electronics

Article
Joint User Scheduling and Resource Allocation in Distributed
MIMO Systems with Multi-Carriers
Yinglan Bu 1 , Jiaying Zong 2 , Xinjiang Xia 3, *, Yang Liu 2 , Fengyi Yang 2 and Dongming Wang 3, *

1 National Mobile Communications Research Laboratory, Southeast University, Nanjing 210096, China;
[email protected]
2 China Telecom Research Institute, Beijing 102209, China; [email protected] (J.Z.);
[email protected] (Y.L.); [email protected] (F.Y.)
3 Purple Mountain Laboratories, Nanjing 211111, China
* Correspondence: [email protected] (X.X.); [email protected] (D.W.)

Abstract: Compared with the traditional collocated multi-input multi-output system (C-MIMO),
distributed MIMO (D-MIMO) systems have the advantage of higher throughput and coverage,
making them strong candidates for next-generation communication architecture. As a practical
implementation of a D-MIMO cooperative network, the multi-TRP (multiple transmission/reception
point) system becomes a hotspot in the research of advanced 5G. Different from previous research
on a cooperative D-MIMO network with single narrowband transmission, this paper proposes a
joint optimization scheme to address the user scheduling problem along with carrier allocation to
maximize the total spectral efficiency (SE) in the downlink of coherent multi-TRP systems with
multi-carriers. We establish a joint optimization model of user scheduling and resource allocation
to maximize the system spectral efficiency under the constraints of power consumption and the
backhaul capacity limits at each RAU (remote antenna unit), as well as the QoS (quality of service)
requirement at each user. Since the optimization model is both non-covex and non-smooth, a joint
optimization algorithm is proposed to solve this non-convex combinatorial optimization problem.
Citation: Bu, Y.; Zong, J.; Xia, X.; Liu, We first smooth the mixed-integer problem by employing penalty functions, and after decoupling
Y.; Yang, F.; Wang, D. Joint User the coupled variables by introducing auxiliary variables, the original problem is transformed into a
Scheduling and Resource Allocation series of tractable convex optimization problems by using successive convex approximation (SCA).
in Distributed MIMO Systems with Numerical results demonstrate that the proposed joint optimization algorithm for user scheduling
Multi-Carriers. Electronics 2022, 11, and resource allocation can reliably converge and achieve a higher system SE than the general
1836. https://doi.org/10.3390/ multi-TRP system without carrier allocation, and this advantage is more pronounced under a higher
electronics11121836
backhaul capacity or higher power consumption constraints.
Academic Editor: Luis Castedo
Keywords: D-MIMO system; multi-carrier system; resource allocation; user scheduling; SCA
Received: 29 April 2022
Accepted: 7 June 2022
Published: 9 June 2022

Publisher’s Note: MDPI stays neutral 1. Introduction

with regard to jurisdictional claims in
The improvement of data rate is always the focus of attention in the development of
published maps and institutional affil-
wireless communication. Fifth-generation mobile communication technology, which has
iations.
been deployed and implemented on a large scale, can achieve a peak data rate of 20 Gbps
but has yet to be able to meet the higher demand of user-experienced data rates for 2030+
in the future [1].
Copyright: © 2022 by the authors.
Looking back on the past few decades, the vast majority of increases in communication
Licensee MDPI, Basel, Switzerland. capacity came from the densification of network infrastructure [2]. However, limited by the
This article is an open access article cell edge effect, inter-cell interference will gradually dominate as the distance between base
distributed under the terms and stations decreases and will seriously effect the user-experienced data rates [3]. The D-MIMO
conditions of the Creative Commons proposed in [4] had a distributed architecture, which can effectively manage the inter-cell
Attribution (CC BY) license (https:// interference and thereby improve the network spectral efficiency and coverage [5,6].
creativecommons.org/licenses/by/ In the D-MIMO system, multiple RAUs are distributed widely in the coverage area
4.0/). and connected to the baseband processing unit (BPU) via backhauls. They cooperate to

Electronics 2022, 11, 1836. https://doi.org/10.3390/electronics11121836 https://www.mdpi.com/journal/electronics

Electronics 2022, 11, 1836 2 of 16

serve users by transmitting and receiving signals processed collaboratively at the central
BPU, for which it can achieve higher spatial diversity/multiplexing gain [7–10].
In previous studies, user scheduling (namely user association) has always been an
important research direction to improve the system efficiency in cooperation communi-
cations [11–13]. A dynamic cooperation clustering (DCC) model was proposed in [14],
which constructed user-centric clustering association by setting antenna selection diagonal
matrices. Considering the difference in channel quality between different users and RAUs
over different carriers, the best RAUs and carriers can be matched for each user through
joint optimization at the BPU, and multi-user interferences can be reduced as well.
A similar scheme, named multistream carrier aggregation (MSCA), initially appeared
for heterogeneous networks, which aggregates component carriers belonging to multiple
base stations for users to maximize the available configuration, thereby, improving the
system data rate of service [15–17]. In [15,16], the optimization problem of energy efficiency
in MSCA systems was considered by jointly investigating the user association, channel
allocation and also the power allocation problems. Ref. [17] proposed a low-complexity
algorithm to determine precoding vectors for MSCA systems and an efficient scheme to
assign carriers to particular users based on their throughput requirements and the purpose
of minimizing the total feedback overhead as well.
In addition, the discussion of resource allocation also appears in orthogonal frequency-
division multi- plexing (OFDM) systems [18–20]. Ref. [18] aimed at energy-efficient power
and channel allocation, along with user association optimization in OFDM systems. In [19],
an iterative beamforming and scheduling strategy was proposed to maximize the system
weighted sum rate, combined with the heuristic Hungarian algorithm for channel allocation
on a fixed precoding pattern. However, Refs. [18,19] only considered the coordinated
interference management but not the cooperative service of multiple base stations.
Most of the existing resource allocation algorithms of the distributed cooperative
network do not consider multi-carrier scenarios in a coherent way or only optimize the
system performance with single narrowband transmission. Though when compared with
non-coherent transmissions, the coherent transmission requires strict phase synchronization
across all RAUs serving a user, it can actually achieve higher system spectral efficiency
with lower decoding complexity, since the non-coherent transmission needs no phase
synchronization but successive interference cancellation at users to decode the individual
streams [21,22].Compared with incoherent cooperative networks, coherent cooperative
networks can achieve higher system spectral efficiency [21]. However, the current research
on the joint matching of users, RAUs and carriers is still confined to the traditional cellular
network architecture, with restrictions on carrier selection or in a non-coherent way.
There is currently no complete resource allocation solution for a widely coordinated
D-MIMO network with multi-carriers. Reasonable resource allocation and user-scheduling
schemes for the coherent multi-carrier D-MIMO system are significant to improve the total
spectral efficiency of D-MIMO systems. As one of the features of enhanced massive MIMO,
multiple transmission and reception point (multi-TRP) can be regarded as a practical
implementation of cooperative D-MIMO. In the established Release18 work package,
coherent joint transmission in multi-TRP scenarios and two timing advances are listed as
important evolution directions as well [23–25].
User scheduling and carrier allocation can be jointly optimized since they both belong
to integer programming. Under constraints of backhaul capacity and power consumption
budgets, this paper establishes a joint optimization model to solve the resource allocation
and user scheduling problem in multi-TRP systems and assign appropriate RAUs and
carriers to each user in a user-centric manner by setting integer matching variables. With
the coherent and collaborative service provided by RAUs, the user-RAU-carrier matching
variables are highly coupled in the SINR expression, making the model a non-convex and
non-smooth optimization problem.
We first smooth the original problem by using penalty functions and introduce auxil-
iary variables to deal with the coupled variables. A concave lower bound of the objective
Electronics 2022, 11, 1836 3 of 16

