pr\u00e9sentent<\/p>\n\n\n\n
Autoris\u00e9 \u00e0 pr\u00e9senter ses travaux en vue de l\u2019obtention du Doctorat de l’Institut Polytechnique de Paris, pr\u00e9par\u00e9 \u00e0 T\u00e9l\u00e9com SudParis en :<\/p>\n\n\n\n
le LUNDI 16 D\u00e9CEMBRE 2024 \u00e0 14h00<\/p>\n\n\n\n
\u00e0<\/p>\n\n\n\n
Amphi 11
9 rue Charles Fourier – 91000 \u00c9vry-Courcouronnes<\/p>\n\n\n\n
Membres du jury :<\/strong><\/p>\n\n\n\n M. Denis CONAN<\/strong>, Associate Professor, T\u00e9l\u00e9com SudParis, FRANCE – Directeur de these Invit\u00e9s :<\/strong><\/p>\n\n\n\n M. Djamel BELAID<\/strong>, Professeur, T\u00e9l\u00e9com SudParis<\/p>\n\n\n\n M. Ajay KATTEPUR<\/strong>, Senior researcher, Ericsson AI Research<\/p>\n\n\n\n R\u00e9sum\u00e9 :<\/strong><\/p>\n\n\n\n The proliferation of Internet of Things (IoT) devices has led traditional spaces to become smarter and more interconnected. This has resulted in sophisticated IoT systems that are typically composed of devices that sense physical phenomena and generate data that may be processed by computing nodes before being consumed by applications. Such applications typically define specific Quality-of-Service (QoS) requirements that have to be met (e.g., availability, accuracy, latency). For this purpose, IoT systems are usually configured to ensure that the QoS requirements of deployed applications are respected. This involves adjusting multiple parameters such as network settings, processing resources, and tuning data exchange systems. However, modern smart spaces are inherently dynamic and unpredictable. Changes in the number of IoT devices, network conditions, and available computational resources create a continuously evolving environment. Thus, to ensure that IoT systems operate autonomously, it is essential to design self-adaptive mechanisms to maintain QoS requirements of applications in dynamic situations. This thesis proposes a middleware-based, hybrid Artificial Intelligence (AI) self-adaptation approach for enabling autonomous IoT operations in dynamic environments. By combining logic-based approaches with data-driven AI techniques, we design effective and explainable self-adaptation solutions for IoT systems. This is achieved through three main contributions. First, queueing networks are leveraged to compose QoS models that represent IoT systems under different situations and\/or configurations. By simulating QoS models, we are able to generate performance metrics datasets that can be used in self-adaptive approaches to dynamically adapt to changing situations by selecting the configuration that best satisfies the QoS requirements defined by applications. The second contribution enables AI-driven adaptive data flow management in IoT enhanced spaces. By combining logic-based and data-driven AI techniques such as AI planning and Reinforcement Learning, we design a framework involving intelligent agents capable of taking adaptation decisions at runtime. Possible adaptation actions include data flow configurations (e.g., priorities, drop rates) and resource control (e.g., network resources, computing resources). Finally, the last contribution enables more proactive and explainable autonomous systems through Causal Reinforcement Learning. To achieve this, we rely on the Causality framework to provide a formal analysis of the performance of IoT systems. Subsequently, causal models enable a more effective decision-making process that enables adaptation agents to take efficient adaptation actions in dynamic environments. We validate our approach by developing a prototype implementation of an IoT system and experimenting with case studies considering different types of IoT environments. Our QoS models are evaluated and compared with a prototype implementation for validating the accuracy of the generated performance metrics datasets. We then evaluate the effectiveness of our solution by leveraging data from real deployments to ensure that our approach is valid in real-life settings.