Amadeus S.a.s.

Guillaume Le Grand

Jean-Sebastien Thireau

Olivier Amadieu

Jean-Marie Cazorla

Yannick Devaux

Jean-Philippe Aubry

Eric Jean Joseph Martin

Johann Marchi

Nicolas Legrand

Guillaume Lecourt

Blerina Topciu

A method at an aggregator includes: storing previous search results resulting from previous client search requests, and for each previous search result, a previous handling indicator, indicating a relevance of the previous search result to the client; receiving, from the client, a search request containing search parameters; in response to the search request, selecting a subset of previous search results based on correspondence between attributes of the previous search results and the search parameters, and on the previous handling indicators; providing, to a supplier, the search request and auxiliary search parameters corresponding to the selected previous search results and indicating characteristics of the selected previous search results, for generation of current search results at the supplier employing the auxiliary search parameters as inputs; receiving, from the supplier, the current search results generated at the supplier; and returning at least one of the current search results to the client.

Computational load mitigation with predictive search request enrichment

Pre-computed query results stored at a database server are re-computed by a computation machine. The pre-computed query results are composed of a plurality of shares. Each share includes a certain number of pre-computed query results. The computation resources of the computation machine needed to re-compute a pre-computed query result of a first share depend on whether or not other pre-computed query results of the first share are re-computed during a given time interval forming a current re-computation cycle. The computation machine receives a request to re-compute pre-computed query results of at least a first share, determines a number of pre-computed query results in the first share to be re-computed in the current re-computation cycle based on a Reinforcement Learning algorithm, and re-computes the determined number of pre-computed query results in the first share.

Re-computing pre-computed query results

Search requests are processed in a distributed database environment which includes a frontend module, one or more clients, a cache, a live data source, and multiple data records. Each data record includes at least one stable data piece stored in the cache and at least one volatile data piece. The frontend module receives a search request from a client and retrieves from the cache a set of data records. Each stable data piece in the set of data records corresponds to search criteria indicated in the search request. Current volatile data pieces, which correspond to the stable data pieces of the retrieved set of data records, are retrieved by the frontend module from the live data source for the retrieved set of data records. A completed set of data records is generated by combining the stable data pieces and the current volatile data pieces, and returned to the requesting client.

Processing search requests

Growing community of inventors

Le Cannet, France

Guillaume Le Grand