" Materialized integration views constructed by integrating data from multiple distributed data sources help to achieve better access, reliable performance, and high availability for a wide range of applications. In this dissertation, we propose parallel, adaptive, and grouping techniques to address scalability challenges in high-performance integration view computation and maintenance due to increasingly large data sources and high rates of source updates. State-of-the-art parallel integration view computation makes the common assumption that the maximal pipelined parallelism leads to superior performance. We instead propose segmented bushy parallel processing that combines pipelined parallelism with alternate forms of parallelism to achieve an overall more effective strategy. Experimental studies conducted over a cluster of high-performance PCs confirm that the proposed strategy has an on average of 50\% improvement in terms of total processing time in comparison to existing solutions. Run-time adaptation becomes critical for parallel integration view computation due to its long running and memory intensive nature. We investigate two types of state level adaptations, namely, state spill and state relocation, to address the run-time memory shortage. We propose lazy-disk and active-disk approaches that integrate both adaptations to maximize run-time query throughput in a memory constrained environment. We also propose global throughput-oriented state adaptation strategies for computation plans with multiple state intensive operators. Extensive experiments confirm the effectiveness of our proposed adaptation solutions. Once results have been computed and materialized, it's typically more efficient to maintain them incrementally instead of full recomputation. However, state-of-the-art incremental view maintenance require O($n^2$) maintenance queries with n being the number of data sources that the view is defined upon. Moreover, they do not exploit view definitions and data source processing capabilities to further improve view maintenance performance. We propose novel grouping maintenance algorithms that dramatically reduce the number of maintenance queries to (O(n)). A cost-based view maintenance framework has been proposed to generate optimized maintenance plans tuned to particular environmental settings. Extensive experimental studies verify the effectiveness of our maintenance algorithms as well as the maintenance framework. "

Ces dernières années, un nouveau type de système de production nommé système de production géographiquement dispersé (GDPS) est prôné par de nombreuses entreprises manufacturières internationales. Par cette vision « dispersée », il présente un certain nombre d'avantages tels que l'économie des coûts du produit livré (puisque proche des clients), l'amélioration de la qualité des services (délais de livraison courts, services après-vente de haute qualité) favorisant la pérennité et la compétitivité des entreprises dans un contexte de compétition mondiale. Cependant l’exploitation multi-sites d’un GPDS est confronté à de nombreux défis concernant les normes, les réglementations, la maîtrise des flux de production, et en particulier la planification et l'optimisation de la maintenance en raison de la dispersion géographique des sites de production. Sur ce dernier point et plus globalement la définition d’une stratégie de maintenance adaptée au GDPS, peu d'études ont été menées compte tenu de la jeunesse du sujet et de la complexité des GDPSs (ex. multi-sites, multi-composants). Cette thèse se positionne donc sur ce sujet émergeant avec comme objectif de développer une stratégie de maintenance de regroupement dynamique pour un GDPS en tenant compte de dépendances à la fois aux niveaux composants et sites de production (dépendances économique et géographique) et des impacts des contextes dynamiques (à savoir, taux de détérioration variable des composants, modification des itinéraires de maintenance, possibilités de maintenance, etc.) auxquels il est soumis. Dans cette stratégie, les itinéraires de maintenance et l'ordonnancement sont considérés conjointement dans un modèle global. Le modèle vise à trouver un plan optimal de maintenance et de routage des ressources de maintenance. A cette fin, une structure de coûts et un modèle de dépendance qui prend en compte conjointement la dépendance économique et géographique sont formulés. Ils servent de base à l'élaboration du modèle global de planification et d'ordonnancement de la maintenance et du routage. De plus, pour la recherche de la solution optimale, des algorithmes d’optimisation basés sur l'algorithme génétique et l'algorithme Branch and Bound sont proposés. Enfin, une étude numérique est investiguée pour évaluer la performance, les avantages et aussi les limites de la stratégie proposée. / In the recent years, the Geographically Dispersed Production System (GDPS) with a number of advantages such as saving the product delivered costs (closed to the clients), improving quality of services (short delivery time, high quality after-sales services) has been extensively developed by many manufacturing companies to ensure their competitiveness. In operation, the GPDS faces many challenges concerning standards, regulation, production management, and especially maintenance planning and optimization due to the geographical dispersion of production sites. However, few studies have been developed for maintenance strategies of GDPSs. To face this challenge, the main objective of this thesis is to develop a dynamic grouping maintenance strategy for a GDPS with consideration of dependencies between at both component and site level (economic, geographical dependencies) and impacts of dynamic contexts (i.e. varying deterioration rate of components, change of maintenance routes, maintenance opportunities, etc.). In this strategy, maintenance routing and scheduling are jointly considered in a global model. The model aims at finding an optimal maintenance and routing plan. For this purpose, a cost structure and a dependence model jointly considering economic and geographical dependence are formulated. They are used as a basis for the development of the global model of maintenance routing and scheduling. In addition, to find a joint optimal maintenance and routing plan, advanced algorithms using jointly Genetic Algorithm and Branch and Bound are proposed. Finally, a numerical study is investigated to evaluate the performance and the advantage as well as limits of the proposed maintenance strategy.