function is derived then, and the original problem is transformed into a series of iterative
convex optimization problems by approximating the non-convex constraints with the suc-
cessive convex approximation (SCA) method. Specifically, the contributions of this work
are summarized as follows.
1. This paper comprehensively considers resource allocation and user scheduling for co-
herent multi-TRP systems with multi-carriers and generates a radio resource allocation
scheme for coherently cooperative networks. We develop a mixed-integer program-
ming model that jointly optimizes the user-RAU association and carrier allocation,
as well as the downlink power allocation problems to maximize the system spectral
efficiency under constraints of backhaul capacity and power consumption limits.
2. Owing to the NP-hardness of mixed-integer programming, we transform the original
non-smooth non-convex optimization problem into a series of iterative convex opti-
mization problems through penalty functions and the SCA method, and then a joint
optimization algorithm of user scheduling and resource allocation for the coherent
multi-TRP system with multi-carriers is proposed, of which the superiority in system
spectral efficiency is verified by numerical results compared to the general multi-TRP
system without channel selection.
The reminder of this paper is organized as follows. In Section 2, we give a mathemati-
cal description of the system spectral efficiency and establish a corresponding performance
optimization model. In Section 3, the original optimization problem is analyzed and trans-
formed into a tractable form, and a joint user scheduling and resource allocation algorithm
is proposed. Numerical results and corresponding analysis are shown in Section 4, and the
final conclusion is given in Section 5.

2. System Model and Problem Formulation

As shown in illustration Figure 1, the multi-TRP system with multi-carriers that
we consider includes L single antenna RAUs, indexed by l ∈L,{1, . . ., L}, cooperatively
serving K single antenna users, indexed by k∈K,{1, . . ., K }, in a coherent way. Data signals
sent by RAUs and their corresponding transmission power are uniformly determined by
the central BPU, and each RAU is connected to the BPU through backhaul links. There are
N available carriers in this network, indexed by n∈N ,{1, . . ., N }, which are reasonably
assumed orthogonal to each other. Each carrier can be allocated to any user-RAU pair
freely, and the specific association with RAUs and carriers of each user is scheduled by the
BPU as well.
Although centralized processing will bring more computational complexity overhead
to the BPU who has full knowledge of channel state information (CSI), it consequently has a
better performance gain. Since RAUs are widely distributed over the coverage area in the D-
MIMO system, the distance between users and RAUs is correspondingly shortened, which
thereby causes better and more uniform channel conditions than that in co-located MIMO
systems. We assume that the channel over each carrier between RAUs and users obeys
flat fading, which means the CSI remains constant during the coherent time. Considering
the tradeoff between computational complexity and system performance, we adopt the
conjugate beamforming scheme as in [26], and the corresponding signal received at user k
over carrier n is
√
pk0 ,n ρl,k0 ,n h∗l,k0 ,n sk0 ,n
yk,n = ∑ hl,k,n ∑ + vk,n
l ∈L k0 ∈K hl,k0 ,n
√ 0√
= ∑ λk,kl,n pk,n ρl,k,n sk,n + ∑ ∑ λk,k l,n pk0 ,n ρl,k0 ,n sk0 ,n + vk,n , (1)
l ∈L k0 6=k,k0 ∈K l ∈L

k,k 0 hl,k,n h∗l,k0 ,n

where λl,n = , hl,k,n ∈C is the channel between RAU l and user k on carrier n.
|hl,k0 ,n |
vk,n is the
complex additive white Gaussian noise (AWGN) following the distribution of
CN 0, σ2 received at user k over carrier n. sk,n denotes the symbol intended for user k over
Electronics 2022, 11, 1836 4 of 16

carrier n, and pk,n ∈R expresses the downlink

 transmission power allocated for symbol sk,n
at RAUs. It is reasonably assumed that sk,n |k, n∈K, N are independent
  and identically
2
distributed variables, which are zero mean and normalized as E sk,n = 1. ρl,k,n ∈{0, 1}
denotes the binary matching variable, which follows
(
1, if RAU l allocates carrier n to user k ,
ρl,k,n = (2)
0, otherwise.

Figure 1. Illustration of the multi-TRP system with multi-carriers.