<\/p>\n\n\n\n La prolif\u00e9ration de l’Internet des objets (IoT) a transform\u00e9 les espaces traditionnels en environnements plus intelligents et interconnect\u00e9s. Cela a donn\u00e9 lieu \u00e0 des syst\u00e8mes IoT sophistiqu\u00e9s, g\u00e9n\u00e9ralement compos\u00e9s de dispositifs qui d\u00e9tectent des ph\u00e9nom\u00e8nes physiques et g\u00e9n\u00e8rent des donn\u00e9es pouvant \u00eatre trait\u00e9es par des n\u0153uds de calcul avant d’\u00eatre consomm\u00e9es par des applications. Ces applications d\u00e9finissent g\u00e9n\u00e9ralement des exigences sp\u00e9cifiques de qualit\u00e9 de service (QoS) qui doivent \u00eatre respect\u00e9es (par exemple, disponibilit\u00e9, pr\u00e9cision, latence). \u00c0 cette fin, les syst\u00e8mes IoT sont souvent configur\u00e9s pour garantir que les exigences de QoS des applications d\u00e9ploy\u00e9es soient respect\u00e9es. Cela implique l’ajustement de plusieurs param\u00e8tres tels que les configurations r\u00e9seau, les ressources de traitement, et l’optimisation des syst\u00e8mes d’\u00e9change de donn\u00e9es. Cependant, les espaces intelligents modernes sont intrins\u00e8quement dynamiques et impr\u00e9visibles. Les variations dans le nombre de dispositifs IoT, les conditions du r\u00e9seau, et les ressources de calcul disponibles cr\u00e9ent un environnement en constante \u00e9volution. Ainsi, pour garantir que les syst\u00e8mes IoT fonctionnent de mani\u00e8re autonome, il est essentiel de concevoir des m\u00e9canismes d’auto-adaptation pour maintenir les exigences de QoS des applications dans des situations dynamiques. Cette th\u00e8se propose une approche d’auto-adaptation \u00e0 base d’intergiciel et d’intelligence artificielle (IA) hybrid pour permettre des op\u00e9rations autonomes des syst\u00e8mes IoT dans des environnements dynamiques. En combinant des approches bas\u00e9es sur la logique avec des techniques d’IA bas\u00e9es sur les donn\u00e9es, nous concevons des solutions d’auto-adaptation efficaces et explicables pour les syst\u00e8mes IoT. Cela est r\u00e9alis\u00e9 \u00e0 travers trois contributions principales. Tout d’abord, des r\u00e9seaux de files d’attente sont utilis\u00e9s pour composer des mod\u00e8les de QoS qui repr\u00e9sentent les syst\u00e8mes IoT sous diff\u00e9rentes situations et\/ou configurations. En simulant ces mod\u00e8les de QoS, nous sommes en mesure de g\u00e9n\u00e9rer des ensembles de donn\u00e9es de performances qui peuvent \u00eatre utilis\u00e9s dans des approches auto-adaptatives pour s’adapter dynamiquement aux situations changeantes en s\u00e9lectionnant la configuration qui satisfait le mieux les exigences de QoS d\u00e9finies par les applications. La deuxi\u00e8me contribution permet une gestion adaptative des flux de donn\u00e9es pilot\u00e9e par l’IA dans les espaces IoT. En combinant des techniques d’IA bas\u00e9es sur la logique et sur les donn\u00e9es, telles que l’AI planning et l’apprentissage par renforcement, nous concevons un cadre impliquant des agents intelligents capables de prendre des d\u00e9cisions d’adaptation en temps r\u00e9el. Les actions d’adaptation possibles incluent les configurations de flux de donn\u00e9es (par exemple, priorit\u00e9s, taux de rejetde donn\u00e9es) et le contr\u00f4le des ressources (par exemple, ressources r\u00e9seau, ressources de calcul). Enfin, la derni\u00e8re contribution permet de rendre les syst\u00e8mes autonomes plus proactifs et explicables gr\u00e2ce \u00e0 l’apprentissage par renforcement et la th\u00e9orie de causalit\u00e9. Pour ce faire, nous nous appuyons sur le cadre de la causalit\u00e9 afin de fournir une analyse formelle des performances des syst\u00e8mes IoT. Par la suite, les mod\u00e8les causaux permettent un processus de prise de d\u00e9cision plus efficace, permettant aux agents d’adaptation de prendre des actions d’adaptation efficaces dans des environnements dynamiques. Nous validons notre approche en d\u00e9veloppant une impl\u00e9mentation prototype d’un syst\u00e8me IoT et en menant des exp\u00e9riences sur des \u00e9tudes de cas consid\u00e9rant diff\u00e9rents types d’environnements IoT. Nos mod\u00e8les de QoS sont \u00e9valu\u00e9s et compar\u00e9s \u00e0 une impl\u00e9mentation prototype pour valider la pr\u00e9cision des ensembles de donn\u00e9es de performances g\u00e9n\u00e9r\u00e9s. Nous \u00e9valuons ensuite l’efficacit\u00e9 de notre solution en exploitant des donn\u00e9es issues de d\u00e9ploiements r\u00e9els pour nous assurer que notre approche est valide dans des contextes r\u00e9els.<\/p>\n","protected":false},"excerpt":{"rendered":" L’Ecole doctorale : Ecole Doctorale de l’Institut Polytechnique de Paris et le Laboratoire de recherche SAMOVAR – Services r\u00e9partis, Architectures, Mod\u00e9lisation, Validation, Administration des R\u00e9seaux pr\u00e9sentent l\u2019AVIS DE SOUTENANCE de Monsieur Houssam HAJJ HASSAN Autoris\u00e9 \u00e0 pr\u00e9senter ses travaux en vue de l\u2019obtention du Doctorat de l’Institut Polytechnique de Paris, pr\u00e9par\u00e9 \u00e0 T\u00e9l\u00e9com SudParis en […]<\/p>\n","protected":false},"author":4,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ocean_post_layout":"","ocean_both_sidebars_style":"","ocean_both_sidebars_content_width":0,"ocean_both_sidebars_sidebars_width":0,"ocean_sidebar":"","ocean_second_sidebar":"","ocean_disable_margins":"enable","ocean_add_body_class":"","ocean_shortcode_before_top_bar":"","ocean_shortcode_after_top_bar":"","ocean_shortcode_before_header":"","ocean_shortcode_after_header":"","ocean_has_shortcode":"","ocean_shortcode_after_title":"","ocean_shortcode_before_footer_widgets":"","ocean_shortcode_after_footer_widgets":"","ocean_shortcode_before_footer_bottom":"","ocean_shortcode_after_footer_bottom":"","ocean_display_top_bar":"default","ocean_display_header":"default","ocean_header_style":"","ocean_center_header_left_menu":"","ocean_custom_header_template":"","ocean_custom_logo":0,"ocean_custom_retina_logo":0,"ocean_custom_logo_max_width":0,"ocean_custom_logo_tablet_max_width":0,"ocean_custom_logo_mobile_max_width":0,"ocean_custom_logo_max_height":0,"ocean_custom_logo_tablet_max_height":0,"ocean_custom_logo_mobile_max_height":0,"ocean_header_custom_menu":"","ocean_menu_typo_font_family":"","ocean_menu_typo_font_subset":"","ocean_menu_typo_font_size":0,"ocean_menu_typo_font_size_tablet":0,"ocean_menu_typo_font_size_mobile":0,"ocean_menu_typo_font_size_unit":"px","ocean_menu_typo_font_weight":"","ocean_menu_typo_font_weight_tablet":"","ocean_menu_typo_font_weight_mobile":"","ocean_menu_typo_transform":"","ocean_menu_typo_transform_tablet":"","ocean_menu_typo_transform_mobile":"","ocean_menu_typo_line_height":0,"ocean_menu_typo_line_height_tablet":0,"ocean_menu_typo_line_height_mobile":0,"ocean_menu_typo_line_height_unit":"","ocean_menu_typo_spacing":0,"ocean_menu_typo_spacing_tablet":0,"ocean_menu_typo_spacing_mobile":0,"ocean_menu_typo_spacing_unit":"","ocean_menu_link_color":"","ocean_menu_link_color_hover":"","ocean_menu_link_color_active":"","ocean_menu_link_background":"","ocean_menu_link_hover_background":"","ocean_menu_link_active_background":"","ocean_menu_social_links_bg":"","ocean_menu_social_hover_links_bg":"","ocean_menu_social_links_color":"","ocean_menu_social_hover_links_color":"","ocean_disable_title":"default","ocean_disable_heading":"default","ocean_post_title":"","ocean_post_subheading":"","ocean_post_title_style":"","ocean_post_title_background_color":"","ocean_post_title_background":0,"ocean_post_title_bg_image_position":"","ocean_post_title_bg_image_attachment":"","ocean_post_title_bg_image_repeat":"","ocean_post_title_bg_image_size":"","ocean_post_title_height":0,"ocean_post_title_bg_overlay":0.5,"ocean_post_title_bg_overlay_color":"","ocean_disable_breadcrumbs":"default","ocean_breadcrumbs_color":"","ocean_breadcrumbs_separator_color":"","ocean_breadcrumbs_links_color":"","ocean_breadcrumbs_links_hover_color":"","ocean_display_footer_widgets":"default","ocean_display_footer_bottom":"default","ocean_custom_footer_template":"","ocean_post_oembed":"","ocean_post_self_hosted_media":"","ocean_post_video_embed":"","ocean_link_format":"","ocean_link_format_target":"self","ocean_quote_format":"","ocean_quote_format_link":"post","ocean_gallery_link_images":"on","ocean_gallery_id":[],"footnotes":""},"categories":[286,549],"tags":[],"class_list":["post-6744","post","type-post","status-publish","format-standard","hentry","category-fractualites-ennews-fr","category-seminaire-acmes","entry"],"_links":{"self":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/6744","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/users\/4"}],"replies":[{"embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/comments?post=6744"}],"version-history":[{"count":2,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/6744\/revisions"}],"predecessor-version":[{"id":6746,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/posts\/6744\/revisions\/6746"}],"wp:attachment":[{"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/media?parent=6744"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/categories?post=6744"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/samovar.telecom-sudparis.eu\/index.php\/wp-json\/wp\/v2\/tags?post=6744"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
M. Georgios BOULOUKAKIS<\/strong>, Associate Professor, T\u00e9l\u00e9com SudParis, FRANCE – Co-encadrant de these
Mme Raffaela MIRANDOLA<\/strong>, Full professor, Karlsruhe Institute of Technology, ALLEMAGNE – Examinateur
M. Philippe ROOSE<\/strong>, Professeur, LIUPPA, Universit\u00e9 de Pau et des Pays de l\u2019Adour, FRANCE – Examinateur
M. Thomas LEDOUX<\/strong>, Full professor, IMT Atlantique, FRANCE – Rapporteur
M. Khalil DRIRA<\/strong>, Directeur de recherche, LAAS-CNRS, Toulouse, FRANCE – Rapporteur<\/p>\n\n\n\n\u00ab Autonomie dans les syst\u00e8mes Internet des Objets : une approche d’IA hybride \u00e0 base d’intergiciel pour l’auto-adaptation \u00bb<\/h2>\n\n\n\n
pr\u00e9sent\u00e9 par Monsieur Houssam HAJJ HASSAN<\/h2>\n\n\n\n