Public defence, Computer Science, MSc Sara Ranjbaran

Title of the thesis: Cooperative Resource Scheduling in Heterogeneous Edge-IoT Networks

Thesis defender: Sara Ranjbaran
Opponent: Full Professor Luigi Atzori, University of Cagliari, Italy
Custos: Senior Advisor Antti Ylä-Jääski, Aalto University School of Science

Better rules for cooperation in connected computing systems 

Smart homes, cities and industries increasingly depend on connected devices that collect data and provide digital services. These services require computing power, storage and communication capacity, often spread across devices and service providers. Sharing these resources could help meet growing demand, but cooperation is not automatic: participants have different capabilities, costs and interests. 

Sara Ranjbaran’s doctoral thesis investigates how algorithms and incentives can support cooperative resource sharing in edge computing and the Internet of Things. Edge computing brings data processing closer to devices and users, helping services respond quickly. The research asks how resources should be allocated, why their owners would choose to share them, and how decisions can be coordinated across a network. 

The central finding is that effective cooperation requires allocation rules suited to both the technical environment and participants’ incentives. The dissertation connects distributed computing, economic mechanism design and online learning to address these challenges through complementary approaches. 

One theoretical contribution shows how certain distributed allocation algorithms can be combined with locally computed payments that encourage honest reporting of private information. Another develops efficient distributed matching under shared preferences, preserving incentives for honest reporting among participants who propose matches. These results extend understanding of when resource allocation can be coordinated through local communication. 

The applied contributions develop auction mechanisms for trading combinations of resources and incentive models that help resource owners form cooperative groups and share revenues. The research also introduces an online learning approach for deciding where to host Digital Twins—digital representations of physical systems—as service demand changes. Its analysis provides a performance guarantee relative to the best fixed placement selected after the requests are known. Mathematical analysis and simulations establish the properties and evaluate the behaviour of the approaches within their respective models. 

The findings provide design principles for future resource-sharing platforms and services in smart cities, industrial monitoring and other connected environments. The overall conclusion is that cooperation must be designed: choosing suitable allocation rules, participation incentives and coordination methods is essential to making distributed resources work together.

Keywords: Internet of Things (IoT); edge computing; cooperative resource scheduling; mechanism design; distributed algorithms; incentive mechanisms; stable matching; combinatorial auctions; online learning; digital twins.

Thesis available for public display 7 days prior to the defence at Aalto University's public display page. 

Contact Information:
sara.ranjbaran@aalto.fi