The received signal in expression (1) is illustrated as two parts, the desired signal
k,k √ k,k0 √
∑ l,n pk,n ρl,k,n sk,n , and the interference-plus-noise term
λ ∑ ∑ λl,n pk0 ,n ρl,k0 ,n sk0 ,n +
l ∈L k0 6=k,k0 ∈K l ∈L
vk,n . Clearly they are uncorrelated with each other, and thereby the signal- to-interference-
plus-noise ratio (SINR) for user k over carrier n is given by

H Λk,k ρ
pk,n ρk,n n k,n
γk,n = 0 , (3)
∑ pk0 ,n ρkH0 ,n Λk,k
n ρk0 ,n + σ
2
k0 6=k,k0 ∈K

where ρk,n = [ρ1,k,n , ρ2,k,n , . . .,hρ L,k,n ] T is the collective vector of matching variables for user
0 0 0 0 0
i
k on carrier n. Let λk,k
n = λ1,nk,k k,k
, λ2,n , . . ., λk,k k,k
L,n , Λn is a positive semidefinite matrix
0 0 H k,k 0
 
formulated by Λk,k n = λk,k
n λn . Unlike previous studies on joint user scheduling
and carrier allocation, both the numerator and denominator of Equation (3) contain a
quadratic form consisting of matching variables, because with the coherent transmission
all associated RAUs will transmit the same symbol to a user. Therefore, the achievable
SE of user k over carrier n can be written as Rk,n = ln(1 + γk,n ). According to the OFDM
transmission framework, the total achievable SE at user k across all available carriers is
given by Rk,total = ∑ Rk,n .
n∈N
As can be observed in Equation (3), the SINR of each user over each carrier is deter-
mined by both the corresponding power allocation for downlink transmission at RAUs and
the user-RAU-carrier three-sided matching relationships. Considering the requirement of
further improving the data rate with limited frequency band resources in the future, the
system spectral efficiency, which is also one of the important performance indicators in
communication networks, is employed as the utility function for the optimization problem
that we consider.
We denote Ω as a L×K × N dimensional matrix gathering elements of ρl,k,n and P as a
K × N dimensional matrix gathering elements of pk,n . Accordingly, the mathematical model
Electronics 2022, 11, 1836 5 of 16

of the joint user scheduling and resource allocation optimization problem for a single time
slot is formulated by

maximize
Ω,P
∑ ∑ Rk,n (4a)
k ∈K n∈N
subject to C1: ∑ Rk,n ≥ Rmin,k , ∀k∈K , (4b)
n∈N
C2: ∑ ∑ pk,n ρ2l,k,n ≤ Pmax,l , ∀l ∈L , (4c)
n∈N k∈K
C3: ∑ ∑ ρl,k,n Rk,n ≤Cmax,l , ∀l ∈L , (4d)
n∈N k∈K
C4: ρl,k,n ∈{0, 1} , ∀l, k, n∈L, K, N . (4e)

In the above optimization problem (4), constraint C1 guarantees the QoS requirement
for each user, and C2 and C3 restrict the maximum power consumption and hackhaul
budget across all frequency bands at each RAU to Pmax,l and Cmax,l , respectively. The binary
nature of the matching variable ρl,k,n is ensured in C4.
Problem (4) is actually a non-convex combinatorial optimization problem, and its
three-sided matching feature makes this NP-hard problem more difficult to solve. Thus,
we propose an iterative optimization algorithm which transforms the original problem into
a tractable form.

3. Problem Analysis and Proposed Approach

Since variable ρk,n and pk,n are coupled together inside both the numerator and
denominator of the SINR expression, the original problem is actually a complex non-convex
optimization problem with combinatorial variables. Referring to [20], we first smooth the
original problem by setting penalty functions, and auxiliary variables are introduced to
disassemble the expression of coupled variables.
After obtaining the concave lower bound of the objective function, the SCA method is
employed to transform the original non-convex problem into a series of iterative convex op-
timization problems, which can be easily solved. The main idea of our proposed approach
is to find corresponding concave or convex approximations for the objective function and
constraints, which contain parameters that have a mapping relationship with the results of
the previous iteration. These parameters are constantly updated in the iterative process of
solving the approximate convex optimization problem, until the solution of each iteration
finally converge.

3.1. Equivalent Reformulation of Binary Constraints

The integer variable in problem (4) makes it a NP-hard problem that cannot be solved
in polynomial time. According to [27], the original mixed-integer optimization problem
can be transformed into a continuous form by utilizing penalty functions. We then rewrite
constraint C4 as
C5: ρl,k,n ∈[0, 1], ∀l, k, n∈L, K, N , (5)
C5 redefines ρl,k,n as a continuous variable varying within [0, 1]. To ensure the binary
feature of the original matching variable ρl,k,n , penalty terms are added to the objective
function, which guarantees the equivalence of the optimization problems before and after
relaxation of integer variables. Referring to [27], the employed penalty function is set as

L K N
f (Ω) = − ∑ ∑ ∑ (ln(ρl,k,n + e) + ln((1 − ρl,k,n ) + e)) , (6)
l =1 k =1 n =1
Electronics 2022, 11, 1836 6 of 16

where e is a very small positive constant, which can make the continuous optimization
problem after transformation consistent with the original integer programming by properly
setting its value. The equivalent optimization problem is then transformed into

maximize
Ω,P
f (Ω) + ∑ ∑ Rk,n
k ∈K n∈N
subject to C1,C2,C3, (7)

C5: ρl,k,n ∈[0, 1] , ∀l, k, n∈L, K, N .

3.2. Approximation of the Objective Function

The objective function of problem (7) comprises two parts, the achievable SE expression
involving SINR γk,n and the penalty function f (Ω), making this maximization objective
function non-concave and difficult to solve. Referring to the path-following optimization
algorithm in [28], the original objective function can be approximated to its concave lower
bound for an iterative solution. By taking the first-order Taylor expansion of the penalty
function at feasible point Ω(m) , the following inequality can be obtained that

L K N     
f (Ω) ≥ fˆ(Ω) = − ∑ ∑∑
(m) (m)
ln ρl,k,n + e + ln 1 − ρl,k,n + e
l =1 k =1 n =1
  !
1 1 
 ρl,k,n − ρ(m)

+ (m)
− (m)
 l,k,n . (8)
ρl,k,n + e 1 − ρl,k,n + e

This inequality relation holds because the penalty function f (Ω) is actually a concave
function, and the linear function obtained by the first-order Taylor expansion at any feasible
point of a concave function is greater than itself. Such inequality relation is exactly reversed
for convex functions, as we shall see later. The objective function in problem (7) can then be
approximated by its lower bound

f obj = fˆ(Ω) + ∑ ∑ Rk,n , (9)

k ∈K n∈N

with fˆ(Ω) as an affine substitute for the original penalty function f (Ω). We further intro-
duce two groups of auxiliary variables xk,n | xk,n ∈R, k, n∈K, N and yk,k0 ,n |yk,k0 ,n ∈R+ ,
k, k0 , n∈K, K, N } to obtain a lower bound approximation of Rk,n in the objective function:
 
