A joint optimization method of base station activation and user access control

Abstract

The application discloses a kind of joint optimization method of base station activation and user access control, the application is activated by joint control of base station and user access relationship, the advantage of two control directions is integrated, both consider the resource consumption problem brought by base station activation, and also consider the access relationship of user, guarantee the normal and reliable operation of system.In the algorithm running process, by setting appropriate parameter value, the relationship of base station transmitting power, the number of activated base station and the number of access user can be well balanced, the access relationship of user is placed in the first consideration position, under the premise of meeting the service quality level required by user, then the activation problem of base station is balanced.In addition, the application introduces ADMM framework, compared with prior art, the complexity of solving optimization problem is reduced, and the same optimization result can be obtained.And in the process of system construction, the interference between users is considered, overcome the problem that channel interference is not considered in prior art.

Description

Technical Field

[0001] This invention discloses a joint optimization method for base station activation and user access control, which relates to the field of wireless communication technology, and particularly to the field of network optimization technology. Background Technology

[0002] User access control is a method for adjusting network resource allocation to avoid network congestion and improve service quality. It is widely used in mobile communications, the Internet, and other fields. Compared to the contention-based relationship of traditional user access, current technologies generally optimize access control in conjunction with power control. Joint optimization of user access control and power control can improve network capacity and coverage while ensuring service quality, achieving effective utilization and allocation of network resources.

[0003] Base station activation control is a demand-based strategy that reduces network overhead and improves network efficiency by dynamically shutting down redundant base stations and activating them only when needed.

[0004] Based on the foregoing, it's clear that user access control and base station activation control are often interdependent, their outcomes affecting the overall network performance. For example, user access control requires consideration of factors such as the number and location of base stations, as well as the coverage area and bandwidth each base station can provide. These factors influence base station activation control decisions. Conversely, inappropriate base station activation control decisions may result in some base stations having insufficient coverage or bandwidth, further impacting user access control decisions and ultimately affecting overall network performance. Therefore, in practical applications, user access control and base station activation control typically require joint optimization.

[0005] In existing technologies, such as Joint Optimization of Base Station Activation and User Association in Ultra Dense Networks Under Traffic Uncertainty, the related technologies that typically require joint optimization of user access control and base station activation control mainly have two problems: The high complexity of joint optimization algorithms and the failure to consider interference between users lead to defects in existing joint optimization methods.

[0006] Content of this invention The purpose of this invention is to provide a joint optimization method for base station activation and user access control, which solves the problems of high complexity and failure to consider interference between users in existing joint optimization algorithms.

[0007] To achieve the above-mentioned technical objectives and effects, the invention is implemented through the following technical solution: A joint optimization method for base station activation and user access control includes: assuming there are K base stations, Each base station is equipped with N antennas, and these base stations work together to provide services to M single-antenna users. ; definition It is the transmission waveform from base station k to user m. It is the joint transmission waveform of all base stations to user m; definition It is the channel vector between base station k and user m. It is the joint channel between all base stations and user m; definition It is the transmit waveform matrix of base station k to all users. It is the set of transmitted waveforms of the system; definition It is the joint channel matrix of base station k for all users. It is the system's set of channel vectors; The hypothetical system is modeled as the following optimization problem: ; in , , |·| represents the cardinality of a set. Let λ represent the transmit power threshold of base station k, λ1 represent the energy required to maintain the normal operation of a base station, and λ2 represent the penalty for rejecting a user's request. This represents the user access indicator factor vector, which is composed of the access indicator factors of M users.

[0008] Furthermore, the signal-to-interference-plus-noise ratio (SIR) at the user end is defined as follows:

[0009] The numerator represents the transmitted signal, and the denominator represents interference, including noise interference and interference from other users. After transforming the above equation using second-order cone programming, we obtain the following:

[0010] This represents the service quality requirements expected by user m. This represents an indicator factor used to indicate whether user m meets the quality of service requirements. This indicates that user m can access the network. This indicates that user m has been denied access. The operation represents the Hermitian transpose of a complex vector or matrix, which involves taking the complex conjugate of each element and then transposing it; ℑ represents the operation of extracting the imaginary part of a complex number. The system should also satisfy the following constraints.

