Towards xApp Conflict Evaluation with Explainable Machine Learning and Causal Inference in O-RAN

2025-10-14

AI Summary

In O-RAN architectures, concurrent execution of multiple xApps frequently induces conflicts among Radio Access Network Control Parameters (RCPs), degrading key performance indicators (KPIs). To address this, we propose the first xApp conflict assessment framework integrating interpretable machine learning with causal inference. Specifically, we employ SHAP values to quantify the local impact of individual RCPs on KPIs; apply causal discovery algorithms to construct a directed acyclic graph (DAG) capturing causal relationships between RCPs and KPIs; and estimate average and conditional average treatment effects (ATE/CATE) to rigorously quantify the causal impact of RCP adjustments. This enables precise identification of conflicting RCP combinations that jointly influence the same KPI. The framework delivers interpretable, evidence-based conflict attribution and supports operators in formulating data-driven, differentiated mitigation strategies. Extensive evaluation—spanning both simulation and real-world deployments—demonstrates its effectiveness, robustness, and operational deployability.

Technology Category

• Machine Learning: Causal Learning
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Planning, Routing, and Scheduling: Plan Execution and Monitoring

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

The Open Radio Access Network (O-RAN) architecture enables a flexible, vendor-neutral deployment of 5G networks by disaggregating base station components and supporting third-party xApps for near real-time RAN control. However, the concurrent operation of multiple xApps can lead to conflicting control actions, which may cause network performance degradation. In this work, we propose a framework for xApp conflict management that combines explainable machine learning and causal inference to evaluate the causal relationships between RAN Control Parameters (RCPs) and Key Performance Indicators (KPIs). We use model explainability tools such as SHAP to identify RCPs that jointly affect the same KPI, signaling potential conflicts, and represent these interactions as a causal Directed Acyclic Graph (DAG). We then estimate the causal impact of each of these RCPs on their associated KPIs using metrics such as Average Treatment Effect (ATE) and Conditional Average Treatment Effect (CATE). This approach offers network operators guided insights into identifying conflicts and quantifying their impacts, enabling more informed and effective conflict resolution strategies across diverse xApp deployments.

Problem

• Evaluating conflicting control actions from multiple xApps
• Identifying causal relationships between RCPs and KPIs
• Quantifying conflict impacts using causal inference metrics

Innovation

• Uses explainable machine learning for conflict detection
• Applies causal inference to quantify parameter impacts
• Models interactions as causal graphs for analysis