We present in this paper a task allocation mechanism constrained by a distributed, complex environment namely in our case a home network. Thus, added to multi agent systems constraints, we have restrictions in terms of communication (available throughput) and resource (bandwidth). We have a multi agent system with a knowledge base, and we have to ensure a proper quality of service in home Network adjusting the routing in real time by applying alternative route to the main ones. Our approach uses the first price sealed bid type auction: when an agent does not have an alternative route, it launches an auction. The agent offering the best price will be the next hop of the route.