Network Lifetime in Wireless Networks

Mobile wireless networks offer numerous possibilities for applications in the current society. Since the radio transmitters become smaller, more and more mobile networks in which persons carry the mobile devices are developed. An example for such an network is given by a data communication network for soldiers or emergency workers, in which positions are exchanged in order to allow all persons to know the positions of their mates. Since each person has to carry its own communication equipment, every decrease in weight of the mobile communication equipment is useful. Since the energy supply, the batteries, forms a main part of the weight of the communication equipment, one aspect of the research to decrease the weight is concerned with developing highly energyefficient communication algorithms that maximize the ‘network lifetime’. In this thesis, we explore the Network Lifetime Problem (NLP). The goal is to maximize the network lifetime, which is defined as the time until the first node runs out of energy. The first part of the thesis gives a classification methodology for algorithms and problems concerning the NLP. This methodology gives insight in the background of the NLP, proposed solutions for the NLP, and also information to compare algorithms fairly. Furthermore, some of the existing algorithms for the NLP are discussed. The main conclusion of the first part is that every algorithm is basically designed for one specific problem. The second part of this thesis is devoted to Multipoint Relay (MPR) flooding and the ability of MPR flooding to maximize the network lifetime. MPR flooding is a technique that optimizes flooding by reducing the number of retransmissions. In a simple flooding approach all nodes have to relay a packet, but in MPR flooding only the selected MPRS. We present a study on MPRs, a study on MPR flooding and a study on MPR selection algorithms, instead of one global view on MPR flooding. One of the results show that in networks in which the nodes are positioned at a grid, the resulting set of selected MPRs of a non-border node is independent of the MPR selection algorithm. We also prove that in several networks the network lifetime cannot be improved and is fixed, independent of the used MPR selection algorithms, because of the definition of MPRs. Based on this result, we study the relation between the network structure and the effect of MPR flooding with a specific MPR selection algorithm on the network lifetime. We show that the performance of MPR flooding depends on a special property of the network, the so called ‘Maximum Forcedness Ratio’. The idea behind the ‘Forcedness Ratio’ is also used in a proposed MPR selection algorithm, the Maximum Willingness and Minimum Forced MPR selection algorithm. This algorithms is comparable with the Maximum Willingness MPR selection algorithm, as used in OLSR, in terms of network lifetime.