Um sistema para sugestão e otimização de conteúdo aplicado ao servidor multimídia do SBTVD

Abstract

The open analogic television arrives on approximately 97% of Brazilian homes. With the advent of the Brazilian Digital Television System (SBTVD) users will be able to have an interactive channel. That way, users can access the multimedia applications server, to send or receive emails, access interactive applications, watch movies or specific news. The Digiconv Platform aims to generate digital content to Digital TV, IPTV and mobile devices. This work presents the development and implementation of an architecture divided into two modules, the content suggestion module aims to suggest content to the user, according to his profile, through the use of software agents and ontologies. In this module, a system was developed using content-based filtering, thus the system find similar content to suggest according to the content that the user liked in the past. The content transmission module aims to optimize the transmission parameters, which is addressed as a combinatorial optimization problem known as Part-Selection Problem that is classified as NPComplete. In this module, were used software agents, Tabu Search, Genetic Algorithm and Hybrid Algorithm with the goal to find the best transmission policy of the video parameters (LD, P1, SD, HD and P2) and the audio parameters (stereo audio and 5.1 audio). The validation of this module is done through the harmonic mean of the distribution of the internet link available on the server. The result of this work is a system with two modules, one module suggest content and the other module defines the best content transmission policy.