Proceedings of the Asia-Pacific Advanced Network 2013 v. 36, p. 74-82.

http://dx.doi.org/10.7125/APAN.36.10
ISSN 2227-3026

CQI-MCS Mapping for Green LTE Downlink Transmission

Mustafa Ismael Salman*, Chee Kyun Ng, Nor Kamariah Noordin, Borhanuddin Mohd Ali
and Aduwati Sali
Department of Computer and Communication Systems Engineering, Faculty of Engineering,
University Putra Malaysia, UPM Serdang, 43400 Selangor, Malaysia.
E-Mails: mofisml3@yahoo.com, {mpnck, nknordin, borhan, aduwati}@eng.upm.edu.my

Abstract: Maximizing the data rate and spectral efficiency are the main targets in
LTE cellular systems. However, any increase in data rate or spectral efficiency will
result in an undesirable increase in energy consumption. Hence, both energy
efficiency (EE) and spectral efficiency (SE) necessitate to be optimized. In this paper,
a multi-objective optimization framework is formulated to prove the tradeoff between
both EE and SE in LTE downlink transmission. A new mapping algorithm between
channel quality indicator (CQI) and modulation and coding scheme (MCS) is
proposed by using a multi-criteria decision making technique. The proposed CQIMCS mapping algorithm will search for the most suitable MCS according to the
reported CQI value by considering the tradeoff between SE and EE while keeping
block error rate (BLER) below its threshold value. Simulation results have shown that
the proposed mapping algorithm can provide an efficient compromise solution
compared to the existing 3GPP-LTE mapping. Therefore, the proposed CQI-MCS
mapping can optimize the EE and SE based on different mapping values depending on
the operator preferences.
Keywords: Energy efficiency; spectral efficiency; CQI; MCS.

1. Introduction
The growing demands for wireless communication services pose new challenges in the design
of the coming generation of cellular networks. In 3GPP-LTE networks, ever-higher data rate and
spectral efficiency are required to meet the increasing demands in cellular traffic [1]. However,
high data rates can be achieved at the cost of energy consumption which needs to be controlled
due to green communication trends [2]. Hence, efficient solutions are necessary to optimize both

74

energy efficiency (EE) and spectral efficiency (SE) to meet the green LTE requirements. There
are many technologies and strategies used in wireless communications that reflect the
competition between EE and SE [3]. Since EE and SE are conflicting between each other, a
tradeoff between them would be attained. A general EE-SE tradeoff framework has been
formalized by the authors in [4], and it is shown that EE is quasi-concave in SE.
In 3GPP-LTE networks, Adaptive Modulation and Coding (AMC) is one of the most
important state-of-the-art techniques that has been used to improve system capacity [5]. AMC is
a channel-aware technique that adapts the modulation and coding scheme (MCS) according the
channel condition. The channel condition can be reported back by the user by using channel
quality indicator (CQI). Nevertheless, the modulation order can also be used to control the
energy consumption [6]. In this paper, a new algorithm of mapping is proposed by using a multicriteria decision making technique. The proposed CQI-MCS mapping algorithm considers the
tradeoff between SE and EE with keeping the block error rate (BLER) below its threshold value.
It can optimize any parameter over the other by using a weighting parameter. Simulation results
have shown that the proposed mapping algorithm can provide an efficient compromise solution
compared to the existing 3GPP-LTE mapping. Therefore, the proposed mapping algorithm can
result in different mapping values based on operator priorities.
2. System Model and Problem Formulation
Considering the downlink transmission in a multi-user LTE network where the conventional
OFDM is used to transmit data over several narrowband (15kHz) subcarriers to J number of
users. In frequency domain, each 12 subcarriers are grouped to form the basic scheduling
element of 3GPP LTE which is named as resource block (RB) [7]. For energy-efficient
transmission, the maximum number of transmitted bits per unit energy is needed to be achieved.
According to [8], the EE is monotonically decreasing with the increase of SE in case that the
circuit power consumption is not considered in the calculation of total power consumption, P(R).
However, the relation between EE and SE will be more complicated when the circuit power
consumption is considered. In particular, EE-SE relation will turn to bell shape according to the
following expression
B SE g n
(1)
EE =
Pcn + BN 0 2 SE 1

where gn is the average channel gain on RBn, N0 is the noise power spectral density, and
B is the bandwidth per RB. To show the effect of circuit power on the trade-off between EE and
SE, a LTE downlink transmission with transmit power of 43 dBm and power amplifier efficiency
of 14 % will be considered. The EE-SE tradeoff is shown in Figure 1 with different channel gainto-noise-ratio (CNR) per RB. It is clear that the EE is more sensitive to the change in SE when
the circuit power consumption is minimized.

