Using genetic algorithms to optimize the behaviour of adaptive multimedia applications in wireless and mobile scenarios

Adaptive applications are a key concept to take into account when dealing with multimedia internetworking in wireless and mobile environments in which abrupt changes in the network conditions may occur. In these scenarios not all the problems in the multimedia quality are due to congestion as it happens in fixed networks but also to interference, mobility, and many other issues. This paper presents an adaptation architecture allowing applications to use adaptation techniques at the application layer to minimize the impact of adverse and changing network conditions. This allows maintaining the QoS perceived by the user in an acceptable level. In addition, we optimize the adaptive behavior using genetic algorithms to estimate the optimal triggers for the adaptation function. We present some results comparing the fitness of using such adaptive techniques against the traditional multimedia internetworking.