Slide 1: Good morning everyone.

Im… and Today, we are presenting our thesis project on 'Power

Allocation Using Network Slicing Enabled NOMA in 5G Networks.

Slide 2: We will begin with an introduction to telecommunications and wireless communications.

Then, we will move on to the background, problem statement, related papers, our proposed
system model, simulation results, and conclude with our findings and future work.

Slide OMA:

Orthogonal Multiple Access, or OMA is a key concept in wireless networks that ensures efficient
and interference-free communication.

As we see a base station is communicating with two users, User 1 and User 2.

This setup illustrates a typical OMA scenario where the base station allocates resources such as
time or frequency orthogonally among users, it ensures that there is no interference between
users, simplifying the design and improving the reliability of communication.

So the total available bandwidth (w) is equally divided with each user. This division ensures that
both users have a fair share of the available resources.

However, a notable drawback is that OMA can lead to inefficient use of resources, especially
when user demand varies dynamically. Fixed allocation might result in underutilization of the
allocated bandwidth.

Moreover, as the number of users increases, the available resources need to be divided into
smaller portions, potentially limiting the data rates for each user.

In summary, OMA architecture provides a clear framework for resource allocation in wireless
networks, ensuring interference-free communication and simplicity, but also presenting
challenges in resource utilization and scalability.
Slide Stackelberg:

Today, we're exploring Dynamic Games, also known as Sequential Move Games. In this
strategic model, players, called followers, make decisions one after another in response to
a leader.

The game unfolds in two main steps:

1. The leader acts first, making decisions without knowing the follower's subsequent
moves. The leader's choices significantly influence how the follower will respond,
aiming to optimize their own goals.
2. Next, the follower observes the leader's actions and then makes their own
decisions. The follower's choices are reactive, shaped by the leader's earlier
moves.

Slide Problem Statement:

Next, I'd like to present the problem statement

The main aim is to minimize latency and maximize throughput by optimizing power allocation in
5G networks.

To achieve this, we propose employing network slicing enabled NOMA.

On one hand, we have network slicing techniques like eMBB, URLLC that enhance the flexibility
and efficiency of network resource utilization.

At the same time, NOMA or Non-Orthogonal Multiple Access offers promising potential benefits
in terms of enhanced spectral efficiency and improved connectivity within the 5G context.

To address this problem of joint power allocation and resource management, we model it as a
Stackelberg game, where the base stations act as leaders optimizing key decision variables
related to resource allocation and power control. The user devices then act as followers,
adjusting their values accordingly based on the leader's strategies.

Slide NOMA Clustering:

Now, let's move on to the next step, which involves grouping users into clusters based on their
proximity and channel quality. In this step, users within each slice are grouped into clusters to
optimize resource utilization. This is achieved using the NOMA protocol and the Nearest
Neighbor Nonlinear Fusion (NNNF) algorithm. The NNNF algorithm is a method that clusters
users by calculating distances using metrics like Signal-to-Noise Ratio (SNR) or channel gain, then
refines these clusters using nonlinear fusion techniques to optimize their structure. NOMA
allows multiple users to share the same frequency resource block (RB) by employing different
power levels, which requires Successive Interference Cancellation (SIC) for message decoding.
The key points of this step include user pairing, typically set to pairs (m=2) to reduce complexity
and avoid error propagation

Iterations:

Now were are going to see how the BS iterates to optimize power,

Here is the initial state where there is one strong user and one weak user connected to the
base station.

And then after multiple iterations, likely in the first phase, the power allocation between
the strong and weak users converges to an equal distribution.

In the second phase of iterations, the power allocation is further optimized. Both users
receives a higher power allocation, potentially to maximize throughput for eMBB
services, and to minimize latency for URLLC services.

Implementation: We used MATLAB to implement the code, which can also be run on Octave. In
the initial setup phase. In the initial setup the base station sets the coordinates and generates
random user positions with their respective channel coefficients.

Moving to the next step, slice management plays a crucial role. It facilitates the creation and
management of network slices within the communication system, supporting different quality-
of-service criteria for various services.

The third step involves user clustering based on the NOMA principle. Users are grouped by
clustering them based on their channel conditions using the Nearest Neighbor Nonlinear Fusion
(NNNF) algorithm.

Finally, in the optimization phase, a game algorithm is implemented between the base station
and the user clusters. This algorithm optimizes the power allocation levels, with the objectives of
minimizing latency for delay-sensitive services like URLLC and maximizing throughput for high
data rate services like eMBB. By following these systematic implementation steps, from initial
setup to slice management, user clustering, and optimization, the MATLAB code effectively
models and solves the power allocation problem, leveraging the benefits of network slicing and
NOMA in a 5G network context.

Results:

● These figures illustrate the iterative process of optimizing power allocation for strong
and weak users in the URLLC and eMBB slices over 10 iterations. In the URLLC slice, the
maximum power ranges between 0 and 35 dBm, while in the eMBB slice, the range of
maximum power extends from 0 to 40 dBm.

● In the initial iterations, the base station (BS) iterates to distribute power equally
between the weak and strong users. Consequently, the power allocated to strong users
decreases initially, while that allocated to weak users slightly increases. In the
subsequent iterations, both the power allocated to strong users and weak users
increases progressively.

● This iterative optimization process aims to balance the power allocation dynamically to
meet the quality of service (QoS) requirements of both URLLC and eMBB services
effectively

This figure illustrates how power allocation affects latency in the URLLC slice and
throughput in the EMBB slice.

In the first plot, as power increases in the URLLC slice, latency decreases for both strong
and weak users. Higher power leads to stronger signals, crucial for applications needing
ultra-reliable low latency.

The second plot shows that increasing power in the EMBB slice enhances throughput for
both user groups. More power allows for higher data rates and better performance in
mobile broadband.

These graphs highlight the balance between power usage and important performance
measures like latency and throughput, offering insights for optimizing power in different
5G scenarios

This figure compares how efficiently Orthogonal Multiple Access (OMA) and Non-
Orthogonal Multiple Access (NOMA) use spectrum as more users join the system.
NOMA consistently outperforms OMA in spectral efficiency, especially with more users.
NOMA achieves this by handling multiple users on the same frequencies using power-
domain multiplexing.

The figure also shows how network slicing enhances NOMA's performance. Efficient
resource allocation across slices boosts NOMA's spectral efficiency even further
compared to non-sliced NOMA.

First, we focused on optimizing power allocation in 5G networks using Non-Orthogonal

Multiple Access (NOMA) and network slicing techniques. These advanced methods
enhance our network's efficiency and performance.

Secondly, we enhanced Ultra-Reliable Low-Latency Communications (URLLC) and

Enhanced Mobile Broadband (eMBB) to improve latency and throughput. This is crucial
for meeting the diverse needs of modern applications, from critical communications to
high-speed data services.

Lastly, we explored Stackelberg Game Theory to model hierarchical interactions and

improve resource allocation. This theoretical approach helps us manage network
resources more effectively, optimizing overall network performance.