Site selection for transmitters in wireless networks is a complex, time-consuming process. Most often, transmitters are located in one of two ways: manually, or through the use of simple geometric models. Unfortunately, each of these methods disregards the most important geographic information aecting the performance of transmitters. This paper introduces a hybrid genetic algorithm designed to automate the site selection process and to account for the relevant geography. The algorithm has

more »

... l features not found in other genetic models. These features are introduced, speci®cally, to facilitate the design of wireless networks: a unique genetic make-up representing the complex wireless networks: a combination of genetic and deterministic search operators to optimize and speed up search, and spatial operators necessary for the accurate network design in the context of a genetic model. The paper serves three functions: (1) description of the fundamental concepts underlying the design of the algorithm; (2) discussion of the details of representation of wireless networks in an evolutionary context; and (3) demonstration of the performance of the algorithm with several network models in a complex spatial environment. #