2
xk,n
Rk,n ≥ R̄k,n = ln1 + , (10)
 
∑ yk,k0 ,n + σ2
k0 6=k,k0 ∈K

where xk,n and yk,k0 ,n satisfy

H
pk,n ρk,n Λk,k 2
n ρk,n ≥ xk,n , ∀k, n∈K, N , (11)
0 0
pk0 ,n ρkH0 ,n Λk,k
n ρk0 ,n ≤ yk,k0 ,n , ∀k, k , n∈K, K, N . (12)

After simple transposition of terms, inequalities (11) and (12) are turned into
2
xk,n
H
ρk,n Λnk,k ρk,n ≥ , (13)
pk,n
0
ρkH0 ,n Λk,k
n ρk0 ,n 1
≤ . (14)
yk,k0 ,n pk0 ,n
Electronics 2022, 11, 1836 7 of 16

However, after introducing additional variables, the lower bound expression R̄k,n is
still non-concave. Referring to [28,29], the following inequalities will be frequently used in
the derivation below.

!
| b |2  
ln 1 + ≥ F a, b, a(m) , b(m)
a
 2
 2 n ∗ o 2 
(m) (m) (m) b (m) 2
b b R b b | b | + a
= ln1 + (m)  − (m) + 2 − , (15)
  
a a a(m) 2
a(m) a(m) + b(m)
 
a − a (m)
1  1 
≥ F̄ a, a(m) = (m) −
2 , (16)
a a a(m)
   H  H  
c H Dc ≥ F̂ c, c(m) , D = c(m) Dc(m) + 2R c(m) D c − c(m) . (17)

b∈C and a∈R+ in (15) and (16) are complex and positive real variables, respectively,
while b(m) ∈C and a(m) ∈R+ are constants. In (17), c is a complex variable vector, and
D = dd H is a positive semidefinite matrix, where d is any complex constant vector.
According to the inequality (15), a concave lower bound of R̄k,n can be written as
!
∑ ∑
(m) (m)
R̄k,n ≥ F yk,k0 ,n + σ2 , xk,n , yk,k0 ,n + σ2 , xk,n , (18)
k0 6=k,k0 ∈K k0 6=k,k0 ∈K

which satisfies
 
(m) 2
 
xk,n
!
∑0 ∑0
(m) (m) (m) (m)
yk,k0 ,n + σ2 , xk,n , yk,k0 ,n + σ2 , xk,n
 
F = ln
1 + (m)
. (19)
k0 6=k,k ∈K k0 6=k,k ∈K ∑ yk,k0 ,n +σ 2
0 0
k 6=k,k ∈K

Similarly, following from the inequalities (17) and (16), respectively, inequalities (13) and (14)
can then be approximated by
2
xk,n  
(m)
C6: ≤ F̂ ρk,n , ρk,n , Λnk,k , ∀k, n∈K, N , (20)
pk,n
0
ρkH0 ,n Λk,k
n ρk0 ,n

(m)

C7: ≤ F̄ pk0 ,n , pk0 ,n , ∀k, k0 , n∈K, K, N . (21)
yk,k0 ,n
Since both the LHS of (20) and (21) are typical convex quadratic-over-linear functions,
and the RHS of (20) and (21) are linear functions, it is clear that C6 and C7 are both convex
constraints. By now, the objective function in problem (7) has already been approximated
into a convex form, and the (m + 1)-th iteration can be written as
!
∑ ∑ F 0 ∑0 yk,k0 ,n + σ , xk,n , 0 ∑0 yk,k0 ,n + σ , xk,n + fˆ(Ω)
2 (m) 2 (m)
maximize
Ω,P,
{xk,n ,y 0 |k,k0 ,n∈K,K,N }
k ∈K n∈N k 6=k,k ∈K k 6=k,k ∈K (22)
k,k ,n

subject to C1,C2,C3,C5,C6,C7.

3.3. Convex Approximation of Constraints Involving Variable Products

Despite the above transformation of the objective function, the non-convex constraints
C1, C2 and C3 still make problem (22) an intractable one. To tackle this issue, the SCA
Electronics 2022, 11, 1836 8 of 16

method is utilized to approximate the non-convex constraints further. Again by employing

the inequality (15), constraint C1 can be approximated by
!
∑ ∑0 ∑0
(m) (m)
C8: F yk,k0 ,n + σ2 , xk,n , yk,k0 ,n + σ2 , xk,n ≥ Rmin,k , ∀k∈K , (23)
n∈N k0 6=k,k ∈K k0 6=k,k ∈K

which is found to be a convex constraint due to the concavity of the LHS (left-hand side) of
this inequality.
 Owing to the existence of the variable product in the LHS of C2, auxiliary variables
αl,k,n |αl,k,n ∈R+ , l, k, n∈L, K, N need to be introduced again to deal with the coupled
variables, and the function pk,n ρ2l,k,n in C2 is upper bounded by

pk,n ρ2l,k,n ≤ αl,k,n , ∀l, k, n∈L, K, N . (24)

Thus, the original C2 can be approximated by

C9: ∑ ∑ αl,k,n ≤ Pmax,l , ∀l ∈L . (25)

n∈N k ∈K

Still we need to transform the inequality (24) due to its non-convexity. After simple
transposition of terms, (24) can be rewritten as

ρ2l,k,n 1
≤ , (26)
αl,k,n pk,n

and both of its left- and right-hand sides (RHS) are convex functions.
 Following from
(m)
the inequality (16), the RHS of (26) can be lower bounded by F̄ pk,n , pk,n , which means
that (26) can be approximated by

ρ2l,k,n 
(m)

C10: ≤ F̄ pk,n , pk,n , ∀l, k, n∈L, K, N . (27)
αl,k,n

Since both sides of (25) are linear functions, and (27) has the same structure as C6, it is
obvious that C9 and C10 are convex constraints as well.
Similar to C2, due to the existence of coupled variables, it is necessary to introduce
auxiliary variables tk,n |tk,n ∈R+ , k, n∈K, N as well. Here, we have
 
