[Defence of thesis] 5/12/2025 - Samira HABLI: "Allocation of 5G Resources in Smart Cities" (UPR LIA)

Ms Samira HABLI will publicly defend her thesis entitled: "Allocation of 5G Resources in Smart Cities", supervised by Mr Rachid ELAZOUZI and Essaid SABIR, under joint supervision with Hassan II University (Morocco), on Friday 05 December 2025 at 2 pm.

Laboratory

UPR 4128 LIA - Avignon Computing Laboratory

Composition of the jury

Mr Rachid ELAZOUZI	Avignon University	Thesis supervisor
Mr Essaid SABIR	TÉLUQ University	Thesis co-supervisor
Mr Mohammed SADIK	Ecole Nationale Supérieure d'Electricité et de Mécanique de Casablanca Hassan II University	Thesis co-director
Ms Johanne COHEN	CNRS/LISN	Rapporteur
Mr Oussama HABACHI	LIMOS University of Clermont Auvergne	Rapporteur
Ms Lila BOUKHATEM	Paris-Saclay University	Examiner
Ms Nancy PARROT	Orange Innovation	Examiner
Mr Francesco DE PELLEGRINI	Avignon University	Examiner

Summary

The rapid evolution of fifth-generation (5G) and next-generation communications networks is profoundly transforming the way digital services are delivered, consumed and managed. These next-generation networks must support a wide range of applications, from ultra-low latency services such as autonomous driving, industrial automation or augmented reality, to data-intensive services based on cloud computing. This diversity places increasing demands on quality of service (QoS), scalability and equity of access to resources. To meet these challenges, network architectures are evolving towards distributed models, taking advantage of the complementary capabilities of cloud and fog computing. In this complex and competitive ecosystem, Service Providers (SPs) interact with Infrastructure Providers (InPs) to acquire computing resources, while defining pricing and quality strategies to attract and retain end-users. This thesis proposes an analytical framework based on game theory to model and study the economic and operational dynamics of pricing and resource allocation in multi-player environments. In the first part, we analyse the competition between non-cooperative SPs that purchase computing resources from a cloud-based InP. Each SP seeks to maximise its profit by taking into account its own price and quality choices, as well as those of its competitors. This behaviour is modelled using a non-cooperative game, and a distributed learning algorithm is proposed to allow SPs to progressively adjust their strategies in response to market conditions. The second part extends the analysis to a hybrid infrastructure combining fog and cloud. Fog nodes, located close to users, provide better responsiveness but have limited resources. Conversely, cloud servers offer high capacity at the cost of higher latency. SPs naturally tend to favour fog resources to improve QoS. However, uniform pricing can lead to inequitable allocation, with some economically more powerful SPs being able to monopolise available resources. In order to remedy this situation, we propose a pricing mechanism differentiated according to usage, favouring a fairer distribution and more efficient use of fog resources. Finally, the last part of the thesis explores the issues associated with net neutrality by studying the role that a content provider (CP) can play in competition between Internet service providers (ISPs). Using Bayesian games, we analyse scenarios in which the CP can choose to publicly or privately disclose information about the popularity of its content, thereby influencing the competitive balance between ISPs. Overall, this thesis provides a unified analytical framework to better understand the interaction between pricing, quality of service and fairness in next-generation smart grids. It contributes both to the enrichment of the theoretical foundations of game-theoretic resource management and to the design of sustainable, fair and user-centric mechanisms for the future digital infrastructures of smart cities.

Keywords Allocation, 5G, Resource