[0011]

[0012]

[0013] Furthermore, a series of approximate substitutions are made to the objective function of the optimization problem, and a global optimal or suboptimal solution is found by iteratively solving a series of subproblems. ;

[0014] ; in ; and This represents the results of a and b in the previous iteration.

[0015] Furthermore, based on the ADMM algorithm framework, the optimization problem becomes as follows: ;

[0016]

[0017] in , . It is the m-th row of matrix U. yes The m-th element, yes 1 x M sub-vectors, i.e.

[0018] Furthermore, the equality constraints from the optimization problem are added to the objective function, and the augmented Lagrangian function is defined in the following form:

[0019] in , , It is a Lagrange multiplier, and ρ>0 indicates a penalty term; b kThis is an auxiliary variable introduced to handle the zero-norm constraint of base station activation, representing the upper bound of the squared Frobenius norm of the transmit power of the k-th base station.

[0020] Furthermore, based on the alternating direction multiplier method of ADMM, the variable {W, U, V, a, b, c} is divided into two parts {W, c} and {U, V, a, b}, and processed separately.

[0021] (1) First, update {W, c}, and then divide it into M subproblems to solve. , , m = 1, 2, .... M.

[0022] renew The subproblems are expressed as follows:

[0023] in, It is the m-th column of U. It is the m-th column of φ. It is the m-th column of Ψ. It is the m-th column of V; This represents the operation of extracting the real part of a complex number, i.e., extracting the real value of the complex number; the above formula is about For a quadratic function, it is easy to find the value that minimizes the objective function. Value, that is

[0024] because It is known in advance and can be calculated ahead of time. Therefore, in each ADMM iteration, the parameters are updated. The complexity is .

[0025] renew The subproblems are expressed as follows:

[0026]

[0027] The above formula is about For a quadratic function, it is easy to find the value that minimizes the objective function. Value, that is

[0028] in Update parameters The complexity is .

[0029] (2) Next, update {U, V, a, b}, which can be divided into M subproblems to be solved. and And solving K subproblems and .

[0030] renew and The subproblems are expressed as follows, defined :

[0031]

[0032]

[0033] in It is a matrix The m-th line, It is a matrix The m-th row. The above problem is a quadratic constrained quadratic programming problem. According to the KKT conditions, the analytical solution to the above problem can be obtained as follows: if

[0034] So

[0035]

[0036]

[0037] otherwise

[0038] in This represents the Lagrange multipliers. (Update) and The complexity is renew and The subproblems are expressed as follows:

[0039]

[0040] Based on the KKT conditions, the analytical solution to the above problem can be obtained as follows:

[0041] in This represents the Lagrange multipliers. (Update) and The complexity is (3) Last updated

[0042] Beneficial effects: By jointly controlling base station activation and user access relationships, the advantages of both control methods are combined. This approach considers both the resource consumption associated with base station activation and the user access relationships, ensuring the normal and reliable operation of the system. During algorithm operation, by setting appropriate parameter values, the relationship between base station transmission power, the number of activated base stations, and the number of access users can be effectively balanced. User access relationships are given primary consideration, and the base station activation issue is addressed only after meeting the service quality requirements of users.

[0043] Furthermore, this invention introduces the ADMM framework, which, compared to the existing technology using the CVX toolkit to solve optimization problems, reduces the complexity of solving optimization problems because each step of the ADMM framework algorithm can yield an analytical solution, while still achieving the same optimization results. Moreover, the system construction process considers interference between users, overcoming the problem of existing technologies not considering channel interference.

[0044] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0045] Figure 1 This is a logic diagram of a joint optimization method for base station activation and user access control according to an embodiment of the present invention; Figure 2 This is a diagram of the joint optimization model for base station activation and user access control as described in an embodiment of the present invention. Detailed Implementation

[0046] To more clearly illustrate the technical solutions of the embodiments of the present invention, the present invention will be described in detail below with reference to the accompanying drawings.

[0047] Example To improve overall system performance, the system should provide users with a QoS (Quality of Service) level that meets their needs. Based on this, the system needs to shut down as many base stations as possible to reduce resource consumption. Secondly, if all base stations are activated, the system still cannot provide services to all users. In this case, the system needs to reject as few users as possible to ensure service is available to the remaining users.