H Λk,k ρ
pk,n ρk,n
 n k,n 
≤tk,n , ∀k, n∈K, N ,
1 +
ln 0 (28)
∑ pk0 ,n ρk0 ,n Λk,k
H
n ρk0 ,n +σ 2
k0 6=k,k0 ∈K

to approximate the constraint C3 by

∑ ∑ ρl,k,n tk,n ≤Cmax,l , ∀l ∈L . (29)

n∈N k ∈K

However, both of (28) and (29) are yet to be convex, and thus further approximations
are needed. The inequality (28) is equivalent to

H Λk,k ρ
pk,n ρk,n n k,n
0 ≤etk,n − 1 . (30)
∑ pk0 ,n ρkH0 ,n Λk,k
n ρk0 ,n + σ2
k0 6=k,k0 ∈K
Electronics 2022, 11, 1836 9 of 16

Since
 both the numerator and denominator in (30) contain the product of coupled vari-
ables, x̄k,n | x̄k,n ∈R, k, n∈K, N and ȳk,k0 ,n |ȳk,k0 ,n ∈R+ , k, k0 , n∈K, K, N are introduced


as auxiliary variables again, which satisfy

H
pk,n ρk,n Λk,k 2
n ρk,n ≤ x̄k,n , ∀k, n∈K, N , (31)
0 0
pk0 ,n ρkH0 ,n Λk,k
n ρk0 ,n ≥ ȳk,k0 ,n , ∀k, k , n∈K, K, N . (32)

When the following inequality is satisfied,

2
x̄k,n
≤etk,n − 1 , (33)
∑ ȳk,k0 ,n + σ2
k0 6=k,k0 ∈K

the inequality (30) is also satisfied. Therefore, the LHS of (30) can be approximated by
2
x̄k,n
its convex upper bound function , and after taking the first-order Tay-
∑ ȳk,k0 ,n + σ2
k0 6=k,k0 ∈K
(m)
lor expansion of the RHS of (33) at a feasible point tk,n , the inequality (28) can then be
approximated to
2
x̄k,n (m) (m)  
tk,n tk,n (m)
C11: ≤ e + e t k,n − t k,n −1, ∀k, n∈K, N . (34)
∑ ȳk,k0 ,n + σ 2
k0 6=k,k0 ∈K

Inevitably, inequalities (31) and (32) also need to be transformed since they are both
non-convex. After transposition, inequalities (31) and (32) are turned into
2
x̄k,n
H
ρk,n Λk,k
n ρk,n ≤ , (35)
pk,n
0
ρkH0 ,n Λk,k
n ρk0 ,n 1
≥ . (36)
ȳk,k0 ,n pk0 ,n

Both of the left- and right-hand sides of (35) are convex functions, and thus an affine
lower bound of the
 RHS of  (35) can be obtained by taking its first-order Taylor expansion
(m) (m)
at feasible point x̄k,n , pk,n , and the inequality (35) can then be approximated by

(m) 2
 
(m) x̄k,n
H
2x̄k,n
C12:ρk,n Λk,k
n ρk,n ≤ x̄k,n −  ( pk,n ) , ∀k, n∈K, N . (37)
(m) (m) 2

pk,n pk,n

In (36), both the left- and right-hand sides are convex functions as well. Referring
to [29], the following inequality can be used to find an affine lower bound of the LHS
of (36):
 H  
( m )
H
2R c Dc c(m) Dc(m) a
c H Dc  
≥ F̃ c, c(m) , a, a(m) , D = −
2 , (38)
a a(m) a(m)

where a∈R+ is a positive real variable, and a(m) ∈R+ is a constant. c denotes a complex
variable vector, and D = dd H represents a positive semidefinite matrix, where d can be
Electronics 2022, 11, 1836 10 of 16

any complex constant vector. Following from the inequality (38), the inequality (36) can be
approximated by

(m) (m) 0
 1
C13: F̃ ρk0 ,n , ρk0 ,n , ȳk,k0 ,n , ȳk,k0 ,n , Λk,k
n ≥ , ∀k, k0 , n∈K, K, N . (39)
pk0 ,n

Clearly C11, C12 and C13 all have the same form of a convex function minus an affine
function; thus, it is easy to see that these three constraints are all convex as well.
Now, let us focus on the inequality (29). according to the arithmetic inequality, we
have [30]
(m)
ψl,k,n 2 t2k,n
ρl,k,n tk,n ≤ ρl,k,n + (m)
, (40)
2 2ψl,k,n
tk,n (m)
where the equal sign holds if and only if ψl,k,n = . Since the RHS of (40) is convex for
ρl,k,n
the fact that it contains only two convex functions adding up, the inequality (29) can be
approximated by

(m)
ψl,k,n t2k,n
C14: ∑ ∑ 2
ρ2l,k,n + (m)
≤Cmax,l , ∀l ∈L . (41)
n∈N k ∈K 2ψl,k,n

(m)
(m) (m) tk,n
Updating the value of ψl,k,n in the m-th iteration according to ψl,k,n = (m)
can make
ρl,k,n
both sides of (40) finally equal to each other after multiple iterations.
After the transformations above, problem (22) has been completely transformed to a
series of iterative convex optimization problems. The (m + 1)-th iteration can be written as
!
∑ ∑ F ∑ yk,k0 ,n + σ , xk,n , ∑ y 0 + σ , x
(m) 2 (m)
maximize 2
+ fˆ(Ω) k,k ,n k,n
Ω,P,
 k∈K n∈N k0 6=k,k0 ∈K k0 6=k,k0 ∈K
xk,n , yk,k0 ,n , x̄k,n , ȳk,k0 ,n , αl,k,n , (42)
tk,n |l, k, k0 , n∈L, K, K, N
subject to C5,C6,C7,C8,C9,C10,C11,C12,C13,C14 .
By exploiting current algorithms, such as the interior point method, the optimal
solution of problem (42) can be easily obtained. With initial values reasonably set and
parameters updated cyclically according to the solution of each iteration, the result of each
iteration will finally converge to a local optimum. Detailed procedures of our proposed
joint optimization algorithm for user scheduling and resource allocation are summarized
in Algorithm 1:

Alogrithm 1. Joint optimization of user scheduling and resource allocation

n
(m) (m) (m) (m) (m) (m) (m) (m)
1: Set m=0, generate initial values of ρk,n , pk,n , xk,n , yk,k0 ,n , x̄k,n , ȳk,k0 ,n , tk,n , αl,k,n ,
o
(m)
ψl,k,n |l, k, k0 , n∈L, K, K, N with constraints satisfied.
2: Repeat n
3: Solve problem (42) to obtain the optimal solution ρ∗k,n , p∗k,n , xk,n
∗ , y∗ ∗ ∗ ∗
k,k0 ,n , x̄k,n , ȳk,k0 ,n , tk,n ,
o
α∗l,k,n |l, k, k0 , n∈L, K, K, N .
( m +1) ( m +1) ( m +1)
4: Update parameters for the next iteration: ρk,n = ρ∗k,n , pk,n = p∗k,n , xk,n ∗ ,
= xk,n
( m +1) ( m +1) ∗ , ȳ(m+1) ( m +1) ( m +1) ( m +1)
yk,k0 ,n = y∗k,k0 ,n , x̄k,n = x̄k,n k,k0 ,n = ȳ∗k,k0 ,n , tk,n = t∗k,n , αl,k,n = α∗l,k,n , ψl,k,n =
∗
tk,n
0
, l, k, k , n∈L, K, K, N .
ρ∗l,k,n
5: Set m = m + 1.
6: Until Convergence
Electronics 2022, 11, 1836 11 of 16

3.4. Complexity Analysis

It can be observed that the convex optimization problem (42) is actually a second-
order cone program (SOCP), and it contains ñ = 2LKN + 4KN + 2K2 N variables in total.
Referring to [31], if the interior point methodis used to solve this SOCP problem, the
computational complexity is approximately O (ñ)3.5 . Given that Algorithm 1 requires
T iterations
 to converge, the overall computational complexity can then be written as
O (ñ)3.5 T .

4. Numerical Results
In this section, a series of numerical simulation results are evaluated to demonstrate
the effectiveness of the proposed algorithm in this paper. Compared with the general multi-
TRP system without carrier resource allocation, the multi-TRP system with multi-carriers
that performs dynamic resource allocation according to the CSI can achieve higher spectral
efficiency gains. We consider a circular area with radius R = 60 m, where all RAUs are
wrapped-around to avoid boundary effects, and users are randomly distributed over the
coverage area and can associate with any RAU over more than one carriers. The number
of RAUs is set to L = 2, user number K = 4 and available carrier number N = 2. The
following channel coefficient model is considered:
q
hl,k,n = β l,k,n gl,k,n , (43)

where gl,k,n ∼CN (0, 1) is the small scale fading coefficient, and β l,k,n is the large scale fading
coefficient involving path loss and shadowing effects, which would be modeled by

10log10 ( β l,k,n ) = −128.1 − 37.6log10 (dl,k ) + xl,k,n , (44)

where
2

dl,k denotes the distance between RAU l and user k in kilometers, and xl,k,n ∼N 0,
σshad is the shadow fading with σshad = 8 dB. The noise power at receivers is given by
σ2 = 10((nd −30)/10) B, where nd = −174 dBm/Hz is the noise power density, and B is the
single carrier bandwidth. The total frequency bandwith of the system is NB = 10 MHz.
Without loss of generality, the relative stopping criteria of the proposed algorithm is set
to 5×10−5 , and the maximum iteration number is set to 60 to ensure sufficient convergence.
When the maximum power consumption budget at each RAU set to 10 mW, under different
backhaul capacity limits of 50, 100 and 150 Mbps, convergence of the proposed joint
optimization algorithm for user scheduling and resource allocation is demonstrated in
Figure 2.
As can be seen, all gradually increasing curves of the total spectral efficiency of the
multi-carrier multi-TRP system finally converge to a stable value within about seven to
eight iterations. Driven by the penalty function, the matching variables will eventually
converge to 0/1 binary forms.
Figure 3 compares the system spectral efficiency of multi-TRP systems with and with-
out carrier allocation under different settings. Figure 3a depicts carvers of system spectral
efficiency versus transmission power limit Pmax under backhaul capacity constraints of
100 Mbps and 150 Mbps. It can be seen that, with the increase of the power consump-
tion limit at each RAU, the system spectral efficiency of both systems also increases and
gradually flattens out.
Furthermore, with different power consumption constraints, the total spectral effi-
ciency of the multi-TRP system with multi-carriers using the proposed joint optimization
algorithm of user scheduling and resource allocation is always higher than that of the
general multi-TRP system without carrier allocation. This advantage also increases with the
growth of power consumption limits. With the power consumption budget set to 20 mW,
compared with the general multi-TRP system without carrier allocation, our proposed joint
Electronics 2022, 11, 1836 12 of 16

optimization algorithm can help the multi-TRP system with multi-carriers achieve over 6%
performance gain of the total spectral efficiency.
In Figure 3b, how the system spectral efficiency changes with the growth of the back-
haul capacity limit under different power consumption budgets is demonstrated. We
investigate the impact of different backhaul capacity limits from 80 to 200 Mbps on the sys-
tem spectral efficiency under power consumption budgets set to 50 and 500 mW. Consistent
with general inferences, as the backhaul capacity limit increases, the total spectral efficiency
of both multi-TRP systems with and without carrier allocation increases accordingly as
well, and each increasing curve will gradually level off owing to the restriction from other
radio resource settings. As can be seen, the upper limit of system spectral efficiency under
the power consumption budget of 500 mW is apparently higher than that of 50 mW.
In addition, the multi-carrier multi-TRP system with the proposed joint optimization
algorithm can always achieve higher system spectral efficiency than that of the general
multi-TRP system without carrier allocation under different backhaul capacity constraints,
and such gaps also grow as the backhaul capacity constraint increases. Under the backhaul
capacity limit of 200 Mbps, the proposed joint optimization algorithm for user scheduling
and resource allocation can bring about 6% performance gain.
Figure 4 depicts when Pmax is set to 5 W, the system spectral efficiency of the proposed
algorithm versus the number of users under different backhaul capacity limits set to 100,
120 and 150 Mbps. It can be observed from Figure 4 that the system spectral efficiency grow
with the number of users since the proposed algorithm can exploit multiuser diversity
brought by the CSI-based joint user scheduling and carrier allocation. Such result is
consistent with the general conclusion.
We further compare the performance between the proposed optimization algorithm
and another benchmark using the genetic algorithm (GA) for the joint user scheduling
and carrier allocation. As a heuristic algorithm, GA is also one of the popular algorithms
for integer programming. Figure 5 shows the superiority of our proposed algorithm over
GA, with Pmax set to 50 mW. The system spectral efficiency of the optimization algorithm
proposed in this paper is higher than that of GA under different backhaul constraints, and
the advantage is more obvious when the backhaul capacity limit is small.

