Modelling and manipulation of complex and historical data warehouses

This thesis deals with conceptual modelling and data manipulation (using algebras) in decision support systems. Our thesis is based on the dichotomy of two storage spaces: the data warehouse contains extracts from source databases that are useful to decision-makers, and data stores are derived from the warehouse and dedicated to a specific analysis need. At the warehouse level, we define a data model that describes the temporal evolution of complex objects. In our proposal, the warehouse object integrates current, past and archived states, modelling decision-making data and its evolution. The extension of the object concept leads to an extension of the class concept. This extension consists of filters (temporal and archival) to construct past and archived states, as well as a construction function modelling the extraction process (source origin). We also introduce the concept of environment, which defines coherent temporal parts of sizes adapted to the requirements of decision-makers. Data manipulation is an extension of object algebras that takes into account the characteristics of the warehouse representation model. [...]