[0048] To achieve the aforementioned objectives, this invention first establishes a problem model. Specifically, consider a downlink wireless communication system, such as... Figure 1 As shown, a joint optimization method for base station activation and user access control is proposed: assuming there are K base stations, Each base station is equipped with N antennas, and these base stations work together to provide services to M single-antenna users. ; definition It is the transmission waveform from base station k to user m. It is the joint transmission waveform of all base stations to user m; definition It is the channel vector between base station k and user m. It is the joint channel between all base stations and user m; definition It is the transmit waveform matrix of base station k to all users. It is the set of transmitted waveforms of the system; definition It is the joint channel matrix of base station k for all users. It is the system's set of channel vectors; The hypothetical system is modeled as the following optimization problem:

[0049] in , , |·| represents the cardinality of a set. Let λ represent the transmit power threshold of base station k, λ1 represent the energy required to maintain the normal operation of a base station, and λ2 represent the penalty required to reject a user.

[0050] The applicant has considered interference between users. This technology uses the signal-to-interference-plus-noise ratio (SINR) to represent the user-end service quality. The user-end SINR is defined as follows:

[0051] The numerator represents the transmitted signal, and the denominator represents interference, including noise interference and interference from other users. After transforming the above equation using second-order cone programming, we obtain the following:

[0052] This represents the service quality requirements expected by user m. This represents an indicator factor used to indicate whether user m meets the quality of service requirements. This indicates that user m can access the network. This indicates that user m has been denied access. Additionally, the system must satisfy the following constraints.

[0053]

[0054]

[0055] In response to the foregoing, this embodiment performs a series of approximate substitutions on the objective function of the optimization problem, and finds a global optimal or suboptimal solution by iteratively solving a series of subproblems; ;

[0056] ; in ; and This represents the results of a and b in the previous iteration.

[0057] The approximate substitution specifically refers to: When dealing with the zero norm in the objective function, define a continuous function.

[0058] in Given a very large positive number, replacing the zero norm in the objective function with this function yields the following objective function.

[0059] Introducing new variables Replace the above formula The following objective function is obtained.

[0060] And add new constraints.

[0061] The final approximation is the following optimization problem.

[0062]

[0063]

[0064] Furthermore, although the above approximation solves the zero-norm problem, the optimization problem remains non-convex. Based on the continuous upper bound minimization algorithm, the optimization problem is decomposed into solvable sub-convex problems, and the globally stable solution of the original problem is found iteratively. The function is in Performing a first-order Taylor expansion at the given point yields the following equation:

[0065] Substituting the above equation into the approximated objective function yields the subconvex problem.

[0066] Compared with the prior art, the technology in this embodiment takes into account the interference between users, thereby improving the reliability of the algorithm.

[0067] Example 2

[0068] In practice, the algorithms described above are still quite complex and place a heavy load on the control system. In this embodiment, the applicant introduces the ADMM algorithm framework.

[0069] Specifically, when solving subconvex problems, variables are introduced. , , The purpose is to decouple the original optimization variables in the objective function based on the ADMM algorithm framework.

[0070] The solution complexity of the subconvex problem using a centralized solution method is O(n). For large-scale wireless networks, the computational cost is very high. The ADMM low-complexity algorithm can decompose the original problem into multiple subproblems, each of which can be solved analytically, thus significantly reducing the problem's complexity. The specific update steps are as follows: (1) First, update {W, c}, and then divide it into M subproblems to solve. , , m = 1, 2, .... M.

[0071] renew The subproblems are expressed as follows:

[0072] in, It is the m-th column of U. It is the m-th column of φ. It is the m-th column of Ψ. It is the m-th column of V. The above formula is about For a quadratic function, it is easy to find the value that minimizes the objective function. Value, that is

[0073] because It is known in advance and can be calculated ahead of time. Therefore, in each ADMM iteration, the parameters are updated. The complexity is .

[0074] renew The subproblems are expressed as follows:

[0075]

[0076] The above formula is about For a quadratic function, it is easy to find the value that minimizes the objective function. Value, that is

[0077] in Update parameters The complexity is .

[0078] (2) Next, update {U, V, a, b}, which can be divided into M subproblems to be solved. and And solving K subproblems and .