14 Cmax=50Mbps,Pmax=10mW
Cmax=100Mbps,Pmax=10mW
13 Cmax=150Mbps,Pmax=10mW
System spectral efficiency(bps/Hz)

12

11

10

4
5 10 15 20 25 30
Iteration
Figure 2. Convergence of the joint optimization algorithm for user scheduling and resource allocation.
Electronics 2022, 11, 1836 13 of 16

9.8
Multi-TRP without carrier allocation, Cmax=100Mbps
9.6 Multi-TRP without carrier allocation, Cmax=150Mbps
Multi-TRP with joint resource allocation, Cmax=100Mbps

System spectral efficiency(bps/Hz)

9.4 Multi-TRP with joint resource allocation, Cmax=150Mbps

9.2

8.8

8.6

8.4

8.2

7.8
2 4 6 8 10 12 14 16 18 20
Pmax(mW)
(a)

15
Multi-TRP without carrier allocation, Pmax=50mw
Multi-TRP without carrier allocation, Pmax=500mw
14.5 Multi-TRP with joint resource allocation, Pmax=50mw
Multi-TRP with joint resource allocation, Pmax=500mw
System spectral efficiency(bps/Hz)

14

13.5

13

12.5

12

11.5
80 100 120 140 160 180 200
Cmax(Mbps)
(b)

Figure 3. System SE comparison of the multi-TRP systems with and without carrier allocation.
(a) System SE versus Pmax for different Cmax limitations. (b) System SE versus Cmax for different Pmax
limitations.
Electronics 2022, 11, 1836 14 of 16

21
Cmax=100Mbps
20.5 Cmax=120Mbps
Cmax=150Mbps
20

System spectral efficiency(bps/Hz)

19.5

19

18.5

18

17.5

17

16.5

16

15.5
2 3 4 5 6 7 8
Number of users

Figure 4. System SE versus the number of users for different Cmax limitations.

16
Multi-TRP with joint resource allocation
15 Multi-TRP without carrier allocation
Multi-TRP with GA
System spectral efficiency(bps/Hz)

14

13

12

11

10

6
80 100 120 140 160 180 200
Cmax(Mbps)

Figure 5. Performance comparison between the proposed algorithm and GA.

5. Conclusions
In this paper, we considered a coherent multi-TRP system with multi-carriers, where
RAUs collaboratively serve the users in a coherent way and users are allowed to associate
with any RAU over each carrier freely. In order to maximize the system spectral efficiency,
we established a joint optimization problem of user-RAU-carrier three-sided matching and
downlink transmission power allocation for the coherent multi-carrier multi-TRP system
under the constraints of power consumption and backhaul capacity limits at each RAU, as
well as the QoS requirement at each user.
Due to the existence of integer variables and the products of coupled variables, the
original problem is both non-convex and non-smooth and thus intractable. In this paper, a
joint optimization algorithm of user scheduling and resource allocation was proposed to
tackle this problem.
Electronics 2022, 11, 1836 15 of 16

By employing penalty functions and the SCA method, the original non-convex combi-
natorial optimization problem was transformed into a series of iterative convex optimiza-
tion problems that could be easily solved. Numerical results also confirmed that, with the
same settings, the multi-carrier multi-TRP system with the proposed joint optimization
algorithm of user scheduling and resource allocation was able to achieve higher system
spectral efficiency compared with the multi-TRP system without carrier allocation.

