Enhancing Consensus Efficiency and Security in DAG-Based Distributed Ledgers via an Adaptive Tip Selection Algorithm
Year: 2026
Type: conference-paper
Abstract: Directed Acyclic Graph (DAG)-based distributed ledgers offer parallel transaction processing suited to IoT environments, but their tip selection algorithms (TSAs) face a fundamental tension between scalability and security. The widely used Markov Chain Monte Carlo (MCMC) algorithm relies on a static parameter that cannot simultaneously optimize for low tip accumulation and resistance to parasite chain attacks. This paper introduces the Context-Aware Adaptive Tip Selection (CATS) algorithm, which resolves this trade-off through three modules: dynamic bias adaptation driven by real-time topological signals, statistical anomaly detection with adaptive thresholding, and fairness-preserving tip reintegration. We validate CATS through discrete-event simulations against MCMC, URTS, S-URTS, and G-IOTA across varying arrival rates and attack scenarios. Results demonstrate that CATS achieves tip count stability within 5% of URTS while reducing parasite chain tip selection probability to below 0.001 across all tested chain lengths, and provides a recommended reference configuration for practical deployment.
Authors: Diyas Nurullayev, Alibek Orynbek
Institution: Astana IT University
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7168300046