[0079] renew and The subproblems are expressed as follows, defined :

[0080]

[0081]

[0082] in It is a matrix The m-th line, It is a matrix The m-th row. The above problem is a quadratic constrained quadratic programming problem. According to the KKT conditions, the analytical solution to the above problem can be obtained as follows: if

[0083] So

[0084]

[0085]

[0086] otherwise

[0087] in This represents the Lagrange multipliers. (Update) and The complexity is renew and The subproblems are expressed as follows:

[0088]

[0089] Based on the KKT conditions, the analytical solution to the above problem can be obtained as follows:

[0090] in This represents the Lagrange multipliers. (Update) and The complexity is (3) Last updated

[0091] Overall algorithm control logic reference Figure 1 To understand.

[0092] The above are merely some embodiments of this application and are not intended to limit the application in any way. Any simple modifications, equivalent changes, and alterations made to the above embodiments shall still fall within the scope of protection of the technical solution of this application.

Claims

1. A joint optimization method for base station activation and user access control, characterized in that, Assume there are K base stations, Each base station is equipped with N antennas, and these base stations work together to provide services to M single-antenna users. ; definition It is the transmission waveform from base station k to user m. It is the joint transmission waveform of all base stations to user m; definition It is the channel vector between base station k and user m. It is the joint channel between all base stations and user m; definition It is the transmit waveform matrix of base station k to all users. It is the set of transmitted waveforms of the system; definition It is the joint channel matrix of base station k for all users. It is the system's set of channel vectors; The hypothetical system is modeled as the following optimization problem: ； in , , |·| represents the cardinality of a set. Let λ1 represent the transmit power threshold of base station k, λ2 represent the energy required to maintain the normal operation of a base station, and λ3 represent the penalty required to reject a user. The signal-to-interference-plus-noise ratio (SIR) on the user side is defined as follows: The numerator represents the transmitted signal, and the denominator represents interference, including noise interference and interference from other users. After transforming the above equation using second-order cone programming, we obtain the following: This represents the service quality requirements expected by user m. This represents an indicator factor used to indicate whether user m meets the quality of service requirements. This means that user m can access the network. This indicates that user m has been denied access; in addition, the system should also meet the following constraints: By making a series of approximate substitutions to the objective function of the optimization problem and iteratively solving a series of subproblems, a globally optimal or suboptimal solution can be found. ； ； in ； and This represents the results of a and b in the previous iteration; Based on the ADMM algorithm framework, the optimization problem becomes as follows: ； in , It is the m-th row of matrix U. yes The m-th element, yes 1 x M sub-vectors, i.e. 。 2. The joint optimization method for base station activation and user access control according to claim 1, characterized in that, The equality constraints from the optimization problem are added to the objective function, and the augmented Lagrangian function is defined in the following form: in , , It is a Lagrange multiplier, and ρ>0 indicates a penalty term.

3. The joint optimization method for base station activation and user access control according to claim 2, characterized in that, According to the alternating direction multiplier method of ADMM, the variable {W, U, V, a, b, c} is divided into two parts {W, c} and {U, V, a, b}, and they are processed separately. (1) First, update {W, c}, and then divide it into M subproblems to solve. , m = 1, 2, .... M; renew The subproblems are expressed as follows: in, It is the m-th column of U. It is the m-th column of φ. It is the m-th column of Ψ. It is the m-th column of V; the above formula is about For a quadratic function, it is easy to find the value that minimizes the objective function. Value, i.e. because It is known in advance and can be calculated ahead of time. Therefore, in each ADMM iteration, the parameters are updated. The complexity is ; renew The subproblems are expressed as follows: The above formula is about For a quadratic function, it is easy to find the value that minimizes the objective function. Value, that is in Update parameters The complexity is ; (2) Next, update {U, V, a, b}, which can be divided into M subproblems to be solved. and And solving K subproblems and ; renew and The subproblems are expressed as follows, defined : in It is a matrix The m-th line, It is a matrix The m-th row; the above problem is a quadratic constrained quadratic programming problem. According to the KKT conditions, the analytical solution to the above problem can be obtained as follows: if So otherwise in Represents Lagrange multipliers; Update and The complexity is renew and The subproblems are expressed as follows: Based on the KKT conditions, the analytical solution to the above problem can be obtained as follows: in Represents Lagrange multipliers; Update and The complexity is (3) Last updated

Images