75

Figure 1. The relation between EE and SE with different channel condition.

Therefore, both EE and SE should be optimized to meet the green LTE requirements.
Accordingly, the optimization function for green LTE downlink transmission can be formulated
as

Pj ,n [k ]g j ,n [k ]

B j ,n log 2 1 +

J
N
J N
B j ,n N 0
Pj ,n [k ]g j .n
(2)

+ j ,n [k ]
max j ,n [k ]log 2 1 +

P , B[ k ]
(
)
B j ,n N 0 j =1 n =1
Pcn , j [k ] + Pn , j [k ]

j =1 n =1

subject to
J

[k ] 1,
j ,n

j =1

j ,n

n =1
N

Pj ,n [k ]g j.n
R req
log 2 1 +
j , j

B j ,n N 0

p [k ] p
n

max

n =1

where Pcn , j [k ] is the dynamic circuit power consumption, which is related to transmission
bandwidth, as well as the static circuit power consumption.

j ,n

in the first constraint is either 0

or 1 to assure that the nth RB would be allocated once at each scheduling. The second constraint
guarantees the quality of service (QoS) requirements for all users, while the third constraint
shows that the allocated power should be less than the power available at the base station.
According to (2), it is shown that this optimization problem is Non-deterministic Polynomialtime hard (NP-hard), and thus, sub-optimal solution is required.

76

3. CQI-MCS Multi-Criteria Decision Making Algorithm

In LTE downlink transmission, the AMC process is carried out by the eNodeB and depends
on CQI feedback reported by the user equipment (UE). Like HSDPA, the CQI in LTE systems
does not reflect the SINR only. Instead, it is a measurement that reflects the most spectralefficient MCS that can be decoded by the relevant UE with BLER less than 10%. In the proposed
CQI-MCS algorithm the upper-bounded SE per RB can be achieved by
(1 BLER ) log 2 M L
(3)
SE =
0.5 10 3
where 1 is the coding rate, M is the number of constellation points, and L is the number of
subcarriers per RB. According to (3), the relation between the modulation order and transmit
power can be formulated as
Pt g n
0.5 10 3
1 +

(4)
M=
(1 BLER) L BN 0
It is obvious that the lower modulation order can be chosen to transmit log2M bits per symbol
with reduced amount of transmit power but more RBs are required to be allocated. According to
green LTE requirements, there are some competing parameters must be optimized, e.g. EE, SE
and coding rate (CR). Therefore, a multi-criteria decision making (MCDM) approach can be
used to formulate the tradeoff among these conflicting parameters. Prior to dealing the green
LTE multi-criteria problem, some issues must be resolved.
In 3GPP-LTE standard, 15 CQI values are considered as shown in Table 1. Each CQI value
will be mapped into one MCS that can achieve maximum SE. In the proposed CQI-MCS
mapping, each CQI can be mapped into a set of MCSs such that the BLER is kept below 10%.
The resulted set of MCSs represents the set of alternatives in the green LTE decision matrix.
Consequently, the proposed green LTE decision matrix is an (m x n) matrix where m represents
the number of criteria, i.e. in this paper m = 3 to reflect EE, SE and CR, and n represents the
number of alternatives. The relative importance of each criterion is not equal for all cellular
operators and base stations. Typically, each operator can decide its priorities according to many
parameters, such as traffic load and provided services. Therefore, a weighting factor will be
assigned to each criterion according to the decision maker priority. For instance, EE and SE
contribute equally to the green LTE objective but they have an absolute importance over the CR.
This is because the necessary condition for BLER can be achieved by any of the feasible MCSs.
Because of the conflicting property associated to the MCDM problems, there is no single
global optimal solution that can maximize (or minimize) all the objectives simultaneously. In
contrast, there will be a set of Pareto-Optimal solutions on which none of the objective functions
can be improved without worsening the other objective value. In green LTE optimization
problem, however, a compromise MCS solution that achieves the trade-off between the SE and
EE must be acquired. The compromise solution can be defined as the closest solution to the

77

ideal/aspired point that is not belonging to the feasible area, e.g. the ideal point in green LTE
optimization problem represents the maximum value of both EE and SE, and the minimum value
of CR which is not feasible.
The difficulty in finding the compromise solution is how to define the closeness between
feasible Pareto-optimal solutions and the infeasible ideal solution. In this paper, the multicriteria
optimization and compromise solution ranking method of VIKOR is used [9]. VIKOR aims to
formulate the multi-criteria measure to compromise ranking by using Lp-metric which is used in
compromise programming method [10]. Hence, the Lp-metric is formulated as
1/ p

p
n


(5)
= w j f j* f ij f j* f j

j =1

*
where 1 p , i = 1, 2,..., m. , and fj and fj- represent the best and worst values of all criterion

