Multicast Routing in Wireless Sensor Networks with Incomplete Information

In this paper we develop a set of heuristic algorithms for the solution of the Quality of Service (QoS) constrained multicast routing problem with incomplete information in Wireless Sensor Networks (WSN). The QoS constrained multicast routing has already been analyzed and proven to be NP-complete with deterministic link measures. In this paper we investigate this problem in the case of incomplete information when, in aWireless Sensor Network due to information aggregation or randomly fluctuating traffic loads, link measures are considered to be random variables. We develop novel routing algorithms based on these random link metrics. The objective is to choose a tree which fulfills a given QoS criterion (e.g. delay) with maximum probability. We propose a set of algorithms based on Hopfield Neural Networks to solve the QoS constrained multicast routing problem with incomplete information. In this way, a fast tree selection becomes possible.