Simulation of multi-agent missions in an edge-cloud infrastructure to map and analyze a crisis situation: application to an earthquake

The Inria Centre at Rennes University is one of Inria's nine centres and has more than thirty research teams. The Inria Centre is a major and recognized player in the field of digital sciences. It is at the heart of a rich R&D and innovation ecosystem: highly innovative PMEs, large industrial groups, competitiveness clusters, research and higher education players, laboratories of excellence, technological research institute, etc.

Modern crisis management relies on distributed systems combining edge sensors, embedded computing resources, and cloud infrastructures. In this edge–cloud continuum, agents must collaborate in real time to map and analyze the situation and help authorities decide on priority actions. A major challenge is translating a high-level crisis scenario into a multi-agent operational mission, while taking into account:

• the distribution of resources between the edge (proximity, low latency) and the cloud (power, storage),
• quality constraints (time, accuracy, spatial coverage),
• and a sequential decision-making process that allows for dynamic adjustment of sub-missions.

Design and develop a simulator capable of generating and deploying a multi-agent mission based on a crisis scenario, incorporating:

1. A breakdown of the overall objective into sub-objectives down to activities that can be assigned to agents in the edge–cloud continuum;
2. Modeling sequential decision-making mechanisms (e.g., MDP, POMDP, decision trees) to adapt the mission to the evolution of the crisis;
3. The definition of interaction protocols between agents (edge–edge, edge–cloud, cloud–cloud), particularly for reporting information and triggering actions;

Taking into account the environment and its constraints (latency, bandwidth, availability, data transfer costs, etc.).

1. Crisis scenario modeling: Analysis of typical scenarios and extraction of decision variables, uncertainties, and temporal dependencies.
2. Conceptual framework: Proposal for a formalism enabling the translation of a scenario into missions and sub-missions, and integration of a decision structure.
3. Edge–cloud allocation: Definition of a sub-objective allocation strategy that takes resource variability into account.
4. Generation of interaction protocols: specification of information flows and coordination between agents, and management of uncertainty in sequential decisions (e.g., intermittent connectivity, partial observation, noisy data, etc.).
5. Software prototype: Implementation of a proof of concept (Python) enabling:
6. Evaluation: Experiments on simulated scenarios using a real Earthquake Early Warning dataset in a virtual edge-cloud environment, and analysis of the quality of the generated mission (time, accuracy, robustness).

• Edge computing for urgent/low-latency tasks,
• Cloud processing for complex or global analyses.

• the entry of a crisis scenario,
• the production of a multi-agent mission,
• the dynamic adaptation of the mission via a sequential decision-making model.

• Subsidized meals
• Partial reimbursement of public transport costs
• Leave: 7 weeks of annual leave + 10 extra days off due to RTT (statutory reduction in working hours) + possibility of exceptional leave (sick children, moving home, etc.)
• Possibility of teleworking (after 6 months of employment) and flexible organization of working hours
• Professional equipment available (videoconferencing, loan of computer equipment, etc.)
• Social, cultural and sports events and activities

Gratification environ 650€