PEMODELAN KOMPARATIF QUALITY OF SERVICE PADA LINGKUNGAN SOFTWARE DEFINED NETWORK MENGGUNAKAN ALGORITMA ADAPTIF DDPG DAN DQN

Muhammad Rizky Aulia, . (2026) PEMODELAN KOMPARATIF QUALITY OF SERVICE PADA LINGKUNGAN SOFTWARE DEFINED NETWORK MENGGUNAKAN ALGORITMA ADAPTIF DDPG DAN DQN. Skripsi thesis, Universitas Pembangunan Nasional Veteran Jakarta.

[img] Text
ABSTRAK.pdf
Download (169kB)
[img] Text
AWAL.pdf
Download (2MB)
[img] Text
BAB 1.pdf
Restricted to Repository UPNVJ Only
Download (282kB)
[img] Text
BAB 2.pdf
Restricted to Repository UPNVJ Only
Download (478kB)
[img] Text
BAB 3.pdf
Restricted to Repository UPNVJ Only
Download (627kB)
[img] Text
BAB 4.pdf
Restricted to Repository UPNVJ Only
Download (1MB)
[img] Text
BAB 5.pdf
Download (256kB)
[img] Text
DAFTAR PUSTAKA.pdf
Download (218kB)
[img] Text
RIWAYAT HIDUP.pdf
Restricted to Repository staff only
Download (114kB)
[img] Text
LAMPIRAN.pdf
Restricted to Repository UPNVJ Only
Download (592kB)
[img] Text
ARTIKEL KI.pdf
Restricted to Repository staff only
Download (314kB)
[img] Text
HASIL PLAGIARISME.pdf
Restricted to Repository staff only
Download (19MB)

Abstract

The growing complexity of data traffic in campus networks demands a more adaptive bandwidth management mechanism than conventional or static routing approaches. This research aims to identify the most suitable routing strategy for campus networks with dual gateway paths, considering that an inappropriate choice of adaptive algorithm may add system complexity without delivering commensurate performance gains. This research proposes a comparative modeling of Quality of Service (QoS) performance in a Software Defined Network (SDN) environment using two Deep Reinforcement Learning (DRL) algorithms with distinct action-space characteristics: Deep Q-Network (DQN) with a discrete action space and Deep Deterministic Policy Gradient (DDPG) with a continuous action space. The FIK UPNVJ campus network topology was simulated using the Mininet emulator with Ryu as the SDN controller, and compared across four routing scenarios: conventional network, static Dijkstra, DQN, and DDPG. The performance of each scenario was measured using four QoS metrics throughput, latency, packet loss, and jitter referring to the TIPHON and ITU-T Y.1541 standards, and tested for statistical significance. The test results show that under peak-load (burst) conditions, DDPG achieves significantly lower packet loss compared to the other scenarios, as its continuous action space enables simultaneous utilization of both gateway paths a capability unattainable by the discrete action space of DQN or the static routing of Dijkstra. These findings indicate that algorithm suitability is context-dependent: DDPG is best suited for dual-path topologies with fluctuating loads, while Dijkstra and DQN remain relevant under stable load conditions.

Item Type: Thesis (Skripsi)
Additional Information: [No.Panggil:2210511060] [Pembimbing 1: Widya Cholil] [Pembimbing 2: Novi Trisman Hadi] [Penguji 1: Indra Permana Solihin] [Penguji 2: Nindy Irzavika]
Uncontrolled Keywords: Software Defined Network, Quality of Service, Deep Reinforcement Learning, Deep Q-Network, Deep Deterministic Policy Gradient
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions: Fakultas Ilmu Komputer > Program Studi Informatika (S1)
Depositing User: MUHAMMAD RIZKY AULIA
Date Deposited: 02 Sep 2026 03:43
Last Modified: 02 Sep 2026 03:43
URI: http://repository.upnvj.ac.id/id/eprint/52412

Actions (login required)

View Item View Item