caidespries.github.io/project1.html at main · CaideSpries/caidespries.github.io · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
<!DOCTYPE html>
<html lang="en">

<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>Mobile and Wireless Networks</title>
    <link rel="stylesheet" href="assets/css/projects.css">
</head>

<body>
<div class="container">
    <header>
        <h1>Mobile and Wireless Networks Project</h1>
    </header>
    <main>
        <section id="project-overview">
            <h2>Project Overview</h2>
            <div style="text-align: center; max-width: 90%; margin: 1rem auto;">
                <p>In today's connected world, mobile devices often have access to multiple types of wireless networks
                    simultaneously, such as Wi-Fi, 4G, and 2G. These are known as Heterogeneous Networks (HetNets). The
                    challenge is to enable a device to seamlessly switch between these networks to maintain the best
                    possible connection quality. This project investigates how to optimise these switching decisions,
                    known
                    as 'vertical handoffs.' The core of this project was to develop and simulate a network selection
                    algorithm using the TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method
                    to
                    intelligently choose the best network based on multiple criteria.</p>
            </div>
            <figure style="text-align: center;">
                <img src="assets/img/Projects Docs/project1/heterogenous-networks-diagram.webp"
                     alt="Diagram showing heterogeneous networks with different wireless technologies including Wi-Fi, 2G, and 4G"
                     style="max-width:100%; height:auto;">
                <figcaption>Figure 1: Heterogeneous Networks (HetNets) Architecture showing multiple wireless access
                    technologies
                </figcaption>
            </figure>
        </section>
        <section id="technologies-used">
            <h2>Technologies Used</h2>
            <ul>
                <li>Wi-Fi (WLAN)</li>
                <li>2G (GSM/EDGE)</li>
                <li>4G (LTE)</li>
                <li>MADM Techniques (TOPSIS)</li>
                <li>Python for simulation and code setup</li>
            </ul>
        </section>
        <section id="methodology-process">
            <h2>Methodology and Process</h2>
            <div style="text-align: center; max-width: 90%; margin: 1rem auto;">
                <p>The project followed a structured approach to solve the network selection problem. The core of the
                    methodology was the TOPSIS algorithm, which provides a logical framework for ranking the available
                    networks. The process, illustrated in the flowchart below, involves several key steps:</p>
            </div>
            <ul>
                <li>Defining the Criteria: Establishing the metrics for comparison (Allowed Bandwidth, Delay, Power
                    Consumption).
                </li>
                <li>Building the Decision Matrix: Assigning performance values to each network for every criterion.</li>
                <li>Normalizing and Weighting: Processing the data to allow for fair comparison and applying
                    user-defined preferences (weights).
                </li>
                <li>Calculating Closeness: Determining which network is closest to the 'ideal' solution.</li>
                <li>Selecting the Best Network: Choosing the RAT with the highest closeness coefficient.</li>
            </ul>

            <figure style="text-align: center;">
                <img src="assets/img/Projects Docs/project1/flowchart-RAT-selection-decisions.webp"
                     alt="Flowchart showing the RAT selection decision process using TOPSIS methodology, from criteria definition to final network selection"
                     style="max-width:100%; height:auto;">
                <figcaption>Figure 2: Flowchart depicting the TOPSIS-based network selection process</figcaption>
            </figure>
        </section>
        <section id="project-results">
            <h2>Simulation and Results</h2>

            <h3 style="text-align: center;">Part A: The Initial Test with a Realistic Bias</h3>
            <div style="text-align: center; max-width: 90%; margin: 1rem auto;">
                <p>The initial simulation was conducted using a decision matrix based on real-world performance data,
                    where
                    Wi-Fi has superior bandwidth, delay, and power consumption. The following matrix shows the actual
                    values used for each network type across the three key criteria:</p>
            </div>

            <figure style="text-align: center;">
                <img src="assets/img/Projects Docs/project1/original-decision-matrix-real-values.webp"
                     alt="Decision matrix showing realistic values where Wi-Fi outperforms other networks across bandwidth (11 Mbps vs 2 Mbps for 4G and 0.384 Mbps for 2G), delay (10ms vs 40ms for 4G and 100ms for 2G), and power consumption (1W vs 3W for 4G and 1.5W for 2G)"
                     style="max-width:100%; height:auto;">
                <figcaption>Figure 3: Initial Decision Matrix with Realistic Performance Values
                </figcaption>
            </figure>

            <div style="text-align: center; max-width: 90%; margin: 1rem auto;">
                <p>The results confirmed the algorithm's logic, as it consistently selected Wi-Fi, even when user
                    priorities (weights) were changed. This is illustrated in the chart below, which shows that Wi-Fi
                    (RAT_W) was selected by almost all users, regardless of the weight placed on bandwidth. The same
                    behaviour was observed when varying weights were placed on power consumption and delay where Wi-Fi
                    reigned supreme.</p>
            </div>

            <figure style="text-align: center;">
                <img src="assets/img/Projects Docs/project1/weight-selection-criterion-1.webp"
                     alt="Bar chart showing that Wi-Fi (RAT_W) was selected by almost all users regardless of the weight assigned to bandwidth"
                     style="max-width:100%; height:auto;">
                <figcaption>Figure 4: User selection distribution with varying weights applied to bandwidth criterion
                </figcaption>
            </figure>

            <h3 style="text-align: center;">Part B: The "Bonus Task" with a Balanced Matrix</h3>
            <div style="text-align: center; max-width: 90%; margin: 1rem auto;">
                <p>To explore the true impact of user preferences, a second simulation was run with a modified decision
                    matrix where the strengths were distributed more evenly across Wi-Fi (RAT_W), 4G (RAT_L), and 2G
                    (RAT_G). This balanced matrix was designed to better test the algorithm's decision-making process by
                    giving each network type distinct advantages:</p>
            </div>

            <figure style="text-align: center;">
                <img src="assets/img/Projects Docs/project1/modified-decision-matrix-test-logic.webp"
                     alt="Modified decision matrix showing balanced values where each network has different strengths: Wi-Fi leads in bandwidth (5 units), 4G in power consumption (1 unit), and both Wi-Fi and 4G have equal delay advantage (10 units)"
                     style="max-width:100%; height:auto;">
                <figcaption>Figure 5: Modified Decision Matrix with Balanced Performance Values
                </figcaption>
            </figure>

            <div style="text-align: center; max-width: 90%; margin: 1rem auto;">
                <p>This balanced approach revealed a much more dynamic selection process. The charts below show that the
                    best network now changes depending on what the user prioritises—high bandwidth, low delay, or low
                    power consumption.</p>
            </div>

            <figure style="text-align: center;">
                <img src="assets/img/Projects Docs/project1/result-graph-varied-weight-constant-abw.webp"
                     alt="Stacked bar graph showing how network selection changes when bandwidth is prioritised, with a combination of varying weights becoming the preferred choice as the weight increases"
                     style="max-width:100%; height:auto;">
                <figcaption>Figure 6: Network selection results when varying the weight of available bandwidth
                </figcaption>
            </figure>

            <div style="text-align: center; max-width: 90%; margin: 1rem auto;">
                <p><strong>Key insight:</strong> As bandwidth priority increases, network preference shifts from
                    power-efficient 4G to high-bandwidth Wi-Fi, demonstrating how the algorithm adapts to changing user
                    priorities while maintaining minimal 2G selection throughout.</p>
            </div>

            <figure style="text-align: center;">
                <img src="assets/img/Projects Docs/project1/result-graph-varied-weight-constant-delay.webp"
                     alt="Stacked bar graph showing how network selection changes when delay is prioritised, with Wi-Fi and 4G becoming the preferred choice as the weight increases"
                     style="max-width:100%; height:auto;">
                <figcaption>Figure 7: Network selection results when varying the weight of network delay</figcaption>
            </figure>

            <div style="text-align: center; max-width: 90%; margin: 1rem auto;">
                <p><strong>Key insight:</strong> With minimal delay weighting, 4G dominates due to its balanced
                    attributes. As delay importance increases, Wi-Fi becomes the preferred network while 2G selection
                    declines sharply, reflecting the underlying delay characteristics of each network type.</p>
            </div>

            <figure style="text-align: center;">
                <img src="assets/img/Projects Docs/project1/result-graph-varied-weight-constant-power.webp"
                     alt="Stacked bar graph showing how network selection changes when power consumption is prioritised, with 4G becoming the preferred choice as the weight increases"
                     style="max-width:100%; height:auto;">
                <figcaption>Figure 8: Network selection results when varying the weight of power consumption
                </figcaption>
            </figure>

            <div style="text-align: center; max-width: 90%; margin: 1rem auto;">
                <p><strong>Key insight:</strong> Wi-Fi dominates when power efficiency is less important due to its
                    balanced performance. As power consumption priority increases, 4G becomes the preferred choice due
                    to its superior power efficiency, while 2G maintains minimal presence throughout.</p>
            </div>
        </section>
        <section id="conclusion">
            <h2>Conclusion</h2>
            <div style="text-align: center; max-width: 90%; margin: 1rem auto;">
                <p>This project successfully demonstrated the effectiveness and complexities of using the TOPSIS
                    algorithm for vertical handoff decisions. The initial simulations, based on realistic data,
                    confirmed that the algorithm correctly identifies an objectively superior network.
                    However, the key insight came from the analysis with a balanced decision matrix. It revealed that
                    the user's specific priorities—whether for speed, responsiveness, or battery life—are a critical
                    factor that can dynamically change the optimal network choice. The results show how changing the
                    decision matrix significantly impacts network selection, and when combined with user preferences,
                    highlights the complexities that warrant further research. Ultimately, the project proves that
                    effective network selection is not just about the raw performance of technologies, but about the
                    intricate interplay between network capabilities and user needs.</p>
            </div>
        </section>
        <section id="project-document">
            <h2>Project Document</h2>
            <div style="text-align: center; max-width: 90%; margin: 1rem auto;">
                <p>For more detailed information, you can access the following resource:</p>
                <div style="display: flex; justify-content: center; gap: 20px; flex-wrap: wrap; margin-top: 1rem;">
                    <a href="assets/docs/Projects Docs/project1/EEE4121F_Project_SPRCAI002.pdf"
                       target="_blank"
                       class="button-link"
                       style="display: inline-block; padding: 10px 15px; background-color: #3498db; color: white; text-decoration: none; border-radius: 4px;">
                        View Project Document (PDF)
                    </a>
                </div>
            </div>
        </section>
    </main>
    <footer>
        <a href="index.html">Back to Homepage</a>
    </footer>
</div>
</body>

</html>