Author Contributions: Conceptualization, Y.B. and X.X.; Formal analysis, Y.B. and X.X.; Funding
acquisition, D.W.; Methodology, Y.B. and X.X.; Project administration, Y.L. and F.Y.; Software, Y.B.;
Supervision, J.Z.; Validation, Y.B.; Writing—original draft, Y.B.; Writing—review and editing, Y.B. and
X.X. All authors have read and agreed to the published version of the manuscript.
Funding: This work was supported in part by the National Key Research and Development Program
under Grant 2020YFB1807200, the National Natural Science Foundation of China (NSFC) under Grant
61871122, the Support Plan for Postdoctoral Talents in Jiangsu Province and the Project funded by the
China Postdoctoral Science Foundation under Grant No. 2021M702500.
Data Availability Statement: Not applicable.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. You, X.; Wang, C.X.; Huang, J.; Gao, X.; Zhang, Z.; Wang, M.; Huang, Y.; Zhang, C.; Jiang, Y.; Wang, J.; et al. Towards 6G wireless
communication networks: Vision, enabling technologies, and new paradigm shifts. Sci. China Inf. Sci. 2021, 64, 1–74. [CrossRef]
2. Zhang, J.; Björnson, E.; Matthaiou, M.; Ng, D.W.K.; Yang, H.; Love, D.J. Prospective multiple antenna technologies for beyond 5G.
IEEE J. Sel. Areas Commun. 2020, 38, 1637–1660. [CrossRef]
3. You, X.; Wang, D.; Wang, J. Distributed MIMO and Cell-Free Mobile Communication; Springer: Berlin/Heidelberg, Germany, 2021.
4. Roh, W.; Paulraj, A. MIMO channel capacity for the distributed antenna. In Proceedings of the IEEE 56th Vehicular Technology
Conference, Vancouver, BC, Canada, 24–28 September 2002; IEEE: Piscataway, NJ, USA, 2002; Volume 2, pp. 706–709.
5. Wang, D.; Wang, J.; You, X.; Wang, Y.; Chen, M.; Hou, X. Spectral efficiency of distributed MIMO systems. IEEE J. Sel. Areas
Commun. 2013, 31, 2112–2127. [CrossRef]
6. Choi, W.; Andrews, J.G. Downlink performance and capacity of distributed antenna systems in a multicell environment. IEEE
Trans. Wirel. Commun. 2007, 6, 69–73. [CrossRef]
7. Zhu, P.; Bao, J.; Zhang, L.; Li, J. A pilot allocation algorithm based on coalitional game theory for distributed MIMO systems.
IEEE Access 2019, 7, 105996–106001. [CrossRef]
8. Li, J.; Wang, D.; Zhu, P.; You, X. Impacts of practical channel impairments on the downlink spectral efficiency of large-scale
distributed antenna systems. Sci. China Inf. Sci. 2019, 62, 1–14. [CrossRef]
9. Li, J.; Lv, Q.; Yang, J.; Zhu, P.; You, X. Spectral and Energy Efficiency of Distributed Massive MIMO with Low-Resolution ADC.
Electronics 2018, 7, 391. [CrossRef]
10. Lv, Q.; Li, J.; Zhu, P.; Wang, D.; You, X. Downlink Spectral Efficiency Analysis in Distributed Massive MIMO with Phase Noise.
Electronics 2018, 7, 317. [CrossRef]
11. Dao, H.T.; Kim, S. Power Allocation and User-AP Connection in Distributed Massive MIMO Systems. IEEE Commun. Lett. 2020,
25, 565–569. [CrossRef]
12. Seki, Y.; Saito, T.; Adachi, F. A study on UE clustering and antenna selection for distributed MIMO cooperative transmission
using multi-user MMSE-SVD. In Proceedings of the 2019 25th Asia-Pacific Conference on Communications (APCC), Ho Chi
Minh City, Vietnam, 6–8 November 2019; IEEE: Piscataway, NJ, USA, 2019; pp. 73–77.
13. Dong, G.; Zhang, H.; Jin, S.; Yuan, D. Energy-efficiency-oriented joint user association and power allocation in distributed
massive MIMO systems. IEEE Trans. Veh. Technol. 2019, 68, 5794–5808. [CrossRef]
14. Bjornson, E.; Jalden, N.; Bengtsson, M.; Ottersten, B. Optimality properties, distributed strategies, and measurement-based
evaluation of coordinated multicell OFDMA transmission. IEEE Trans. Signal Process. 2011, 59, 6086–6101. [CrossRef]
15. Chen, Q.; Yu, G.; Yin, R.; Li, G.Y. Energy-efficient user association and resource allocation for multistream carrier aggregation.
IEEE Trans. Veh. Technol. 2015, 65, 6366–6376. [CrossRef]
16. Chavarria-Reyes, E.; Akyildiz, I.F.; Fadel, E. Energy-efficient multi-stream carrier aggregation for heterogeneous networks in 5G
wireless systems. IEEE Trans. Wirel. Commun. 2016, 15, 7432–7443. [CrossRef]
17. Tsinos, C.; Galanopoulos, A.; Foukalas, F. Low-complexity and low-feedback-rate channel allocation in CA MIMO systems with
heterogeneous channel feedback. IEEE Trans. Veh. Technol. 2016, 66, 4396–4409. [CrossRef]
18. Venturino, L.; Zappone, A.; Risi, C.; Buzzi, S. Energy-efficient scheduling and power allocation in downlink OFDMA networks
with base station coordination. IEEE Trans. Wirel. Commun. 2014, 14, 1–14. [CrossRef]
19. Khan, A.A.; Adve, R.S.; Yu, W. Optimizing downlink resource allocation in multiuser MIMO networks via fractional programming
and the hungarian algorithm. IEEE Trans. Wirel. Commun. 2020, 19, 5162–5175. [CrossRef]
Electronics 2022, 11, 1836 16 of 16

20. Zarandi, S.; Khalili, A.; Rasti, M.; Tabassum, H. Multi-objective energy efficient resource allocation and user association for
in-band full duplex small-cells. IEEE Trans. Green Commun. Netw. 2020, 4, 1048–1060. [CrossRef]
21. Ammar, H.A.; Adve, R.; Shahbazpanahi, S.; Boudreau, G.; Srinivas, K.V. Downlink resource allocation in multiuser cell-free
MIMO networks with user-centric clustering. IEEE Trans. Wirel. Commun. 2021. [CrossRef]
22. Pan, C.; Ren, H.; Elkashlan, M.; Nallanathan, A.; Hanzo, L. The non-coherent ultra-dense C-RAN is capable of outperforming its
coherent counterpart at a limited fronthaul capacity. IEEE J. Sel. Areas Commun. 2018, 36, 2549–2560. [CrossRef]
23. Muruganathan, S.; Faxer, S.; Jarmyr, S.; Gao, S.; Frenne, M. On the system-level performance of coordinated multi-point
transmission schemes in 5G NR deployment scenarios. In Proceedings of the 2019 IEEE 90th Vehicular Technology Conference
(VTC2019-Fall), Honolulu, HI, USA, 22–25 September 2019; IEEE: Piscataway, NJ, USA, 2019; pp. 1–5.
24. Shi, Z.; Chen, W.; Huang, Y.; Tian, J.; Fang, Y.; You, X.; Guo, L. Multiple-input multiple-output enhancement and beam
management. In 5G NR and Enhancements; Elsevier: Amsterdam, The Netherlands, 2022; pp. 413–506.
25. Lin, X. An Overview of 5G Advanced Evolution in 3GPP Release 18. arXiv 2022, arXiv:2201.01358.
26. Nasir, A.A.; Tuan, H.D.; Ngo, H.Q.; Duong, T.Q.; Poor, H.V. Cell-free massive MIMO in the short blocklength regime for URLLC.
IEEE Trans. Wirel. Commun. 2021, 20, 5861–5871. [CrossRef]
27. Lucidi, S.; Rinaldi, F. Exact penalty functions for nonlinear integer programming problems. J. Optim. Theory Appl. 2010,
145, 479–488. [CrossRef]
28. Nguyen, V.D.; Duong, T.Q.; Tuan, H.D.; Shin, O.S.; Poor, H.V. Spectral and energy efficiencies in full-duplex wireless information
and power transfer. IEEE Trans. Commun. 2017, 65, 2220–2233. [CrossRef]
29. Xia, X.; Zhu, P.; Li, J.; Wu, H.; Wang, D.; Xin, Y. Joint optimization of spectral efficiency for cell-free massive MIMO with
network-assisted full duplexing. Sci. China Inf. Sci. 2021, 64, 1–16. [CrossRef]
30. Beck, A.; Ben-Tal, A.; Tetruashvili, L. A sequential parametric convex approximation method with applications to nonconvex
truss topology design problems. J. Glob. Optim. 2010, 47, 29–51. [CrossRef]
31. Ye, Y. Interior Point Algorithms: Theory and Analysis; John Wiley & Sons: Hoboken, NJ, USA, 2011; Volume 44.