New Optimization Models for Directional and Biological Wireless Sensor Networks.

摘要

Wireless Sensor Networks (WSNs) are embedded networks made up of tiny wireless devices referred to as sensors. They can be distinguished based on the type of sensing elements used in their sensors. Two important classes of WSNs are directional and biological WSNs. The former is given its name because the sensing region of a directional sensor is viewed as a sector in a two-dimensional plane. The latter, however, is given its name because its sensing elements are biological materials like enzymes.;The first objective of the work presented in this dissertation is to develop optimization models for the planning of directional WSNs. Toward that end, the existing literature is critiqued and gaps are identified. Then, three fundamental planning problems are presented. The problems are formulated as Integer Linear Programming (ILP) models. The wide applicability of the proposed models and their effectiveness are also discussed.;The second objective in this dissertation is to develop stochastic optimization models for the control of biological WSNs. The framework of Markov Decision Processes (MDPs) is used for building the optimization models. The solution of an MDP model is a policy which can be used to operate the biological WSNs in such a way that the maximum safe temperature level is never exceeded. When compared to existing policies, the policies produced by the proposed MDP models are superior in terms of network lifetime and temperature increase. Several techniques for handling large-size MDP models are also investigated.;One of the major challenges introduced by directional WSNs is the fact that the existing mathematical models developed for conventional WSNs cannot directly be used for solving the directional sensor placement and configuration problems. Therefore, new optimization models which capture the primary parameters characterizing directional sensors are necessary. On the other hand, a major challenge introduced by biological WSNs is the heat generated as a result of power consumption and wireless radiation. When biological WSNs are operated in temperature-sensitive environments like the human body, the temperature of the surrounding tissues might rise. The tissues might be damaged if the maximum safe temperature level is exceeded. Therefore, thermal management techniques are indispensible.