coupled with the optimization objective function, h m Constrain equation to Sigma m≠j q m,j p j -C m +z m Coupled with the optimization objective function, k m,j Constraining equation to q m,j -u m,j Coupled with an optimization objective function, < 0 > m,n,u Constraining the equation to s m,n,u -v m,n,u Coupled with the optimization objective function, t m,n Constraining equation to f m,n -w m,n 0 is coupled to the optimization objective function;
s46, in the process of the kth iteration, the dual problem of the step S45 is solved by using a gradient descent method at each operator, and therefore the resource proportion distribution coefficient f is obtained m,n And s m,n,u The convex optimization tool is utilized to obtain the resource proportion coefficient f distributed to the slice n by the operator m m,n And the spectrum resource allocation ratio s of slice n to user u m,n,u 。
CN202010151220.XA 2020-03-06 2020-03-06 Distributed resource allocation method for inter-operator spectrum sharing based on network slice Active CN111356141B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010151220.XA CN111356141B (en) 2020-03-06 2020-03-06 Distributed resource allocation method for inter-operator spectrum sharing based on network slice

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010151220.XA CN111356141B (en) 2020-03-06 2020-03-06 Distributed resource allocation method for inter-operator spectrum sharing based on network slice

Publications (2)

Publication Number Publication Date
CN111356141A CN111356141A (en) 2020-06-30
CN111356141B true CN111356141B (en) 2022-08-30

Family

ID=71196052

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010151220.XA Active CN111356141B (en) 2020-03-06 2020-03-06 Distributed resource allocation method for inter-operator spectrum sharing based on network slice

Country Status (1)

Country Link
CN (1) CN111356141B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112738811B (en) * 2021-01-08 2022-06-24 重庆理工大学 Spectrum sharing method for network slices in cognitive capacity collection network
CN114158078B (en) * 2021-12-15 2023-06-16 中国联合网络通信集团有限公司 Network slice management method, device and computer readable storage medium

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106455078A (en) * 2016-10-31 2017-02-22 东南大学 Equilibrium strategy-combined wireless virtual network resource allocation method
CN108599913A (en) * 2018-04-26 2018-09-28 重庆邮电大学 The coexistence method of LTE-U and WiFi under a kind of multi-operator scenario

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106332088B (en) * 2015-06-28 2020-01-17 上海无线通信研究中心 Frequency spectrum sharing method based on user fairness among different operators
CN106549806B (en) * 2016-10-26 2019-06-18 清华大学 A kind of network slice manager and its management method
CN108601087B (en) * 2018-04-27 2022-07-08 哈尔滨工业大学深圳研究生院 Wireless communication resource allocation method based on network slice

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106455078A (en) * 2016-10-31 2017-02-22 东南大学 Equilibrium strategy-combined wireless virtual network resource allocation method
CN108599913A (en) * 2018-04-26 2018-09-28 重庆邮电大学 The coexistence method of LTE-U and WiFi under a kind of multi-operator scenario

Also Published As

Publication number Publication date
CN111356141A (en) 2020-06-30

Similar Documents

Publication Publication Date Title
CN111314889B (en) Task unloading and resource allocation method based on mobile edge calculation in Internet of vehicles
Yi et al. Cooperative communication-aware spectrum leasing in cognitive radio networks
CN111372314A (en) Task unloading method and task unloading device based on mobile edge computing scene
CN111356141B (en) Distributed resource allocation method for inter-operator spectrum sharing based on network slice
CN109121151A (en) Distributed discharging method under the integrated mobile edge calculations of cellulor
CN102752864B (en) User experience-oriented resource allocation method in multi-user and multi-service system
CN102665282B (en) Distribution method of multi-user parallel transmission resource in wireless heterogeneous network
CN111182495B (en) 5G internet of vehicles partial calculation unloading method
CN106954234B (en) User connection and virtual resource allocation method in ultra-dense heterogeneous network
Pham et al. Fairness-aware spectral and energy efficiency in spectrum-sharing wireless networks
CN108632077B (en) Power business data transmission modeling process and transmission channel determination method
CN108832979B (en) Multi-objective optimization resource allocation algorithm for MU-MIMO system in channel under-rank condition
CN108901075A (en) A kind of resource allocation methods based on GS algorithm
CN109618351B (en) Resource allocation method in heterogeneous network based on stackelberg game
CN103079278A (en) Method for allocating downlink resources of OFDMA (Orthogonal Frequency Division Multiple Access)-WLAN (Wireless Local Area Network) system based on user satisfaction degrees
CN107302801B (en) QoE-oriented double-layer matching game method in 5G mixed scene
CN104702394B (en) A kind of power line communication resource allocation methods fair based on service delay
CN108430104A (en) A kind of method and its system of optimized for energy efficiency and resource allocation
CN108616996B (en) Spectrum and power distribution method, device and system for Femtocell network
CN112367523B (en) Resource management method in SVC multicast based on NOMA in heterogeneous wireless network
Kim et al. Resource allocation in cluster based ultra dense network
CN103402265A (en) Spectrum allocation method based on fuzzy logic and communication priority
CN110691383B (en) Resource allocation method and device
CN102984717A (en) Method and system for system capacity estimation
CN110062399B (en) Cognitive heterogeneous cellular network spectrum allocation method based on game theory

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
GR01 Patent grant
GR01 Patent grant