Skripsi

OPTIMASI PERFORMANCE METODE PROOF OF PRIORITY (PoP) MENGGUNAKAN ALGORITMA KONSENSUS BLOCKCHAIN PADA SISTEM TRANSAKSI ORDER DI INDUSTRI MANUFAKTUR

---

Digitalization is currently being continuously promoted by the government to address future challenges, which are classified into three main sectors: digital infrastructure, digital economy, and digital society. One of the key efforts undertaken is the continuous transformation toward a digital economy. Blockchain technology, which has the potential to drive this transformation, requires robust mechanisms capable of addressing future challenges. Blockchain is characterized by its decentralized, immutable, and transparent nature. However, one of the major challenges in blockchain systems lies in the scalability of consensus algorithms, particularly related to execution time and computational overhead. This study formulates several research questions: (1) How can feature selection preprocessing be conducted to obtain optimal features using machine learning algorithms? (2) How can the Proof of Priority (PoP) method be optimized within blockchain consensus algorithms in the manufacturing industry? (3) How can the performance of blockchain consensus algorithms be measured in blockchain-based transaction systems? Based on the literature review, feature selection approaches have the potential to improve the performance of order transaction systems. To validate this assumption, experiments were conducted using several feature selection methods, namely Information Gain, Gain Ratio, and Chi-Square. To evaluate classification accuracy, multiple classification algorithms were employed, including Naïve Bayes, Random Tree, and Random Forest. The results indicate that feature selection methods improve the performance of blockchain-based order transaction systems; however, further optimization is required to comprehensively measure performance. Subsequently, this study performed feature selection optimization using the Chi-Square method, which resulted in the selection of 3 features out of 14 original features. The optimized feature selection demonstrated improved detection performance, enhancing the Naïve Bayes classifier with an accuracy of 97.65%, recall of 0.999, precision of 0.999, and a false positive rate (FPR) of 0.002. The execution time obtained using a comparative PoW method after feature selection was 0.14 seconds. Experimental results further show that the Proof of Priority (PoP) method significantly reduces execution time across various difficulty levels. At difficulty level 1, execution time decreased from 0.25 seconds to 0.02 seconds; at level 2, from 0.70 seconds to 0.55 seconds; at level 3, from 0.35 seconds to 0.37 seconds; and at level 4, from 0.65 seconds to 0.39 seconds. On average, PoP reduced execution time by 38% compared to conventional Proof of Work (PoW). This study provides a theoretical contribution by integrating machine learning techniques with blockchain consensus mechanisms, as well as a practical contribution by offering an efficient solution for order transaction systems in the manufacturing sector.

---

Availability

Detail Information

Series Title

-

Call Number

T1910292026

Publisher

Indralaya : Prodi Doktor Ilmu Teknik, Fakultas Teknik Universitas Sriwijaya., 2026

Collation

xv, 152 hlm.; ilus.; tab.; 29 cm.

Language

Indonesia

ISBN/ISSN

-

Classification

004.678 06

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Blockchain
Prodi Doktor Ilmu Teknik

Specific Detail Info

-

Statement of Responsibility

TUTI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment