Ericsson is exploring the application of quantum computing principles to a practical challenge facing mobile networks: optimizing Radio Access Network (RAN) planning. The company is investigating combining classical algorithms with quantum computing to address the increasing complexity of 5G deployments and prepare for the demands of 6G, representing a shift toward potential network improvements. This hybrid approach aims to solve complex network planning and optimization tasks more efficiently, enabling “smarter, more accurate decisions,” according to Ericsson’s blog post. The research team, spanning Quantum Compute, Cloud Intelligence, and Cognitive Network Solutions within Ericsson’s Business Area Cloud Software & Services, is focused on exploring pathways toward fully autonomous, real-time next-generation networks by overcoming the limitations of traditional, static planning methods.
Rising Network Complexity Demands New Planning Methods
Scaling networks now presents challenges far beyond those faced in previous generations. Dynamic user behavior, fluctuating traffic, and increasingly dense topologies are creating a computational burden that conventional tools struggle to manage. Operators are not simply seeking to maintain seamless service, but to do so while efficiently managing resources, adapting to evolving usage patterns, and proactively optimizing performance, all within strict budgetary constraints. Even with increasing automation, existing heuristic-based tools often rely on simplified models or approximations to maintain computational feasibility, potentially leading to suboptimal network configurations when conditions change rapidly. Traditional optimization cycles, requiring hours or even days to recompute configurations, are proving inadequate for the speed demanded by modern users. This has forced operators to prioritize practicality over exhaustive exploration of the vast solution space. A particularly acute example of this challenge lies in Tracking Area (TA) planning.
As networks become denser and mobility patterns less predictable, conventional TA designs struggle to deliver optimal results. Tracking Areas define the regions where idle-mode users are tracked, and their size presents a critical trade-off. “If TAs are too small, frequent TA updates increase signaling overhead; if too large, paging messages must be broadcast across unnecessarily wide areas, consuming radio and core resources,” highlighting the delicate balance required for efficient resource allocation and user experience. Advanced computational approaches are therefore essential to navigate these complexities.
This represents a fundamental shift in telecommunications, moving away from heuristic-driven planning toward intelligent, automated, and computationally advanced network management. “As networks grow in scale, complexity, and diversity of services, the ability to optimize configurations efficiently and proactively will directly influence performance, cost, and customer satisfaction.” Research is exploring a hybrid quantum-classical approach, combining the strengths of both computing paradigms to tackle problems that exceed the capabilities of classical systems alone. This isn’t simply applying quantum computing as a theoretical exercise; it’s about researching a practical solution for a pressing, real-world problem, and the next sections will explore how this hybrid approach provides a promising path forward.
Hybrid Quantum-Classical Intelligence for Autonomous Networks
The pursuit of fully autonomous networks is driving telecommunications toward an unexpected convergence: quantum computing. While still in its nascent stages, work at Ericsson explores the application of quantum principles to Radio Access Network (RAN) planning, representing a significant departure from theoretical exploration and a research direction for practical implementation within existing infrastructure. Ericsson is exploring the development of hybrid quantum-classical intelligence systems, combining the strengths of both computational approaches to address the escalating complexities of modern mobile networks. This isn’t simply about adding quantum processing power; it’s about fundamentally reimagining how networks are planned and optimized. Rising network complexity, fueled by 5G deployments and preparations for 6G, is pushing traditional planning methods to their limits.
Operators now grapple with dynamic user behavior, fluctuating traffic demands, and increasingly dense network topologies, creating challenges that stretch conventional tools. “TA planning must balance signaling efficiency, mobility behavior, and paging performance,” and this research aims to address this directly. Ericsson is investigating a hybrid approach, acknowledging the current limitations of Noisy Intermediate-Scale Quantum (NISQ) devices. Rather than attempting end-to-end quantum planning, they are integrating classical algorithms with quantum techniques. Classical systems initially generate a feasible TA layout, which is then refined through quantum exploration of alternative partitions. This synergy, they believe, provides both stability and the ability to delve into complex solution spaces. “Because today’s quantum devices are still in the NISQ…stage, they cannot yet handle end-to-end network planning alone,” necessitating this combined methodology. The workflow involves three stages: classical data processing and initial clustering, quantum-driven refinement, and optimization. Classical computation structures and potentially reduces the problem, while quantum techniques could evaluate candidate configurations efficiently by sampling diverse possibilities in parallel.
Faster, smarter, and more scalable planning enables operators to proactively respond to dynamic traffic patterns and evolving service requirements, ensuring optimal user experiences while controlling operational costs.
Tracking Area Planning Challenges in Dense Networks
Ericsson researchers are exploring how to address the escalating complexities of Radio Access Network (RAN) planning, with a particular focus on optimizing Tracking Area (TA) configurations within increasingly dense mobile networks. Muhammad Asad Ullah, Senior Researcher in Quantum Compute, and colleagues are investigating a hybrid approach that integrates classical algorithms with the emerging capabilities of quantum computing to overcome limitations inherent in traditional network management strategies. Rising network complexity is pushing traditional tools to their limits; massive data volumes and the intricate interplay between coverage, capacity, and mobility create substantial computational hurdles. Operators are compelled to balance seamless service delivery with efficient resource management, all while controlling operational costs. “One clear example of this evolution is Tracking Area (TA) planning,” the researchers note, highlighting the specific difficulties in designing these regions for efficient user tracking.
The team’s work centers on leveraging quantum computing’s ability to explore a vast solution space simultaneously, a capability that surpasses the sequential approach of classical computers. “Imagine trying to find the fastest route through a city with thousands of roads. A classical computer checks one route at a time, while a quantum computer can consider many routes simultaneously—making it much faster at finding the best solution,” they explain. This is particularly relevant to TA planning, where the number of possible arrangements grows exponentially, rendering exhaustive exploration impractical.
As evident, the hybrid Louvain-quantum (Louvain-QUBO) approach achieved the strongest overall performance: inter-TA handovers were reduced by 36.4%, while the maximum paging traffic and the maximum number of cells per TA were also decreased by 25.0% and 13.6%, respectively; all indicating a more efficient TA planning.
Quantum Computing Enables Parallel Solution Exploration
The pursuit of fully autonomous, real-time next-generation networks is now being advanced by an unexpected alliance: classical computing paired with the emerging power of quantum processors. As operators scale 5G deployments and look toward 6G capabilities, traditional static planning methods struggle to adapt to dynamic user behavior and increasingly dense network topologies. This isn’t simply about faster processing; it’s about fundamentally altering how network configurations are determined. TA planning, defining regions for tracking idle-mode users, presents a delicate balance; too-small areas increase signaling overhead, while overly large areas waste resources. Muhammad Asad Ullah, Senior Researcher, Quantum compute, Mbarka Soualhia, Senior Researcher, Cloud intelligence, Rana Pratap Sircar, Senior Research Manager, Juan Ramiro, Head of Research & Innovation, Cognitive Network Solutions, and Adriano Mendo, Research Specialist, are all contributors to this research. Recognizing the limitations of current “NISQ (Noisy Intermediate–Scale Quantum)” devices, the approach is inherently hybrid.
This synergy provides both stability from classical processing and the exploratory depth of quantum computation, enabling high-quality TA planning at scales relevant to future networks. This approach aims to reduce the problem’s scale before exploring complex relationships with a quantum layer. Although current evaluations utilize quantum simulators, the underlying circuit structures are designed to be compatible with maturing quantum processors, indicating potential for future hardware implementation.
NISQ Era Drives Hybrid Pipeline Development
Rather than awaiting fault-tolerant quantum computers, Ericsson is exploring the integration of quantum computing principles with existing classical algorithms, specifically for Radio Access Network (RAN) planning, a move that prioritizes immediate gains over theoretical perfection. Rising complexity in network management stems from the demands of 5G scaling and preparation for 6G, where traditional planning methods struggle with dynamic user behavior and ultra-dense topologies. “Tracking Areas (TA) define the regions in which idle-mode users are tracked.” The approach involves a three-stage hybrid workflow. Initial data processing and clustering are handled classically, leveraging techniques like spectral clustering to create baseline TA candidates. This reduces the problem’s scale while preserving essential relationships. “Although the evaluations in this work were conducted using a quantum simulator, the formulation and circuit structures are designed to be compatible with maturing quantum processors.” This hybrid approach acknowledges the current limitations of NISQ hardware, benefiting from classical preprocessing and post-processing to deliver stable and exploratory depth, ultimately enabling high-quality TA planning at scales suited to next-generation networks.
