博士研究生岗位 F/M：电信基础设施对气候变化的韧性研究——下一代网络的建模、影响分析及适应策略

该项目旨在开发一种系统化的自动化数据驱动方法，以评估并提升电信网络应对气候变化影响的韧性。其中将涉及整合天气与事件数据，构建统计模型以识别故障成因，并在气候情景下模拟未来网络行为。

Main activities: Context and Problem Statement Telecommunication infrastructures (fixed and mobile access networks, backhaul networks) are critical components of modern socio-economic functioning. However, as highlighted by ARCEP in its note on network resilience [6], these physical infrastructures are highly exposed to the direct and indirect impacts of climate change. Extreme weather events (heatwaves, storms, floods, wildfires) as well as long-term trend shifts (rising average temperatures, sea level rise) degrade service quality, increase failure rates, and threaten to weaken isolated areas [1, 2]. While initial scientific research and exploratory studies (such as Master's thesis work based on local open data) have established empirical correlations between weather conditions and network incidents on a small scale, several major scientific roadblocks remain to generalize these approaches to a national and predictive scale: Spatio-temporal heterogeneity and granularity of data: Crossing complex time series (weather metrics, RIPE Atlas active/passive probes, ARCEP/operator incident reports, physical network topology) requires processing noisy, incomplete, and asynchronous data. Causality vs. Correlation: Distinguishing failures directly caused by climate factors from those resulting from cascading failures (e.g., power grid outages, traffic saturation during crises). Prospective uncertainty: Anticipating the vulnerability of future network generations (5G standalone, 6G, fiber densification) under various IPCC climate scenarios for the 2030–2050 horizon. Scientific Objectives of the Thesis The main objective of this thesis is to build a systematic, automated, data-driven methodology to assess, predict, and enhance the resilience of telecommunication networks in the face of climate hazards. The research will focus on the following axes: 1. Data Aggregation, Coupling, and Advanced Mapping Extending the cross-analysis between meteorological time series (Météo-France / Copernicus), incident data (ARCEP), and fixed/mobile access quality (RIPE Atlas, crowdsourced data) across the entire French territory. Developing standardized metrics to measure the weather-climate vulnerability index of a piece of equipment or a network node. 2. Explanatory Modeling for Causality Detection Designing statistical models capable of identifying root causes of incidents and detecting cascading failure propagation (e.g., the impact of a heatwave on cooling systems in cell sites or optical distribution frames / NROs). Accounting for cross-dependencies with other critical infrastructures (specifically the ENEDIS/RTE power distribution network). 3. Prospective Inference and Resilience Optimization Strategies Simulating network behavior under various future extreme climate scenarios (IPCC / Météo-France projections). Proposing redeployment strategies, hardware hardening, or emergency dynamic routing to minimize the risk of global disconnection during extreme events. Scientific Roadblocks and Expected Contributions Scaling and automation: Moving from a local exploratory analysis (departmental scale) to an automated data processing pipeline for high-volume streams at a national scale. Systemic resilience: Providing a unified theoretical framework combining physical infrastructure characteristics (material exposure) with the dynamic behavior of data traffic flows. Public and industrial decision support: Delivering a risk assessment tool that regulators (ARCEP, ANSSI) and operators (Orange, SFR, Bouygues Telecom, Free, Axione) can use to guide network hardening investments. Bibliographic References [1] Padmanabhan, R., et al. "Residential links under the weather." Proc. of ACM SIGCOMM , 2019. [2] Durairajan, R., Barford, C., & Barford, P. "Lights out: Climate change risk to internet infrastructure." Proc. of ANRW , 2018. [3] Météo-France — Open data platform (meteo.data.gouv.fr). [4] RIPE NCC — RIPE Atlas Infrastructure (atlas.ripe.net). [5] ARCEP — Incident tracking and mobile network status (arcep-dev.github.io/siteshs/). [6] ARCEP — Networks of the future: The resilience of electronic communications networks, Summary Note.

技能： 所需的技术技能与水平： 语言： 关系技能： 其他受重视的优点： 成功的关键： 计算机科学、数据科学或通信领域的硕士或工程学位。 在数据分析、机器学习/深度学习和时间序列分析方面具备扎实的技能。 对电信网络与互联网协议有扎实的理解。 对跨学科挑战（计算机科学、网络、气候、环境）具有浓厚的兴趣。

福利包： 补贴餐食 75%的公共交通票费可获得报销 休假规定：每年7周年假 + 因RTT（法定工作时间减少）额外享有10天假期 + 可能享受特殊假期（如照顾生病的子女、搬家等）。 可实行远程工作（在办公室工作数月后），且工作时间安排具有灵活性 提供专业设备（视频会议、计算机设备借用等） 社会、文化和体育活动及项目 职业培训机会 社会保障覆盖范围 薪酬： 税前每月薪资为2300欧元 36个月的固定期限合同，博士学生身份 为期2个月的不可续签试用期 申请该职位

申请说明： 防御安全：该职位可能位于受限区域（ZRR），该区域依据第2011-1425号法令关于保护国家科技潜力的规定而划定（PPST）。根据2012年7月3日关于PPST的法令，单位主管在获得部长批准后，将授予进入该区域的许可。 关于位于 ZRR 地区的职位所涉及的不利部长级决定，将导致该聘用被取消。 招聘政策：作为其多元化政策的一部分，法国国家信息与自动化研究所的所有职位均向残疾人开放。