Ali Deylamy

Modern mobile networks are becoming increasingly intelligent and automated. In the latest generation, known as Open Radio
Access Networks (O-RAN), network control is no longer managed by a single vendor but instead by Many vendors each
providing a small software applications called xApps. Each xApp focuses on a specific task, such as improving signal
strength, managing handovers, or balancing network load. While this open approach allows for more innovation and
flexibility, it also creates a new problem: different xApps can unintentionally interfere with each other. When this happens, the
network’s performance can suddenly drop for example, users may experience higher latency or lower data speeds but
identifying which xApp caused the problem is extremely difficult.
My research aims to address this challenge by advancing the understanding of intelligent decision-making and fault
diagnosis mechanisms in open and distributed mobile networks. Specifically, I propose to investigate how Digital Twin–based
simulations can be used as a controlled environment to study, model, and validate the causal relationships between network
actions and performance degradations. The project will explore methodologies for identifying, tracing, and quantifying the
impact of autonomous xApp decisions, and for designing a feedback loop that predicts and prevents conflicts before they
propagate to the live network. Rather than developing a deployable system, the objective is to generate new theoretical
insights and reproducible models that improve the explainability, reliability, and adaptability of AI-driven network management
frameworks.
The outcome of this research will be a self-healing network that can detect, explain, and correct its own problems with
minimal human intervention. This work will make future mobile networks more reliable, energy-efficient, and secure,
contributing to the global effort to develop fully autonomous 6G communication systems that serve society with faster,
smarter, and more dependable connectivity.