DYNAMIC ALGORITHM FOR TIP SELECTION IN MULTI-AGENT BLOCKCHAIN ENVIRONMENTS: OPTIMIZATION AND ANALYSIS

Abstract

The rise of decentralized technologies has brought forward novel challenges in maintaining fairness, efficiency, and scalability in distributed ledger protocols. The Internet of Things Applications (IOTA), a Directed Acyclic Graph (DAG)-based structure, offers an alternative to traditional blockchain systems by enabling scalable and feeless transactions, particularly suited for the Internet of Things (IoT). However, ensuring fairness and addressing issues such as orphaned transactions and lazy behavior in such a decentralized network remain significant challenges. This dissertation introduces a novel Partially Observable Markov Decision Process (POMDP)-based Tip Selection Algorithm (TSA) aimed at optimizing fairness in the IOTA Tangle. Unlike existing algorithms, the proposed method reduces orphaned transactions to as low as 0.003% in medium-activity networks and 0.013% in high-activity networks, significantly outperforming other TSAs. POMDP-based TSA also avoids lazy tip selection with 0.000% lazy tips in low to medium network loads, ensuring fairness in tip selection even under network congestion. Through extensive simulation experiments, the performance of the proposed algorithm is compared to other established TSAs, including Random, Unweighted, Weighted, Hybrid-1, Hybrid-2, G-IOTA, and E-IOTA, across various network conditions. Results indicate that the POMDP-based TSA confirms up to 107 transactions at optimal arrivate rate (lambda values), outperforming the next best-performing algorithm (Weighted TSA, confirming 25 transactions) by 328% in efficiency. The algorithm maintains its robust performance even in high-throughput scenarios, demonstrating both scalability and adaptability. These findings contribute to the ongoing development of DAG-based distributed ledger systems by offering a scalable and secure TSA. The proposed algorithm's novel approach to handling orphaned and lazy transactions offers significant improvements over state-of-the-art TSAs, making it a robust solution for IoT-based decentralized applications.