Lip

functions respectively. Lip=1 and Lip= (as Si and Ri in (6)) are used to formulate the ranking
measure. The minimum value of Si represents the maximum group utility, while the minimum
value of Ri represents the minimum individual regret of the opponent. Consequently, the
following steps should be carried out.
- Determine the best fj* and the worst fj- for all green LTE criteria; EE, SE and CR
f j* = max f ij , f j = min f ij , if j-th criterion represents a benefit, i.e. EE and SE.
i

f j* = min f ij , f j = max f ij , if j-th criteria represents a cost, i.e. CR.

i

- Compute the vlues of Si and Ri for all of green LTE competing criteria
n

S i = wi f i* f ij / f i* f i and Ri = max[wi ( f i* f ij )/ ( f i* f i )]
j =1

)(

where wi is the weighting factor to indicate the decision maker preferences.

- Compute the values of Qi by using the following relation
Q = v(S i S * ) / (S S * ) + (1 v )(Ri R * )(R R * )
where
S * = min S i = min[(S i )i = 1,2,.........n],

S = max S i = max[(S i ) i = 1,2,........n]

R * = min Ri = min[(Ri ) i = 1,2,.........n],

R = max Ri = max[(Ri )i = 1,2,........n]

(6)

(7)

v and (1-v) are the weight of the maximum group utility and individual regret respectively.
- Rank the alternatives (MCSs)
Ranking the alternatives is carried out by sorting the Q values in decreasing order. The best
MCS which meets the cellular operator requirement would be the minimum Q value obtained in
the previous step.
Table.1 shows new CQI-MCS mapping compared with the 3GPP LTE standard by
considering the circuit power consumption. As stated previously, the new CQI-MCS mapping
takes into account the tradeoff between EE and SE by using fixed weighting assignment. The

78

weight factor can be chosen according to operator priority and/or the traffic load. In other words,
higher weight can be assigned to the SE when the network is fully loaded. Therefore, different
CQI-MCS mappings can be adapted according to the network status by simply changing the
weighting parameter.
Table 1. The proposed CQI-MCS mapping compared to mapping in 3GPP-LTE.
CQI

MCS-3GPP

SE/RB 3GPP

EEx103 3GPP

MCS-VIKOR

SE/RB VIKOR

EEx103 VIKOR

QPSK-0.078

0.14

1.4

QPSK-0.096

0.18

1.77

2
3
4
5
6
7
8
9
10
11
12
13
14
15

QPSK-0.12
QPSK-0.19
QPSK-0.30
QPSK-0.44
QPSK-0.6
16QAM-0.37
16QAM-0.49
16QAM-0.61
64QAM-0.46
64QAM-0.56
64QAM-0.66
64QAM-0.77
64QAM-0.87
64QAM-0.94

0.224
0.35
0.56
0.82
1.10
1.38
1.79
2.24
2.52
3.08
3.64
4.2
4.76
5.20

2.24
3.54
5.6
8.2
11.01
13.81
17.92
22.4
25.2
30.8
36.4
42
47.6
52.08

QPSK-0.096
QPSK-0.24
QPSK-0.24
QPSK-0.37
QPSK-0.438
QPSK-0.438
QPSK-0.438
16QAM-0.369
16QAM-0.42
16QAM-0.47
16QAM-0.54
16QAM-0.54
64QAM-0.45
64QAM-0.45

0.18
0.448
0.448
0.69
0.82
0.82
0.82
2.02
2.3
3.34
2.95
2.95
2.46
2.46

3.11
4.46
6.9
10.69
16.45
24.14
34.84
56.18
79.83
143.63
152.96
208.8
148
148

4. Simulation Setup and Result Discussion

In this section, simulation results are presented and discussed to show the effectiveness of the
proposed approach. A single-cell downlink transmission is considered in the proposed scenario.
Macro-cell for urban areas is used to represent the LTE channel realization, and the related
parameters are shown in Table 2.
Table 2. Simulation parameters.
Carrier Frequency
Bandwidth
Subcarrier/RB, Spacing
Transmit Power
Radius
Scheduling TTI
Simulation Time
Traffic Flows
Circuit Power
PA eff.

2GHZ
5MHz
12, 15KHz
47dBm
1km
1ms
120ms
Video H264,
10W
14%

79

The number of fixed-position users is chosen to be in the range of [10, 50]. The simulation for
the system level of proposed algorithm is carried out by using the open-source LTE-Sim
framework [11]. In this scenario, equal priority weight is assigned to each of EE and SE. Figure
2 shows the performance of the simulated scenario in terms of EE and SE. The EE can be
enhanced by using the proposed CQI-MCS mapping. The resulted EE performance is compared
with the conventional scenario that considers the standard mapping proposed by 3GPP. It is
shown that the proposed CQI-MCS mapping produces a noticeable enhancement in EE
especially when the number of users is increasing due to multiuser diversity. In other words, the
EE is enhanced by 21% when the number of users reaches to 50. It is also shown that EE can be
further improved to its maximum value by simply adjusting preference weight to maximum
priority as shown in Figure 2a. This enhancement, however, will result in a serious detriment to
SE compared to other schemes as shown in Figure 2b. Therefore, the compromise result can be
obtained by assigning equal priority to EE and SE. Moreover, the preference weighting
parameter can be adjusted according to the operator priority and/or traffic load, and thus, the
MCS can be adapted according to channel traffic load conditions.

(a)

(b)

Figure 2. The relationship between the number of users and (a) EE and (b) SE.
5. Conclusions
In this paper, the tradeoff between EE and SE in LTE downlink transmission is presented.
Also, the influence of AMC on system performances is highlighted. A new CQI-MCS mapping
has been proposed based on multi-criteria decision making algorithm VIKOR. It is shown that
the proposed mapping can optimize the EE-SE tradeoff. Furthermore, this mapping can be
adapted according to network operator priority by simply changing the weighting factor which

80

associated to the competed criteria. Furthermore, the proposed mapping can be self-optimized
according to traffic load condition, and accordingly, the performance of the whole system will be
enhanced. Unlike the conventional AMC proposed by 3GPP, the proposed CQI-MCS mapping
will let the MCSs to be adapted according to channel and traffic load conditions.
References
1.
2.

LTE Mobile Broadband Standard Available: http://www.3gpp.org/LTE

Feng, D.; Jiang, C.; Lim, G.; Cimini Jr, L.J.; Feng, G.; Li, G.Y. A Survey of EnergyEfficient Wireless Communications. IEEE Communications Surveys & Tutorials, 2013, vol.
15, no. 1, pp. 167-178.
3. Salman, M.; Ng, C.K.; Noordin, N.K. Energy- and Spectral-Efficient Wireless Cellular
Networks. Joel J.P.C. Rodrigues et al. (Eds.), First International Conference on Green
Communications and Networking, (GreeNets 2011), Colmar, France, October 5-7, 2011,
Lecture Notes of the Institute for Computer Sciences, Social Informatics and
Telecommunications Engineering (LNICST), 2012, vol. 51, pp. 171-185, Springer Dordrecht
Heidelberg.
4. Xiong, C.; Li, G.Y.; Zhang, S.; Chen, Y.; Xu, S. Energy-and Spectral-Efficiency Tradeoff in
Downlink OFDMA Networks. IEEE Transactions on Wireless Communications, 2011, vol.
10, pp. 3874-3886.
5. Yu J.; Yu, F. Link Adaptive Technology in Wireless Channel. The 8th International
Conference on Electronic Measurement and Instruments (ICEMI '07), 2007, pp. 2-154-2158.
6. Cui, S.; Goldsmith, A.J.; Bahai, A. Energy-Constrained Modulation Optimization. IEEE
Transactions on Wireless Communications, 2005, vol. 4, no. 5, pp. 2349-2360.
7. Zyren J.; McCoy, W. Overview of the 3GPP Long Term Evolution Physical Layer. White
Paper, 2007. Freescale Semiconductor, Inc.
8. Miao, G.; Himayat, N.; Li, Y.; Bormann, D. Energy Efficient Design in Wireless OFDMA.
IEEE International Conference on Communications (ICC '08), 2008, pp. 3307-3312.
9. Opricovic, S.; Tzeng, G.H. Compromise Solution by MCDM Methods: A Comparative
Analysis of VIKOR and TOPSIS. European Journal of Operational Research, 2004, vol.
156, pp. 445-455.
10. Yu, P.L. A Class of Solutions for Group Decision Problems. Management Science, 1973,
vol. 19, pp. 936-946.
11. Piro, G.; Grieco, L.A.; Boggia, G.; Capozzi, F.; Camarda, P. Simulating LTE Cellular
Systems: An Open-Source Framework. IEEE Transactions on Vehicular Technology, 2011,
vol. 60, no. 2, pp. 498-513.

81

 2013 by the authors; licensee Asia Pacific Advanced Network. This article is an open-access
article distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).

82